1998-99EVALUATION RESULTS OF

THESTUDENT ACHIEVEMENT GUARANTEE

INEDUCATION (SAGE) PROGRAM

Submitted by the SAGE Evaluation Team

School of EducationS

University of Wisconsin–Milwaukee

Alex Molnar

Philip Smith

John Zahorik

Research Assistants

Lee Breese

Karen Ehrle

Anke Halbach

Amanda Palmer

Betsy Schoeller

DECEMBER 1999

Forfurther information, contact Alex Molnar, School of Education

Universityof Wisconsin–Milwaukee, P.O. Box 413, Milwaukee, WI 53201, (414) 229-2220.

This document is available on theSAGE Website: http://www.uwm.edu/SOE/centers&projects/sage/

TABLEOF CONTENTS

INTRODUCTION

SAGE Program……………………………………………………………………………3

Goals of SAGEEvaluation………………………………………………………………..4

Class Size ResearchBackground………………………………………………………….5

Summary of 1996-97 and 1997-98Findings……………………………………………...7

1998-99 SAGE EVALUATION

Descriptions andDefinitions…………………………………………………………….11

Data CollectionInstruments……………………………………………………………..15

ANALYSES OF STUDENT ACHIEVEMENT OUTCOMES 1998-99

Methods Introduction……………………………………………………………….……19

SAGE School/Classroom vs. ComparisonSchool/Classroom Analyses…………….…..22

Effect Sizes………………………………………………………………………………55

Additional Analyses………………………………………………………….…………..56

ANALYSES OF SAGE CLASSROOMS AND SCHOOLS 1998-99

Teaching……………………….……………………………………………….………...62

Rigorous Curriculum…………………………………………………………………...108

ProfessionalDevelopment……………………………………………………………...117

Family Involvement and Lighted Schoolhouse..……………………………………….120

DISCUSSION: MAJOR FINDINGS, LIMITATIONS, AND FUTURE SAGE REPORTS

Major Findings……………………………………………………………….…………124

Limitations……………………………………………………………………………...128

Future SAGE EvaluationReports………………………………………….…………...128

INTRODUCTION

SAGE Program

The Student AchievementGuarantee in Education (SAGE) evaluation is being conducted under contract withthe Department of Public Instruction by the School of Education at theUniversity of Wisconsin–Milwaukee. The purpose of the SAGE evaluation is to determinethe effectiveness of the Student Achievement Guarantee in Education (SAGE)program in promoting academic achievement of students in kindergarten throughthird grade classrooms in schools serving low-income children.

The 1995 SAGE statute [s.118.43] requires participating schools to (1) reduce class size to 15 inkindergarten and grade one in 1996–97, grades kindergarten through two in1997–98, and grades kindergarten through three in 1998–99 to 2000–2001; (2)stay open from early in the morning to late in the day and collaborate withcommunity organizations to provide educational, recreational, community, andsocial services (i.e., the "lighted schoolhouse"); (3) provide arigorous academic curriculum to improve academic achievement; and (4) establishstaff development and accountability mechanisms.

The SAGE evaluationinvolves the 30 schools in 21 school districts that launched the SAGE programin 1996-97 in kindergarten and first grade. Second grade was added in 1997-98,and third grade in 1998-99. The SAGE evaluation compares SAGE schools to agroup of 14-17 non-SAGE Comparison schools located in SAGE districts. Theresults of the 1996-97, 1997-98 and 1998-99 evaluations are generallyconsistent with Tennessee’s Student Teacher Achievement Ratio (STAR) Project(1985-1989), the largest and best-controlled study on class size reduction tothis point. It is worth noting two significant design differences in theTennessee and SAGE studies. First, the Tennessee STAR Project used a trueexperimental design. The SAGE project uses a quasi-experimental design. TheSAGE project evaluation uses naturally occurring classrooms while STAR employedrandom assignment of students to classroom types which were held constant forthe duration of the study. Second, the SAGE evaluation uses a control orcomparison group for the purpose of assessing the impact of SAGE interventions.

Goals of SAGE Evaluation

The SAGE evaluation is intended to determine the impact that the fourinterventions of the SAGE program have on student achievement. To ascertain andto explain this impact, the evaluation addresses the following questions:

SAGE vs. Comparison School – AchievementOutcome Questions

1. What are the achievement levels of SAGE classrooms compared to achievement levels of classrooms in selected Comparison schools?
2. Does participation in a SAGE classroom have a differential impact on the achievement of minority students and white students?
3. Does the impact on achievement of participation in a SAGE classroom change from year to year as students progress from first through third grade?
4. Is the socio-economic status (as measured by participation in the school lunch program) of SAGE participants related to individual achievement gains in first through third grade?
5. Do different types of SAGE classrooms (e.g. one teacher with 15 students vs. two teachers with 30 students) have different impacts on student achievement?

SAGE Schools – Classroom andSchool Questions

1. What are the instructional characteristics of SAGE classrooms?
2. How are SAGE classrooms organized?
3. Does the degree of congruence between SAGE classroom curricula and national professional curriculum standards in reading/language arts and mathematics correlate with the achievement levels in SAGE classrooms?
4. Does participation in the SAGE program result in an increase in the number or change in the type of school and/or community activities housed in the school before and after school hours?
5. What is the structure and focus of professional development activities in SAGE schools?
6. Does the number of years of teaching experience of SAGE teachers correlate with student achievement?

Class Size Research Background

The principal SAGEintervention is a reduction in class size to 15:1 in kindergarten throughthird-grade classrooms. Class size reduction in the early elementary grades hasbecome an increasingly popular issue for policymakers and researchers in recentyears (Grissmer, 1999; U. S. Department of Education and the Laboratory forStudent Success, 1999). Class size reduction has been credited with morelearning opportunities for students, increased opportunities for teachers tomeet children’s individual needs, and less time spent on discipline problems.Parents and teachers like the idea and policymakers are embracing it. Severalstates, among them California, Florida, Indiana, Nevada, Tennessee, andWisconsin have launched class size reduction efforts (Pardini, 1998; Viadero,1998).

Decades of research onclass size reduction have shown small achievement gains for students when, forexample, class size was reduced from 25 to 20 students. In general, though,reductions of just a few students per class do not seem to significantly raiseacademic achievement. However, in the late seventies, an analysis by Glasssuggested that larger reductions produced greater achievement gains (Glass,1978; Pate-Bain, Achilles, Boyd-Zaharias, & McKenna, 1992).

A statewide experiment inTennessee, the largest and best-controlled study on class size reduction tothis point, assigned kindergarten students on a random basis to classes of 15,25 with a teacher and an aide, or 25 with a teacher. The same configurationswere maintained through third grade. Tennessee’s Project STAR (Student/TeacherAchievement Ratio) analyzed student achievement in relation to class size overa four-year period (1985-1989). The project included 17 inner-city, 16suburban, 8 urban, and 39 rural schools. Findings showed that students in thesmaller classes scored higher on the Stanford Achievement Test and on the BasicSkills First (BSF) Test in all four years (K-3) and in all locations. Thegreatest gains on the Stanford Test were made by inner-city small classes.While all students benefited, disadvantaged minority students seemed to benefitmore from smaller class sizes than their peers (Pate-Bain, Achilles,Boyd-Zaharias, & McKenna, 1992).

Studies such as STAR andSAGE can provide crucial information for policymakers. For example, a review ofthe research literature conducted by Bingham (1993) on white-minorityachievement gap reduction and small class size, which included the STARProject, suggests that small class sizes in the early grades represent an effectivestrategy to reduce the achievement gap. Bingham proposes that class sizereduction may offer an early intervention strategy that serves to preventrather than to reduce the achievement gap between white and minority students.            Wenglinsky (1997) studied therelationship between spending and student achievement by analyzing data fromthree separate sources: The National Assessment of Educational Progress, theCommon Core of Data, and the Teacher’s Cost Index of the National Center forEducation Statistics. Wenglinsky’s research suggests that increased spending toreduce class size has a direct positive effect on mathematics achievement forfourth grade students. Further support for small classes in lower elementarygrades is produced by the Lasting Benefits Study (LBS). The LBS trackedstudents who participated in Project STAR in order to determine whetherachievement advantages of students from small classes were maintained afterstudents returned to regular-sized classes in fourth grade. Data from 1990-1994indicate that students who were originally in smaller classes continued toperform better than their peers from classes of 25 with or without a teacher’saide (Mosteller, 1995).

Grissmer (1999) notes thatultimately our confidence in measurements of small class size effects willarise from knowledge of the specific processes inside classrooms. Class sizeeffects are better understood if we can discern the processes used in theseclassrooms that affect student development and achievement. Grissmer advisesresearchers to address fundamental questions about the nature of evidence insmall class size studies. For policy decisions, he notes that "the seemingtransparency of experimental data to policymakers should not be the decidingfactor in their policy judgments" (p. 93). Grissmer further observes thathaving both experimental and non-experimental evidence is rare in theeducational research community. The SAGE evaluation project includes bothexperimental and non-experimental data. Since its inception, the design of theSAGE evaluation project has called for data from student tests and processes inSAGE classrooms. Analyzing processes inside small classrooms with regard toteacher and student behavior have been a crucial part of the SAGE evaluationproject and are reported in the evaluation results along with studentachievement data.

Grissmer (1999), from ananalysis of small class size research, concludes that a key contribution ofexperimental data can be their guiding role in non-experimental studies todevelop a theory of classroom teacher and student behavior that explains higherstudent achievement. In Wisconsin, the SAGE evaluation team has alreadyincorporated this recommendation in its research on class size effects.

Summary of 1996-97 and 1997-98Findings

Achievement Outcome Findings1996-97 and 1997-98

To measure academicachievement, first grade students in SAGE schools and in a group of Comparisonschools were tested in October 1996, and again in May 1997, using theComprehensive Test of Basic Skills (CTBS) Complete Battery, Terra Nova edition,Level 10. After one year, students inSAGE first grade classrooms scored higher on CTBS tests than first gradestudents in Comparison schools. As a group, SAGE students scored significantlyhigher on the post-test in reading, language arts, and mathematics sub-tests ofthe CTBS. The total score of SAGE students was also significantly higher thanthe total score of comparison group students. The achievement advantageassociated with participation in the SAGE program was revealed both in theanalysis of individual student scores and in the analysis of averaged classroomscores.

At the individual level ofanalysis, after controlling for pre-test scores, income, absenteeism, and raceand ethnicity, SAGE first grade students scored higher than Comparison schoolfirst grade students on the CTBS post-test in reading, language arts,mathematics and total score. The results were statistically significant for allbut the reading scores. At the classroom level, the post-test performance ofSAGE first grade students was 4 scale score points higher in language arts, 4.3scale score points higher in reading, 4.6 scale score points higher inmathematics, and 4.6 scale score points higher in the total test score thanComparison school students. Each of these findings was statisticallysignificant.

After adjusting forindividual pre-test results, socioeconomic status (SES) as defined byeligibility for subsidized lunch, and student attendance, participation in SAGEshows a statistically significant advantage of 6.4 scale score points in thetotal score and 8.1 scale score points on the mathematics sub-test.

The classroom level data onthe averaged performance of students in 1996-97 and 1997-98 SAGE classroomssuggested that the lower student-teacher ratio in SAGE classrooms mitigated thenegative achievement consequences of poverty. SAGE classrooms achieved at ahigher level than Comparison school classrooms despite the fact that, as agroup, SAGE classrooms enrolled more students who were eligible for subsidizedlunch. Furthermore, after adjusting for individual pre-test results and SES asdefined by lunch status and student attendance, the post-test scale scoreadvantage increased to 9.8 for SAGE first grade classrooms. The advantage was7.1 on the reading sub-test, 9.0 on the language arts sub-test, and 12.3 on themathematics sub-test. These results were all statistically significant.

Second grade classroomswere looked at in 1997-98, and results were similar to those found for 1996-97first graders. The 1997-98 results suggest that the positive effects of theSAGE program are maintained, but not increased in second grade.

School and Classroom Findings1996-97 and 1997-98

To more fully understandthe SAGE program, it is important to understand how SAGE schools structuredclassrooms and implemented a reduced student-teacher ratio, rigorouscurriculum, staff development, and lighted schoolhouse. Together, thatinformation provides a within-SAGE school and classroom data description oflife in SAGE classrooms and schools.

Classroom Level Findings

Data from 1996-97 and1997-98 suggested that the main change that results from having a reduced sizeclass is individualization. Teachers focus on individual learning needs throughone-to-one, small groups, and total class teaching. This focus on individualswas brought about by knowing students better, having more time for teachingbecause of reduced need for discipline, and being more enthusiastic aboutteaching , all which result from having fewer students.

The type of instructionthat students encountered in SAGE classrooms was predominantly teachercentered. Listening, practicing, receiving help, and answering accounted forthe main portion of the teaching-learning that occurred. Although teachersindicated that more student-centered activities such as creating, manipulating,and problem solving increased somewhat because of reduced class size,student-centered teaching played only a supplemental role in most SAGEclassrooms.

In regard to teaching andlearning among the four main types of SAGE classes and between grades one andtwo, few differences were observed.

School Level Findings 1996-97 and 1997-98

The TeacherQuestionnaire and Principal Interviews, both completed in May 1997 and May1998, were the sources of data regarding rigorous curriculumimplementation. The Teacher Questionnaire contained a section on classroomcurriculum designed to determine the congruence of SAGE classroom curriculawith professional curriculum standards. First grade and second grade teacherresponses indicated that their reading/language arts curriculum and mathematicscurriculum were quite congruent with professional standards. Principalresponses to curriculum-related questions suggested that a rigorous curriculumincluded basic skills, problem solving, and higher-level thinking. Only a fewprincipals seemed to believe that the curriculum of their school was rigorousin total. However, most SAGE principals regarded parts of their curriculum asstrong. A section of the Teacher Questionnaire contained staff developmentquestions. Teachers were asked about their individual level of professionaldevelopment as well as the extent to which their school district provided staffdevelopment programs. About 60 percent indicated that they had a writtendevelopment plan and it was determined by the teachers themselves. Dataregarding implementation of lighted schoolhouse activities for 1996-97and 1997-98 were obtained from the Principal Interviews and year-end reportsrequired by DPI. Principal Interview data suggested that SAGE schools tookresponsibility for the conception and operation of the lighted schoolhouseactivities and that the number of lighthouse activities and number ofparticipants in the activities has progressively increased.

1998-99SAGE EVALUATION

Descriptions and Definitions

Schools

During 1998–99, the SAGEprogram was continued in 30 schools located in 21 school districtsthroughout the state, as shown in Table 1. In addition, the SAGE evaluationconsisted of 17 Comparison schools located in 11 SAGE school districts.

Table 1. SAGE Schools 1998-99

SAGE DISTRICTS AND SCHOOLS
DISTRICT School	DISTRICT School
ADAMS-FRIENDSHIP AREA Adams Elementary	MENOMONEE INDIAN Keshena Primary
BELOIT Robinson Elementary	MENOMONEE AREA River Heights Elementary
CUDAHY Parkview Elementary	MILWAUKEE PUBLIC SCHOOLS Carleton Elementary Fairview Elementary Longfellow Elementary Maple Tree Elementary Maryland Avenue Elementary Sherman Elementary Wisconsin Conservatory
GILMAN Gilman Elementary
GLIDDEN Glidden Elementary
GREEN BAY AREA Jefferson Elementary
JANESVILLE Wilson Elementary	PRENTICE Ogema Elementary Tripoli Elementary
KENOSHA Durkee Elementary	SIREN Siren Elementary
LAC DU FLAMBEAU #1 Lac Du Flambeau Elementary	SOUTH SHORE South Shore Elementary
LACROSSE Franklin Elementary Hamilton Elementary	SUPERIOR Blaine Elementary Cooper Elementary
LAONA Robinson Elementary	SURING Mountain Elementary
MADISON METROPOLITAN Glendale Elementary	WEBSTER Webster Elementary

Students

In 1998-99, the SAGEevaluation involved 9,876 active students in kindergarten (2,303), first grade(2,508), second grade (2,493), and third grade (2,572 ). The characteristics ofstudents in SAGE and Comparison schools are displayed in Table 2.

Table 2. Characteristics of SAGE and Comparison Students 1996-97,1997-98 and 1998-99

Characteristic	Percent of Students SAGE			Percent of Students Comparison
	96-97	97-98	98-99	96-97	97-98	98-99
Gender
Female	48.6	48.4	48.6	49.4	48.5	48.7
Male	51.4	51.6	51.4	50.6	51.5	51.3
Race/Ethnicity
African American	24.8	26.3	22.4	32.9	24.7	19.7
Asian	5.7	5.2	4.8	5.5	5.6	5.9
Hispanic	6.6	6.5	6.4	8.0	10.0	9.5
Native American	11.7	10.3	10.9	1.4	1.5	1.5
White	48.8	43.8	44.2	49.0	52.2	53.4
Other	1.6	2.0	1.8	2.7	2.3	2.3
Subsidized Lunch Eligibility
Free	57.7	54.0	52.7	49.4	43.4	40.7
Reduced	10.9	10.6	11.5	9.9	8.9	10.4
Not Eligible	31.4	35.4	35.8	40.7	47.7	48.8
Repeating Grade	3.2	2.7	2.0	2.6	2.0	1.5
English as Second Language	8.2	7.9	7.5	4.9	6.4	6.7
Referred to M-Team	13.6	9.6	12.7	9.2	6.8	9.1
Exceptional Education Need	13.1	10.0	12.7	9.7	7.1	1.3

During the course ofthe 1998-99 school year, records were compiled on 9,876 students. Many studentswithdrew from SAGE and Comparison schools during the year, while othersenrolled. Those students who remained in their schools for the entire year arelabeled "persisters". The number of students in SAGE and Comparisonschools by grade and school year can be seen in Table 3.

Table 3. Number of Students in SAGE and Comparison Schools by Gradeand School Year

	SAGE			COMPARISON
	1996-97	1997-98	1998-99	1996-97	1997-98	1998-99
Kindergarten	1494	1524	1416	820	676	887
First Grade	1723	1567	1525	1001	985	983
Second Grade	NA	1541	1446	NA	868	1047
Third grade	NA	NA	1531	NA	NA	1041
Totals	3217	4632	5918	1821	2529	3958

As Table 4 shows, the stability of student enrollment is not clear. Althoughthe percentage of ongoing SAGE students was higher than ongoing Comparisonstudents and the percentage of SAGE students withdrawing was lower thanComparison students, the percentage of Comparison students enrolled wasconsiderably higher than SAGE students.

Table 4. Enrollment Changes in SAGE and Comparison Schools by SchoolYear (Percentage of Students)

	SAGE			COMPARISON
	1996-97	1997-98	1998-99	1996-97	1997-98	1998-99
Ongoing	81.4	42.3	44.8	85.3	44.3	39.1
Withdrew	11.0	18.8	27.5	8.9	18.6	20.2
Enrolled	7.6	39.0	27.6	5.8	37.0	40.6

Classrooms

SAGE schools reduced classsize in several ways in order to meet statutory requirements. The SAGElegislation defines class size as "the number of pupils assigned to aregular classroom teacher." In practice, reduced class size has beeninterpreted as a 15:1 student-teacher ratio (number of students per teacher inone classroom). Implementation occurs in the following ways:

•           A Regular classroom refers toa classroom with one teacher. Most regular classrooms have 15 or fewerstudents, but a few exceed 15.

•           A 2-Teacher Team classroom isa class where two teachers work collaboratively to teach as many as 30students.

•           A Shared-Space classroom is aclassroom that has been fitted with a temporary wall that creates two teachingspaces, each with one teacher and about 15 students.

•           A Floating Teacher classroomis a room consisting of one teacher and about 30 students, except duringreading, language arts, and mathematics instruction when another teacher joinsthe class to reduce the ratio to 15:1.

Two other types of classroom organization were also utilized in the SAGEprogram, but to a limited extent. They are the Split Day classroomconsisting of 15 students and two teachers, one who teaches in the morning andone who teaches in the afternoon, and the 3-Teacher Team classroom wherethere are 45 students taught collaboratively by three teachers.

The types of classrooms aredisplayed in Table 5. SAGE classes range in number of students from 7 to 38. Afew SAGE classrooms exceed the 15:1 student-teacher ratio by a few students.The teacher to student ratio for SAGE and comparison classrooms can be seen in

Table 6.

Table 5. Number of SAGE Classrooms by Type and Grade 1998-99

	Regular	2-Teacher Team	Floating Teacher	Shared Space	Split Day	3-Teacher Team	1 Full Time, 1 Part Time Teacher
Kindergarten	65	16	3	0	0	1	0
Grade 1	64	18	0	2	0	0	5
Grade 2	59	18	2	0	0	0	4
Grade 3	66	17	0	4	0	0	1
Mixed K-1	3	0	0	0	0	0	0
Mixed 1-2	2	0	0	1	0	0	0
Mixed 2-3	5	0	0	0	0	0	0

Table 6. Teacher-Student Ratio for SAGE and Comparison Classrooms1998-99 (number of classrooms)

	SAGE Classrooms				Comparison Classrooms
Students Per Teacher	K	First Grade	Second Grade	Third Grade	K	First Grade	Second Grade	Third Grade
7-13 Students	45	58	36	36	5	3	4	4
14-16 Students	33	32	48	48	6	6	6	4
17+ Students	8	4	6	6	34	40	43	41
Average Class Size	13.21	12.82	14.02	14.02	19.04	19.63	20.04	20.51

Data Collection Instruments

Toprovide information about the processes and product of the SAGE program for1996-97, 1997-98, and 1998-99 a number of instruments were used as part of theevaluation. A description of the test and non-test instruments used in 1996-97,1997-98, and 1998-99 follows. The data collection instruments and the plan fortheir use throughout the evaluation are displayed in Table 7.

1. Comprehensive Test of Basic Skills (CTBS) (1996-97, 1997-98, 1998-99). The Comprehensive Test of Basic Skills (CTBS) complete Battery, Terra Nova edition, Level 10, was administered to first grade students in SAGE schools and Comparison schools in October 1996 and May 1997. In 1997-98, Level 10 was administered in October and Level 11 in May to first grade students and Level 12 to second grade students. In 1998-99, Level 10 was administered in October and Level 11 in May to first grade students, Level 12 to second grade students, and Level 13 to third grade students. The purpose of the first grade October administration of the CTBS was to obtain baseline measures of achievement for SAGE schools and Comparison schools. The complete battery includes sub-tests in reading, language arts, and mathematics. The CTBS was chosen as an achievement measure because it is derived from an Item Response Theory (IRT) model that allows comparison of performance across time. Moreover, it is one of a few instruments that attempts to minimize items biased against minorities and educationally disadvantaged students. Kindergarten students were not tested because of (1) concerns over the reliability and validity of standardized test results for kindergarten-aged children and (2) the view expressed by many kindergarten teachers that standardized tests would have a traumatizing effect on their students. The effects of SAGE on kindergarten students will be determined when they are tested as first- grade students the following year.

2. Student Profiles (1996-97,1997-98, 1998-99). This instrument completed in October and May,provided demographic and other data on each SAGE school and Comparison schoolstudent.

3. Classroom Organization Profile(1996-97, 1997-98, 1998-99). Completed in October, this instrument was used torecord how SAGE schools attained a 15:1 student-teacher ratio.

4. Principal Interviews (1996-97,1997-98, 1998-99). These end-of-year interviews elicited principals'descriptions and perceptions of effects of their schools' rigorous curriculum,lighted-schoolhouse activities, and staff development program, as well as anoverall evaluation of the SAGE program.

5. Teacher Questionnaire (1996-97,1997-98, 1998-99). Administered in May, this instrument obtained teachers'descriptions and judgments of the effects of SAGE on teaching, curriculum,family involvement, and professional development. It also was used to assessoverall satisfaction with SAGE.

6. Teacher Activity Log (1996-97,1997-98). This instrument required teachers to record classroom eventsconcerning time use, grouping, content, and student learning activities for atypical day three times during the year.

7. Student Participation Questionnaire(1996-97, 1997-98). In both October and May, teachers used this instrument toassess each student's level of participation in classroom activities.

8. Classroom Observations (1996-97,1997-98). A group of first grade, second and third grade classroomsrepresenting the various types of 15:1 student-teacher ratios and a range ofgeographic areas was selected for qualitative observations to providedescriptions of classroom events.

1. Teacher Interviews (1996-97, 1997-98). Although in-depth teacher interviews were not part of the original SAGE evaluation design, they were added in 1997 because it became apparent that teachers had important stories to tell about their SAGE classroom experiences. The interviews deal with teachers' perceptions of the effects of SAGE on their teaching and on student learning.
2. Principal Questionnaire (1998-99). The Principal Questionnaire was administered to all SAGE principals in spring beginning in 1998-99. It requested them to rate and comment on teaching, rigorous curriculum, staff development, and lighted schoolhouse activities.
3. Case Studies (1998-99). Case studies of teaching in three schools each representing a different type of SAGE class configuration were conducted continuously throughout the school year in 1998-99. At grades one, two and three classrooms were observed in reading-language arts instruction and mathematics instruction and teachers were interviewed. Interviews with the principal and parents were also conducted.

Table 7. SAGE Data Collection by Grade Level, 1996–01

	1996–97	1997–98	1998–99	1999-2000	2000-2001
CTBS Fall, Spring Spring	1	1 2	1 2, 3	2, 3	3
Student Profiles Fall, Spring	1	1, 2	1, 2, 3	2, 3	3
Classroom Organization Profile Fall	1	1, 2	1, 2, 3	2, 3	3
Principal Interviews Spring	yes	yes	yes
Teacher Questionnaire Spring	K, 1	K, 1, 2	K, 1, 2, 3	2, 3	3
Teacher Activity Log Fall, Winter, Spring	K, 1	K, 1, 2	discontinued
Student Participation Questionnaire Fall, Spring	K, 1	K, 1, 2	discontinued
Classroom Observation Fall, Spring	1 (Selected)	1, 2, (Selected)	1, 2, 3 (Selected)	2, 3 (Selected)	3 (Selected)
Teacher Interview Spring	1 (Selected)	1, 2 (Selected)	1, 2, 3 (Selected)	2, 3 (Selected)	3 (Selected)
Principal Questionnaire Spring	NA	NA	yes	yes	yes
School Case Study Continuous	NA	NA	1, 2, 3 (Selected)	2, 3 (Selected)	3 (Selected)

ANALYSES OF STUDENT ACHIEVEMENT OUTCOMES 1998-99

Methods Introduction

Statistics Utilized

The1998-99 SAGE evaluation design utilizes descriptive statistics and multivariateinferential statistics, including linear regression and hierarchical linearmodeling. Descriptive statistics, including means and standard deviations, areincorporated into this report to provide a less complicated, general analysiswhich the non-technical reader can use as a basis to interpret the findings.Regression analyses (at the individual level), specifically the use of ordinaryleast squares regression models, are employed frequently in the 1998-99 report.Regression models enable "control" variables to be entered in blockswith the variable of interest, i.e. the "SAGE/Comparison" variableentered last, thus isolating its effects from the other variables. Hierarchicallinear modeling is pertinent to the SAGE evaluation because this techniquepermits a focus on the class effects of SAGE; that is, these analysesspecifically assess classroom effects rather than those of individuals withinthe classroom. The classroom effects examined by this approach are of primaryimportance to the SAGE evaluation.

General Findings 1996-97

1996-97 quantitativefindings show that first grade classrooms in SAGE schools scored higher on theCTBS Complete Battery, Terra Nova Level 10 than first grade students inComparison schools. In general, when adjusted for pre-test scores, SAGEstudents scored significantly higher on the post-test in the areas of reading,language arts, and mathematics as well as total score. This means that not allSAGE students outperformed all Comparison students. At the individuallevel of analysis, after controlling for pre-test score, SES, attendance, andrace, SAGE first grade students as a group scored statistically significantlyhigher than Comparison school students on the CTBS post-test in the areas oflanguage arts and mathematics as well as total score. At the class level ofanalysis, SAGE classrooms scored significantly higher in language arts,mathematics, and reading as well as total score after adjusting for individualpre-test results, SES, and attendance.

General Findings 1997-98

Analyses of the second year test data indicatedthat students in SAGE first grade classrooms achieved significantly higherscores than students in Comparison school classrooms in all tested areas:mathematics, reading, and language arts. The achievement advantage of studentsin SAGE first grade classrooms in 1996-97 was maintained in second grade in1997-98. The advantage, however, did not appear to have increased significantly.

Score Metrics 1998-99

A brief discussion of the metrics reported in the 1998-99 SAGE evaluation iswarranted. The SAGE report presents the findings using two metrics, scaledscores and normal curve equivalents. A scaled score provides a means forcomparison across subjects or groups on a specific task or trait. A scaledscore provides a common yardstick by which scores may be compared reasonably,subject to subject or group to group. The primary reason scaled scores are usedin the SAGE quantitative analysis is to anchor the scores from test level totest level (level 10, 11, etc.) so that year-to-year results can be compared.

When comparing the scores to those of other individuals (or groups) toobtain meaning, we make a norm-referenced interpretation. Here the use ofnormal curve equivalents is useful. A norm-referenced interpretation involvescomparing a person’s score with those of some relevant group of people. Thenormal curve equivalent scale ranges from 1 to 100 and thus provides acomparative index of the performance of an individual or group to the referencegroup. In this case, the reference group is the Terra Nova norm reference group(for norm referencing population data see (CTB/McGraw-Hill, 1991). Normal curveequivalents are generally not good indicators of longitudinal progress,however. With these scores, the group average could remain at, for example 50,across pre-test and post-test with the reader erroneously concluding that nogain was made. Actually, if the focus group, in this example, did not"gain" more than the reference group and the score would remainconstant although both groups scored higher.

Structure of 1998-99 Report

The descriptive analysesutilize both scale scores and normal curve equivalents. The inferentialanalyses (regressions and hierarchical linear models) utilize only scalescores. For the inferential tests, a significance level of .05 was used andsignificant results are denoted by an asterisk (*). SAGE versus comparison analysesare divided into three major sections: (1) First Grade Results, (2) SecondGrade Results, (3) Third Grade Results. The following are presented within eachof these sections: (1) descriptive statistics (pre-test and post-test), (2)ordinary least squares regressions, (3) analyses of the scores of AfricanAmerican students, and (4) hierarchical linear models.

The quantitative sectionincludes additional analyses looking at years of participation in SAGE,socio-economic status, and types of classroom. The quantitative section alsoincludes "within SAGE" analyses for first through third gradestudents. SAGE student achievement is examined in relation to teacherexperience, student participation, proximity to curriculum, and classorganization.

SAGE School/Classroom vs.Comparison School/Classroom Analyses

First grade Results 1998-99

Descriptive Statistics

Valid Test Scores. The number of first grade students for whom thevalid test scores are available is substantially less than the total number ofstudents. There are four main explanations for this. First, the evaluation teampresented schools with the option of allowing EEN and ESL students to take thetest, even though the test may be inappropriate for these students. Thesescores were invalidated based on a "Nonvalid/Missing Test Report,"developed by the evaluation team and completed for all first grade classes.Second, given withdrawals and enrollments during the school year, a number ofstudents had valid pre-test scores, but no post-test scores and vice versa.Third, some students took the reading and language arts components of the CTBS,or the mathematics component, but not both. Consequently, total scores areunavailable for these students. Finally, some of the students did not completethe pre-test, post-test, or both the pre- and post-tests. The number of validtest scores for the 1998-99 school year are presented in Table 8.

Table 8. Number of 1998-99 First grade Students with Valid TestScores

	Pre-Test		Post-Test
	SAGE	Comparison	SAGE	Comparison
Reading	1409	862	1335	874
Language Arts	1409	862	1335	874
Mathematics	1406	858	1328	869
Total	1396	855	1324	853

Table 9 gives the nationwide Scale Score Descriptive Statistics for thetest. Lowest and Highest obtainable scale scores as well as mean and standarddeviation are provided for each level of each sub-test.

Table 9. Scale Score Descriptive Statistics

	Lowest Obtainable Scale Score	Highest Obtainable Scale Score
Reading
Level 10 (first grade pre-test)	355	626
Level 11 (first grade post-test)	407	701
Level 12 (second grade)	423	722
Level 13 (third grade)	427	750
Mathematics
Level 10 (first grade pre-test)	290	629
Level 11 (first grade post-test)	324	680
Level 12 (second grade)	347	720
Level 13 (third grade)	385	740
Language Arts
Level 10 (first grade pre-test)	325	620
Level 11 (first grade post-test)	400	680
Level 12 (second grade)	424	706
Level 13 (third grade)	455	730

Pre-Test (Baseline) Results. Table 10 provides descriptive statisticsfrom the SAGE evaluation pre-test (baseline) results. Both Scale Scores andNormal Curve Equivalents are presented for both SAGE and Comparison students.Given the longitudinal nature of the SAGE evaluation, scale scores serve as theprimary measure of student achievement.

Table 10. SAGE and Comparison Population Descriptive Statistics onCTBS PRE-TEST Results for 1998-99 First grade Students

	SCALE SCORES				NORMAL CURVE EQUIVALENT
	SAGE		Comparison		SAGE		Comparison
	Mean	S.D.	Mean	S.D.	Mean	S.D	Mean	S.D.
Reading	533.85	37.85	539.54	41.21	44.52	20.77	48.37	21.70
Language Arts	533.59	43.54	536.84	46.40	45.67	21.89	47.67	22.57
Mathematics	492.93	41.19	501.22	40.97	43.40	19.31	47.45	19.50
Total	520.36	35.74	526.12	36.71	43.96	19.82	47.48	20.00

Difference of Means Test. The results from difference of means testsbetween SAGE and Comparison student scale scores from the Fall 1998 CTBS Level10 Pre-Test are reported in Table 11. Comparison school students scoredsignificantly higher than SAGE school students on all sub-tests and the totalscale at the .05 level. We reject the null hypothesis that there is nodifference between SAGE and Comparison school students on the pre-test. Itshould be noted that the composition of the Comparison schools changed when anumber of Comparison schools withdrew from the study, and new schools joinedthe study as Comparison schools. The relative higher performance of the newComparison schools was thus adjusted statistically.

Table 11. Differences of Means Test on CTBS FALL PRE-TEST for 1998-99First Grade Students

	SAGE	Comparison	Difference
Reading	533.85	539.54	-5.69*
Language Arts	533.59	536.84	-3.25*
Mathematics	492.93	501.22	-8.29*
Total	520.36	526.12	-5.76*
* Significant at .05 level

As noted above, studentpopulations varied in SAGE and Comparison schools due to withdrawals andwithin-year enrollments. The post-test results are based only on those first-grade students who remained in their schools for the entire 1998-99 schoolyear. CTBS allows for measurement of performance over time and thereforepre-test and post-test scores are comparable from a measurement position. TheCTBS Complete Battery, Terra Nova Level 10 was administered to first gradestudents in the fall and the CTBS Complete Battery, Terra Nova Level 11 wasadministered to first graders in the spring.

Table 12. SAGE and Comparison Population Descriptive Statistics onCTBS POST-TEST Results for 1998-99 First grade Students

	SCALE SCORES				NORMAL CURVE EQUIVALENT
	SAGE		Comparison		SAGE		Comparison
	Mean	S.D.	Mean	S.D.	Mean	S.D	Mean	S.D.
Reading	578.44	41.64	576.45	39.62	51.42	21.22	50.34	20.25
Language Arts	583.14	42.69	578.82	44.74	54.25	20.39	52.12	21.28
Mathematics	538.51	40.21	530.47	40.85	51.53	19.58	47.59	20.16
Total	566.88	35.71	562.34	36.54	52.82	19.96	50.27	20.44

Results of the difference of means test between SAGE and Comparison schoolson the CTBS Level 11 post-test are presented in Table 13. Statisticallysignificant differences are found in favor of SAGE students for the languagearts sub-test, mathematics sub-test, and for total scale scores on thepost-test.

Table 13. Differences of Means Test on CTBS FALL POST-TEST for1998-99 First grade Students

	SAGE	Comparison	Difference
Reading	578.44	576.45	1.99
Language Arts	583.14	578.82	4.32*
Mathematics	538.51	530.47	8.04*
Total	566.88	562.34	4.54*
*Significant at .05 level

The largest gain difference in SAGE student scores from pre-test topost-test, relative to Comparison school students, was on the mathematicssub-test (16.33) shown in Table 14. The smallest relative gain difference forSAGE students from pre-test to post-test relative to Comparison school studentswas on the reading sub-test (7.57).

Table 14. Change in Mean Score from PRE-TEST to POST-TEST for 1998-99First grade Students

	SAGE Gain	Comparison Gain	Gain Difference
Reading	49.55	41.98	7.57*
Language Arts	44.59	36.91	7.68*
Mathematics	45.58	29.25	16.33*
Total	46.52	36.22	10.30*
* Significant at .05 level

Regression Analysis

Regression Models. The effect of the SAGE program on studentachievement, controlling for other factors, was tested through a series ofordinary least squares regression models for each sub-test and for total scalescores. Control variables were entered into the models in blocks, with theSAGE/Comparison student variable entered into the models last.

The first block of controlvariables included student scores on the pre-test and school attendance,measured as number of days absent, as reported by teachers in Spring 1999. Eligibilityfor subsidized lunch, as an indicator of family income, is also included in thefirst block of control variables. This variable is coded 0 if student isineligible, 1 if student is eligible for reduced price lunch, and 2 if thestudent is eligible for free lunch (this variable is assumed to be intervallevel). The second block of control variables included dummy variables forrace/ethnicity, coded 1 if a student was of a certain race/ethnicity, and 0 ifnot. Dummy variables were included for African Americans and whites. A residualcategory, "other", is included in the constant term in the regressionequations. In the final block, a dummy variable for SAGE or Comparison schoolstudent was entered on the third block. This variable is coded 0 if a studentis from a Comparison school and 1 if a student is from a SAGE school.

Regression Results.Results of the regression analyses are presented in Table 15. For all analyses,SAGE emerges as a significant predictor of student achievement on the post-test,while controlling for pre-test scores, family income, school attendance, andrace/ethnicity. The magnitude of the effect of SAGE on student achievement, asdenoted by the "b" coefficient, varies depending on the CTBSsub-test.

The largest effects of SAGEare found on the math sub-test (12.454), while the smallest effects of SAGE arefound on the reading sub-test (6.010). When all cases are analyzed thegoodness-of-fit of the models (as denoted by the adjusted R square statistic),ranges from .257 (reading sub-scale score) to .514 (total scale score). Thismeans that when predicting the reading score and total score, the variablesincluded in the model explain 26% and 51% of the variance respectively. Most ofthe variance in the post-test scores is, of course, explained by the pre-testscores.

Explained Variance in Achievement Scores. Attendance (as representedby "days absent") emerges as a consistent and statisticallysignificant predictor of performance on all sub-tests and total scale score.Family income as denoted by the "Lunch Eligibility" variable alsoemerges as a consistent and statistically significant predictor of performanceon all sub-tests and total scale score. "Race" shows some relativelylarge effects (as denoted by the b coefficients), but the effects are highlyvariable and are only sometimes statistically significant (race is discussedfurther below). Membership in SAGE schools has a consistently positive,statistically significant effect on achievement on the CTBS.

Table 15. Scale Scores Regression – First Grade Block Three BetaCoefficients

	Reading	Language Arts	Math	Total
Pre-Test Score	.445*	.530*	.601*	.699*
Days Absent	-.395*	-.543*	-.447*	-2.278*
Lunch Eligibility	-6.620*	-3.706*	-1.644*	-.348*
African American	-.168	-2.434	.698	-.424
White	4.076*	4.572*	3.581*	1.527
SAGE	6.010*	7.021*	12.454*	8.490*
Constant	349.224*	306.648*	244.849*	206.638*
Adjusted R Squared	.257	.358	.409	.514
Standard Error	35.53	34.81	31.53	25.29
*Significant at .05 level

African American Students

Among minority students in SAGE and Comparison schools, African Americansclearly comprise the largest group of valid test scores – roughly 22 percent ofSAGE students and 20 percent of Comparison school students. In the analyses tofollow, African American first grade students are first compared across SAGEand Comparison schools on CTBS sub-test and total scale scores. Second, AfricanAmerican first grade students are compared to white students across SAGE andComparison schools on CTBS total scale scores.

SAGE vs. Comparison.Table 16 provides comparisons of means on the CTBS post-test and change scoresfrom pre-test to post-test. On the post-test, African American SAGE studentsscored higher than African American Comparison school students on every sub-testand on total scale score. There are statistically significant differencesbetween SAGE and Comparison students in the language arts, mathematics, andtotal sub-tests. In addition, the differences between SAGE and Comparisonstudents on mean change scores from pre-test to post-test scores are allstatistically significant. African American SAGE students scored lower on theCTBS pre-test than African American Comparison school students, but madesignificantly larger gains than Comparison school students from pre- topost-test and surpassed African American Comparison school students on thepost-test.

Table 16. African American Post-Test and Change Scores, by SAGE orComparison School for 1998-99 First Grade Students

SCALE SCORE	SAGE	COMPARISON	DIFFERENCE
Language Arts
Mean Post-Test	573.58	560.51	13.07*
Mean Change Pre to Post	48.02	37.82	10.20*
Reading
Mean Post-Test	570.75	564.01	6.74
Mean Change Pre to Post	45.21	30.94	14.27*
Mathematics
Mean Post-Test	526.81	515.57	11.24*
Mean Change Pre to Post	48.90	32.25	16.65*
Total
Mean Post Test	557.70	546.98	10.72*
Mean Change, Pre to Post	47.72	34.17	13.55*
*significant at .05 level

African American and White Achievement. African American first gradestudents scored significantly lower than white students on the CTBS pre-testtotal scale score, as shown in Table 17. This result holds for both SAGE andComparison schools. When all cases are analyzed, African- American SAGEstudents achieved greater gains on the CTBS total scale score than white SAGEstudents from pre- to post-test, closing the achievement gap (though the gapremains statistically significant). In contrast, African Americans inComparison schools showed lesser gains and the achievement gap widened in theComparison schools.

Table 17. African American and White Achievement in SAGE andComparison Schools

on Total Scale Scores for 1998-99 First Grade Students

	PRE-TEST	POST-TEST	CHANGE
SAGE SCHOOLS
African American	509.24	557.70	48.46
White	533.23	577.96	44.73
Difference	-23.99*	-20.26*	3.73*
COMPARISON SCHOOLS
African American	512.36	546.98	34.62
White	537.88	574.93	37.05
Difference	-25.52*	-27.95*	-2.43
*significant at .05 level

Hierarchical Linear Modeling

Many social science research analyses involve hierarchical data structures.Hierarchical data structures are those in which individual units are nestedwithin larger units, the latter being the unit of interest. The SAGE data are aprime example: students are nested within classrooms, and it is the classroomeffect that is of particular interest to the SAGE project. Hierarchical datastructures pose special analytical challenges in that data analysis at theindividual level may result in a biasedimpression of the effect of the nesting unit (in the SAGE case, the classroom).An analytical approach known as "hierarchical linear modeling" or HLM(Bryk & Raudenbush, 1992) was specifically designed to accommodate thesetypes of data structures. HLM was used with the SAGE data to provide analternative and less biased account of the effects of SAGE experience on testscores. In these models, variables associated with individual students arereferred to as level-1 variables and those associated with the classrooms arereferred to as level-2 variables.

 Analyses were conducted for each of therelevant criterion post-test scores: reading, mathematics, language arts, andtotal. For all analyses, the level-1 variables were pre-test achievementscores, attendance, and SES measured as eligibility for subsidized lunch. Thepost-test scores were adjusted for these three variables at the individuallevel, therefore the effects may be thought of as being statisticallyindependent of the effects of these variables. Three different level-2 modelsare reported here: Model A specifies "class size" as the only level-2variable, Model B includes both "class size" and "SAGE"variables, and Model C includes both "class SES" and "class size"variables. Class SES was computed as the student SES average within each classand was therefore a measure of class poverty.

 Table 18 provides a summary of the effects ofeach of the level-1 and level-2 variables for each of these analyses. Level-1effects can be interpreted as the weighted average of the within-classroomeffects of the level-1 variables. Level-2 effects can be interpreted as theclassroom effects of the level-2 variables. The level-1 results indicate thatwithin classrooms, lower individual SES is related to lower post-test scoresand higher pre-test scores are related to higher post-test scores. Thecoefficients associated with the level-2 variables can be thought of asclassroom effects.

Table 18. HLM Results for 1998-99 First Grade Students: First GradePre-Test as Control

Source	Total	Reading	Language Arts	Math
Level 1^**
Pre-Test	0.806	0.592	0.575	0.689
SES	-1.063	-3.804	-3.253	-1.478
Attendance	-0.115	-0.208	-0.701	-0.184
Level 2
A. Class Size	-0.855*	-0.539*	-0.734*	-1.265*
B. Class Size	-1.945*	-1.634*	-2.041*	-2.197*
SAGE	-16.116*	-16.427*	-19.304*	-13.850*
C. Class SES	-12.672*	-14.342*	-13.490*	-10.585*
Class Size	-0.795*	-0.461	-0.653*	-1.197*

Model A. Class Size. These models examined the effect of classsize on the adjusted criterion score. Class size equals the number of studentsdivided by the number of teachers. Depending on the test, an increase in classsize of one person can be expected to produce a .54 to 1.27 scale score pointloss in average post-test performance. HLM results indicate the classroom leveleffect of the SAGE program and are therefore more powerful than the regressionresults in understanding of the program.

Model B. Class Size, SAGE. Combining class size and SAGEparticipation in a single analysis isolates the effects that SAGE might havebeyond those produced by lower class size. However, due to the redundancy ofthese two variables in this case, this effect is not apparent and can beassumed to be zero.

It is instructive to view the results of model B in the context of model Aresults. Due to the multicollinearity of the SAGE and class size variable(i.e., they are very highly correlated), in model B the class size effectincreases from that in model A. Concomitantly, the SAGE effect is negative tocompensate for this effect. This is a statistical phenomenon, not aninterpretable result.

Model C. ClassSES, Class Size. These models examined the effect of class size on theadjusted criterion score after the classrooms were SES adjusted, viewed as theeffect of class size once the effects of the classroom SES are removed. Sincesocioeconomic status is known to have an influence on academic test scores, areplacement for this variable was used as both a level-1 and level-2 predictor.The level-2 variable was the average SES for the class and estimates the effectof the overall class SES level beyond that associated with the individual,which is accounted for in the level-1 model. This model combines class SES andclass size. The results indicate that class SES has a significant effect on theclass average post-test performance. The effect of a 1 point class average gainin SES (recall coding: ineligible = 0, reduced = 1, eligibility for free lunch= 2) equates to between a 11 point and 14 point loss on the average post-testscore, depending on the sub-score. SES was measured on a three-point family incomescale, thus a one point difference on average would be quite pronounced. Classsize still has a significant (although smaller) effect on the post-test scoresonce SES has been accounted for in all areas except for reading.

Second Grade Results 1998-99

Descriptive Statistics

Valid Test Scores. Analyses were conducted to assess the impact ofSAGE on the 1998-99 second grade CTBS Complete Battery, Terra Nova Level 12post-test results. There were 1446persisting second grade students (i.e., students present in both the 1997-98SAGE and comparison first grade classrooms and the 1998-99 SAGE and comparisonsecond grade classrooms). Second grade post-test results are compared to thefirst grade pre-test, as well as first grade post-test. Therefore, only thosestudents who took both the first grade pre-test and post-test, as well as thesecond grade post-test, were used in the 1998-99 second grade analysis. Aswould be expected, the number of second grade students having all three validtest scores was substantially less than the total number of students. Thenumber of valid test scores for the Fall 1997 first grade pre-test, the Spring1998 first grade post-test, and the Spring 1999 second grade post-test arepresented in Table 19.

Table 19. Number of 1998-99 Second Grade Students with Valid TestScores

	Fall 1997 First Grade Pre-test		Spring 1998 First Grade Post-test		Second GradeTest
	SAGE	Comparison	SAGE	Comparison	SAGE	Comparison
Reading	1386	865	1247	797	1312	893
Language Arts	1386	864	1248	797	1312	893
Mathematics	1385	859	1264	793	1318	881
Total	1370	846	1241	782	1300	871

Pre-Test (Baseline)Results. Both the first grade pre-test and the first grade post-test servedas baseline measures. Tables 20, 21 and 22 provide descriptive statistics onthe scale scores from the first- grade pre- and post-test.

Table 20. Descriptive Statistics on CTBS First Grade Pre-Test andPost-Test and Second Grade Test (SAGE)

	FIRST GRADE PRE-TEST				FIRST GRADE POST-TEST
	SCALE SCORES		NORMAL CURVE EQUIVALENT		SCALE SCORES		NORMAL CURVE EQUIVALENT
	Mean	S.D.	Mean	S.D.	Mean	S.D.	Mean	S.D.
Reading	533.39	36.41	44.05	20.32	581.26	40.92	52.96	20.60
Language Arts	530.55	43.75	44.11	21.47	587.12	44.51	56.27	20.93
Mathematics	492.43	42.63	43.28	19.70	539.29	40.06	52.00	19.34
Total	519.12	35.35	43.29	19.58	569.50	36.22	54.36	20.06

Table 21. Descriptive Statistics on CTBS First Grade Pre-Test andPost-Test and Second Grade Test (Comparison)

	FIRST GRADE PRE-TEST				FIRST GRADE POST-TEST
	SCALE SCORES		NORMAL CURVE EQUIVALENT		SCALE SCORES		NORMAL CURVE EQUIVALENT
	Mean	S.D.	Mean	S.D.	Mean	S.D.	Mean	S.D.
Reading	535.02	36.20	45.19	19.11	572.25	44.58	48.48	21.53
Language Arts	528.95	43.34	43.24	21.11	575.85	46.01	50.94	21.24
Mathematics	493.00	38.34	43.35	18.13	526.80	41.61	46.00	19.44
Total	519.48	33.33	43.45	18.33	558.55	37.80	48.42	20.57

Table 22. SAGE and Comparison Descriptive Statistics on CTBS Resultsfor 1998-99 Second Grade Test

	SCALE SCORES				NORMAL CURVE EQUIVALENT
	SAGE		Comparison		SAGE		Comparison
	Mean	S.D.	Mean	S.D.	Mean	S.D	Mean	S.D.
Reading	608.96	34.63	603.67	37.92	50.36	17.94	47.78	18.61
Language Arts	611.55	39.39	600.49	38.47	50.38	19.85	44.82	19.12
Mathematics	571.68	39.39	555.61	40.29	49.62	19.92	41.41	19.99
Total	597.72	33.14	586.78	33.66	50.63	18.99	44.35	18.99

Difference of Means Test.The difference of means between SAGE and Comparison students on the secondgrade test can be seen in Table 23. SAGE students scored significantly higherin every sub test and the total score.

Table 23. Differences of Means Test 1998-99 Second Grade Test

	SAGE	Comparison	Difference
Reading	608.96	603.67	5.29*
Language Arts	611.55	600.49	11.06*
Mathematics	571.68	555.61	16.07*
Total	597.72	586.78	10.94*
*significant at .05 level

Results of the difference of means test between SAGE and Comparison schoolson

the second grade post-test can be seen in Table 24. Table 24 shows that whenthe first grade pre-test is used as the baseline score, significant results arefound on the reading sub-scale, language arts sub-scale, mathematics sub-scale,and total score. When the first grade post-test is used as the baseline score,there are significant results for the language arts sub-scale, math sub-scale,and total score, but no significant results are found for the readingsub-scale. This suggests that the statistically significant positive effects ofSAGE occurred in the first grade. These positive effects were maintained andsignificantly increased in second grade in all areas except reading.

Table 24. SAGE and Comparison Gain

	From First grade Pre-Test to Second Grade			From First grade Post-Test to Second Grade
SCALE SCORE	SAGE Gain	Comparison Gain	Gain Difference	SAGE Gain	Comparison Gain	Gain Difference
Reading	71.45	62.73	8.72*	26.90	26.32	.59
Language Arts	74.84	61.32	13.53*	23.88	16.98	6.90*
Mathematics	74.66	55.94	18.71*	32.60	23.45	9.15*
Total	73.55	60.42	13.13*	27.87	22.20	5.68*
*significant at .05 level

Regression Analysis

Regression Models.The effect of the SAGE program on student achievement for second graderswas also tested through a series of ordinary least squares regression modelsfor each sub-test and total scale score. Control variables were again enteredinto the models in blocks, with the SAGE/Comparison student variable enteredinto the models last. In addition, two different regressions were done for eachsub-test and total scale score. The first regression used the first- gradepre-test as a predictor variable and the second regression used the first gradepost-test as a predictor variable.

The first block of controlvariables included student scores on the first grade pre-test or post-test,attendance, and eligibility for subsidized lunch as an indicator of familyincome. As with the first graders (discussed earlier), the second block ofcontrol variables included dummy variables for race/ethnicity. Finally, a dummyvariable for SAGE or Comparison school student was entered on the third block.As with the first graders, this variable is coded 0 if a student is from aComparison school and 1 if a student is from a SAGE school.

Regression Results.Results of the regression analyses are presented in Tables 25 and 26. Wheneither the first grade pre-test or the first grade post-test is used as thepredictor variable, membership in SAGE emerges as a significant predictor ofstudent achievement on the total score and for all sub-tests except reading.The magnitude of the effect of SAGE on student achievement, as denoted by the"b" coefficient, varies depending on the CTBS sub-test.

The largest effects of SAGEare found when the first grade mathematics pre-test is used to predict thesecond grade test (18.650). When all cases are analyzed, the goodness-of-fit ofthe models (as denoted by the adjusted R square statistic), ranges from .26 to.56. Most of the variance, as was the case with the first graders, is explainedby the baseline scores (either the first grade pre-test or the first gradepost-test). "Family Income" , "Attendance", and"Race" show some relatively large effects (as denoted by the bcoefficients) and these effects are usually statistically significant. Mostimportantly, membership in SAGE schools has a consistently statisticallysignificant positive effect on all sub-tests and the total score. A negativerelationship suggests that if these conditions are present, the test score willbe lower.

Table 25. Scale Scores Regression – Second Grade Block Three BetaCoefficients: First Grade Pre-Test as Control

	Reading	Language Arts	Math	Total
Pre-Test Score	.375*	.448*	.557*	.626*
Days Absent	-.304*	-.276	-.368*	-.182*
Lunch Eligibility	-5.548*	-4.183*	-6.259*	-4.032*
African American	-2.807	-7.883*	-10.002*	-6.512*
White	8.624*	3.318	2.803	3.131*
SAGE	7.804*	13.917*	18.650*	13.278*
Constant	419.911*	392.475*	324.864*	289.110*
Adjusted R Squared	.256	.335	.445	.513
Standard Error	30.93	31.77	30.82	23.53
*Significant at .05 level

Table 26. Scale Scores Regression – Second Grade Block Three BetaCoefficients: First Grade Post Test as Control

	Reading	Language Arts	Math	Total
Pre-Test Score	.394*	.445*	.559*	.602*
Days Absent	-.279*	-.205	-.102	.010
Lunch Eligibility	-6.204*	-4.508*	-4.884*	-4.222*
African American	-4.583*	-6.256*	-9.171*	-5.976*
White	6.919*	3.536	5.116*	3.252*
SAGE	4.135*	10.803*	12.259*	8.607*
Constant	390.807*	368.313*	289.386*	269.397*
Adjusted R Squared	.347	.360	.475	.561
Standard Error	28.43	31.03	29.29	22.13
*Significant at .05 level

African American Students

As in first grade, African American second grade students comprise thelargest minority subgroup of valid test scores – roughly 25% of all SAGEstudents and 22% of all Comparison students. In the analyses to follow, AfricanAmerican students are first compared across SAGE and Comparison schools on theCTBS sub-tests and total scale score. Second, African American students arecompared to white students across SAGE and Comparison schools on the CTBS TotalScale Score.

SAGE vs. Comparison.Table 27 provides comparisons of means on the CTBS second- grade test, as wellas change scores from the first grade pre-test to the second grade test andfrom the first grade post-test to the second grade test. On the second gradetest, African American SAGE students scored higher than Comparison schoolstudents on every sub-test and on the total scale score. The differencesbetween African American SAGE students and African American Comparison studentson the second grade test scores are statistically significant. When using thefirst grade pre-test as the baseline score, statistically significant changescores are found on all scores. SAGE African American student performance isstill ahead of slightly larger non-significant gain. Using the first gradepost-test as the baseline score shows statistically significant differencesbetween SAGE and Comparison schools on the mathematics sub-test only.Comparison school students made a larger gain on the reading sub-test, but thegain is not statistically significant.

Table 27. African American Post-Test and Second Grade Change Scores,by SAGE or Comparison

SCORE	SAGE	COMPARISON	DIFFERENCE
Language Arts
Mean Second grade Score	595.85	584.33	11.52*
Mean Change From First grade Pre-Test to Second Grad	72.65	56.27	16.38*
Mean Change From First grade Post-Test to Second Grade	21.09	18.75	2.34
Reading
Mean Second grade Score	595.94	587.81	8.13*
Mean Change From First grade Pre-Test to Second Grade	70.97	57.73	13.24*
Mean Change From First grade Post-Test to Second Grade	20.88	27.19	-6.31
Mathematics
Mean Second grade Score	551.52	534.86	13.66*
Mean Change From First grade Pre-Test to SecondGrade	74.24	47.41	26.83*
Mean Change From First grade Post-Test to SecondGrade	28.79	20.41	8.38*
Total
Mean Second grade Score	581.53	569.01	12.52*
Mean Change From First grade Pre-Test to Second Grade	72.12	54.03	18.09*
Mean Change From First grade Post-Test to Second Grade	24.04	22.05	1.99
*significant at .05 level

African American andWhite Achievement. African American students scored lower than whitestudents on the first grade pre-test total scale score, as shown in Table 28.This result is statistically significant for both SAGE and Comparison schools,though the gap between African Americans and whites is larger in the SAGEschools. The change from the first grade post-test to the second grade testshows that the SAGE African Americans kept pace with white students but did notfurther close the achievement gap in second grade.

Table 28. African American versus White Achievement on Total Scale

	First Grade Pre-Test	First Grade Post-Test	Second Grade	Change From Pre-Test to Second	Change From Post-Test to Second
SAGE
African American	502.90	556.83	581.53	72.12	24.04
White	531.27	580.29	608.67	73.51	27.69
Difference	-28.37*	-23.46*	-27.14*	-1.39	-3.65
Comparison
African American	509.98	541.87	569.01	54.03	22.05
White	528.60	570.18	597.68	64.46	23.57
Difference	-18.62*	-28.31*	-28.67*	-10.43*	-1.52
*significant at .05 level

Hierarchical Linear Modeling

Hierarchical linear modelswere used to evaluate the second grade achievement results using the sameseries of models used to assess first grade results. Two sets of analyses weredone. The first, shown in Table 29, used first grade pre-test as the initialachievement level of the students. The second, shown in Table 30, used firstgrade post-test as the initial achievement level of the students.

HLMresults. Tables 30 and 31 provide asummary of the effects of each of the level-1 and level-2 variables for each ofthese analyses. Level-1 effects can be interpreted as the weighted average ofthe within-classroom effects of the level-1 variables. Level-2 effects can beinterpreted as the classroom effects of the level-2 variables.

Table 29. HLM Results for 1998-99 Second grade Students – FirstGrade Pre-Test as Initial Achievement

Source	Total	Reading	Language Arts	Mathematics
Level 1
Pre-Test	0.7190	0.497	0.496	0.568
SES	-3.859	-9.331	-8.091	-6.094
Attendance	-0.014	0.371	-0.322	-0.350
Level 2
Class Size	-1.021*	-0.758*	-0.807*	-1.507*
B. Class Size	-1.215*	-1.233*	-0.551	-1.799*
SAGE	-2.995	-7.399	3.954	-4.634
C. Class SES	-16.464*	-16.966*	-16.267*	-20.884*
Class Size	-0.948*	-0.654*	-0.746*	-1.398*
*significant at .05 level

Table 30. HLM Results for 1997-98 Second grade Students – FirstGrade Post Test as Initial Achievement

Source	Total	Reading	Language Arts	Mathematics
Level 1
Post-Test	0.657	0.510	0.453	0.636
SES	-5.201	-5.726	-3.913	-2.285
Attendance	0.089	-0.182	0.221	0.304
Level 2
A. Class Size	-1.048*	-0.748*	-0.799*	-1.607*
B. Class Size	-1.282*	-1.364*	-0.532	-1.824*
SAGE	-3.611	-9.634*	4.193	-3.391
C. Class SES	-16.805*	-17.047*	-16.932*	-22.078*
Class Size	-0.985*	-0.658*	-0.761*	-1.486*
*significant at .05 level

Model A. Class Size. Depending on the test, an increase inclass size of one person can be expected to produce a .75 to 1.6 scale scoreloss in average post-test performance. In other words, the more students in aclass, the lower the average class achievement score. The results for allscores show this effect to be significant. Because the HLM adjusts fordifferences in class size, it is a better indication of what is actuallyoccurring at the classroom level. It is noteworthy that the results of Tables29 and 30 for the Model A are similar, indicating that the second grade classsize has little effect on the two year gain (first grade and second grade)versus the one year gain (second grade only). The positive effect of reducedclass size is maintained from first grade post-test to second grade test, butnot increased. It may appear that the regression and HLM results arecontradictory. The HLM results are, however, consistent with the regressionresults because individuals in SAGE classrooms perform better than individualsin Comparison classrooms, and SAGE classrooms have fewer students. As wasdiscussed relative to first grade findings, HLM results indicate the classroomlevel effect of the SAGE program and are therefore more powerful than theregression results in understanding of the program.

Model B. Class Size, SAGE. Combining class size andSAGE participation in a single analysis isolates the effects that SAGE mighthave beyond those produced by lower class size. However, due to the redundancyof these two variables in this case, this effect is not apparent and can beassumed to be zero.

It is instructive to view the results of model B in the context of model Aresults. Due to the multicollinearity of the SAGE and class size variable(i.e., they are very highly correlated), in model B the class size effectincreases from that in model A. Concomitantly, the SAGE effect is negative tocompensate for this effect. This is a statistical phenomenon, not aninterpretable result.

Model C. ClassSES, SAGE. As in the first grade analysis, this model combines class SESand class size. The results indicate that class SES has a significant effect onthe class average post-test performance. Class size still has a significanteffect on the post-test scores once SES has been accounted for, although theeffects are somewhat attenuated.

Third Grade Results 1998-99

Descriptive Statistics

Valid Test Scores. Analyses were conducted to assess the impact ofSAGE on the 1998-99 third grade CTBS Complete Battery, Terra Nova Level 13post-test results. There were 1702persisting students in third grade (i.e., students present in both the 1996-97SAGE and comparison first- grade classrooms and in the 1997-98 SAGE andcomparison second grade classrooms), while there were 482 new third gradestudents (students who were not in the program last year). However, third grade test results are compared tothe first grade pre-test, first grade post-test and the second grade test.Therefore, only those students who took both the first grade pre-test andpost-test, the second grade test and the third grade test were used in the1998-99 third grade analysis. As would be expected, the number of third gradestudents having all four valid test scores was substantially less than thetotal number of students. The number of valid test scores for the Fall 1996first grade pre-test, the Spring 1997 first grade post-test, the Spring 1998 secondgrade test and Spring 1999 third grade test are presented in Table 31.

Table 31. Number of 1998-99 Third Grade Students with Valid TestScores

	Fall 1997 First Grade Pre-test		Spring 1998 First Grade Post-test		Second Grade Test		Third Grade Test
	SAGE	Comp	SAGE	Comp	SAGE	Comp	SAGE	Comp
Reading	981	475	1020	488	1202	590	1322	895
Language Arts	981	475	1020	488	1201	591	1322	895
Mathematics	971	469	1014	482	1206	586	1324	893
Total	959	460	1009	478	1194	576	1313	877

First Grade and SecondGrade (Baseline) Results. First grade pre-test, first grade post-test, andsecond grade test scores served as a baseline. Table 32 provides descriptivestatistics on the scale scores from the first grade pre-test, first gradepost-test, and second grade test. Table 33 provides descriptive statistics forthe third grade test.

Table 32. Descriptive Statistics Means (and Standard Deviations) onCTBS First Grade Pre-Test and Post-Test and Second Grade Test (SAGE andComparison)

	First Grade Pre-Test		First Grade Post-Test		Second Grade Test
	SAGE	Comp	SAGE	Comp	SAGE	Comp
Reading	535.91	533.55	587.21	581.08	608.25	601.38
	(35.29)	(39.75)	(33.49)	(37.12)	(35.43)	(34.83)
Language Arts	534.43	529.53	586.88	577.81	610.80	598.42
	(41.07)	(42.48)	(35.11)	(40.21)	(40.42)	(39.41)
Math	495.72	491.88	550.60	540.23	571.70	560.19
	(37.49)	(37.98)	(40.97)	(41.62)	(41.91)	(38.03)
Total	522.18	518.63	575.05	566.28	597.10	587.27
	(32.55)	(33.65)	(30.35)	(32.83)	(33.75)	(32.55)

Table 33. SAGE and Comparison 1998-99 Descriptive Statistics CTBSThird Grade Test

	SCALE SCORES				NORMAL CURVE EQUIVALENT
	SAGE		Comparison		SAGE		Comparison
	Mean	S.D.	Mean	S.D.	Mean	S.D	Mean	S.D.
Reading	633.26	39.27	626.95	37.89	51.98	19.08	48.71	18.74
Language Arts	629.03	34.62	622.47	35.02	50.34	17.77	46.80	18.11
Mathematics	607.53	36.69	595.29	38.23	50.28	18.17	44.38	18.40
Total	623.42	33.33	615.29	33.56	51.19	18.28	46.73	18.41

Difference of Means Test.Results from the difference of means tests between SAGE and comparison studentscores from the third grade test are reported in Table 34. As Table 34 shows,SAGE students scored significantly higher than Comparison students in all areasof the test.

Table 34. Differences of Means Test 1998-99 Third Grade Test

	SAGE	Comparison	Difference
Reading	633.26	626.95	6.31*
Language Arts	629.03	622.47	6.56*
Mathematics	607.53	595.29	12.24*
Total	623.42	615.29	8.13*
*significant at .05 level

As noted previously,student populations varied in SAGE and Comparison schools due to withdrawalsand within-year enrollments. The third grade results are based only on thosethird graders who were also in SAGE classrooms the entire 1996-97 first grade and1997-98 second grade school years.

The results from thedifference of means tests between SAGE and Comparison student scale scores fromthe Fall 1996 first grade pre-test, Spring 1997 first- grade post-test, andSpring 1998 second grade test and Spring 1999 third grade test are reported inTable 35. These results reflect comparisons on an individual student level, thedifferences in scores between SAGE and Comparison students. When thefirst grade pre-test is used as the baseline score, SAGE students madesignificantly higher gains on the reading sub-scale, mathematics sub-scale, andtotal score compared to Comparison students.

When the first grade post-test is used as the baseline score, the samesignificant results are found on the reading sub-scale, the mathematicssub-scale, and the total score. These results suggest that the positive effectsof SAGE were maintained and significantly increased in second grade for thereading sub-scale, mathematics sub-scale, and total score.

When the second grade test scores are used for a baseline, the samestatistically significant gains can be seen in reading, mathematics and thetotal score. As with the previous testing cycles, it appears that positive SAGEeffects were maintained and again significantly increased in third grade.

Thesmallest relative gain for SAGE students from first grade pre-test to the thirdgrade test was on the language arts sub-scale; this gain was not statisticallysignificant. Comparison schools had larger comparative gains on the languagearts sub-test from second grade to third grade; the gains were notstatistically significant.

Table 35. SAGE and Comparison Gain

	From First Grade Pre-Test to Third Grade Test			From First Grade Post-Test to Third Grade Test			From Second Grade Test to Third Grade Test
SCALE SCORE	SAGE Gain	Comp Gain	Gain Diff.	SAGE Gain	Comp Gain	Gain Diff.	SAGE Gain	Comp Gain	Gain Diff.
Reading	100.00	91.96	8.04*	49.57	44.09	5.49*	26.25	21.85	4.40*
Language Arts	95.75	91.34	4.41	44.33	43.81	.52	19.19	22.31	-3.11
Mathematics	113.55	98.26	15.29*	59.67	51.29	8.38*	36.40	29.42	6.98*
Total	102.91	94.45	8.46*	51.04	47.06	3.98*	27.33	24.33	3.00*
*significant at .05 level

Regression Analysis

Regression Models.The effect of the SAGE program on student achievement for third graders wasalso tested through a series of ordinary least squares regression models foreach sub-test and total scale score. Control variables were again entered intothe models in blocks, with the SAGE/Comparison student variable entered intothe models last. In addition, three different regressions were done for eachsub-test and total scale score. The first regression used the first- gradepre-test as a predictor variable (Table 36), the second regression used thefirst grade post-test as a predictor variable (Table 37), and the thirdregression used the second grade test as a predictor variable (Table 38).

The first block of controlvariables included student scores on the first grade pre-test or post-test,attendance, and eligibility for subsidized lunch as an indicator of familyincome. As with the first and second graders (discussed earlier), the secondblock of control variables included dummy variables for race/ethnicity.Finally, a dummy variable for SAGE or Comparison school student was entered onthe third block. As with the first and second graders, this variable is coded 0if a student is from a Comparison school and 1 if a student is from a SAGEschool.

Regression Results.Results of the regression analyses are presented in Tables 37-39. When eitherthe first grade pre-test, the first grade post-test, or second grade test isused as the predictor variable, membership in SAGE emerges as a significantpredictor of student achievement on the total score and for all sub-tests. Themagnitude of the effect of SAGE on student achievement, as denoted by the"b" coefficient, varies depending on the CTBS sub-test.

The largest effects of SAGEare found when the first grade mathematics pre-test is used to predict thesecond grade test (16.105). When all cases are analyzed, the goodness-of-fit ofthe models (as denoted by the adjusted R square statistic), ranges from .20 to.50. Most of the variance, as was the case with the first and second graders,is explained by the baseline scores (either the first grade pre-test, firstgrade post-test, or second grade test). "Family Income" ,"Attendance", "Race" show some relatively large effects (asdenoted by the b coefficients) and these effects are usually statisticallysignificant. This suggests that a student with a high absentee rate or from alow socio-economic status will have lower test scores. Most importantly,membership in SAGE schools has a consistently statistically significantpositive effect on all sub-tests and the total score.

Table 36. Scale Scores Regression – Third Grade Block Three BetaCoefficients: First Grade Pre-Test as Control

	Reading	Language Arts	Math	Total
Pre-Test Score	.395*	.351*	.505*	.578*
Days Absent	-.175	-.273	-.471*	-.255*
Lunch Eligibility	-5.042*	-3.922*	-3.183*	-3.152
African American	1.406	1.017	-5.889*	.106
White	6.375*	5.256*	-.203	1.441
SAGE	8.088*	5.716*	16.105*	9.114*
Constant	428.446*	447.082*	367.406*	328.689*
Adjusted R Squared	.224	.244	.367	.401
Standard Error	32.83	29.61	28.86	25.06
*Significant at .05 level

Table 37. Scale Scores Regression – Third Grade Block Three BetaCoefficients: First Grade Post-Test as Control

	Reading	Language Arts	Math	Total
Pre-Test Score	.395*	.324*	.385*	.545*
Days Absent	-.208	-.355*	-.507*	-.306*
Lunch Eligibility	-4.702*	-4.732*	-4.274*	-3.656*
African American	1.073	1.459	-6.537	.753
White	9.598*	8.109*	2.255	5.266*
SAGE	7.719*	5.287*.	13.353*	7.249*
Constant	406.288*	444.875*	406.488*	315.941*
Adjusted R Squared	.220	.200	.333	.369
Standard Error	32.58	30.11	29.44	25.33
*Significant at .05 level

Table 38. Scale Scores Regression – Third Grade Block Three BetaCoefficients: Second Grade Test as Control

	Reading	Language Arts	Math	Total
Pre-Test Score	.619*	.431*	.487*	.655*
Days Absent	-.03	-.286*	-.417*	-.206
Lunch Eligibility	-2.654*	-3.294*	-2.387*	-2.001*
African American	-.545	2.057	-2.001	2.152
White	6.428*	7.667*	4.531*	4.876*
SAGE	7.473*	4.735*	13.058*	7.107*
Constant	259.128*	368.882*	333.960*	234.681*
Adjusted R Squared	.392	.315	.399	.499
Standard Error	28.74	28.31	28.16	22.79
*Significant at .05 level

African American Students

As in the first and second grade classrooms, African American third gradestudents comprise the largest subgroup of valid test scores – roughly 22% ofall SAGE students and 23% of all Comparison students. In the analyses tofollow, African American students are first compared across SAGE and Comparisonschools on the CTBS sub-tests and total scale score. Second, African Americanstudents are compared to white students across SAGE and Comparison schools onthe CTBS Total Scale Score.

SAGE vs. Comparison.Table 39 provides comparisons of means on the CTBS third grade test, as well aschange scores from the first grade pre-test to the second grade test and fromthe first grade post-test to the second grade test, and from the second gradetest to the third grade test. On the third grade test, African American SAGEstudents scored higher than African American Comparison school students onevery sub-test and on the total scale score, and the differences betweenAfrican American SAGE students and African American Comparison students on thethird grade test scores are statistically significant.

When using the first grade pre-test as the baseline score, statisticallysignificant change scores are found on all scores. Using the first gradepost-test as the baseline score shows statistically significant differencesbetween African American SAGE students and African American Comparison studentson the math sub-test and total score. When using the second grade test as thebaseline score, the reading and mathematics sub-tests and total scores showstatistically significant change scores.

In general, African American SAGE students outperformed African Americancomparison students, although the difference was not significant in languagearts from the first grade post-test and the second grade test to the thirdgrade test.

Table 39. African American Third Grade Test and Change Scores, bySAGE or Comparison

SCORE	SAGE	COMPARISON	DIFFERENCE
Language Arts
Mean Third grade Score	619.42	607.57	11.85*
Mean Change From First grade Pre-Test to Third Grade	104.20	90.05	14.15*
Mean Change From First grade Post-Test to Third Grade	48.89	44.71	4.18
Mean Change From Second Grade Test to Third Grade	28.73	24.47	4.26
Reading
Mean Third grade Score	623.46	609.50	13.96*
Mean Change From First grade Pre-Test to Third Grade	105.64	92.43	13.21*
Mean Change From First grade Post-Test to Third Grade	50.69	43.75	6.94*
Mean Change From Second grade Test to Third Grade	29.87	18.94	10.93*
Mathematics
Mean Third grade Score	592.68	570.73	21.95*
Mean Change From First grade Pre-Test to ThirdGrade	122.39	93.67	28.72*
Mean Change From First grade Post-Test to ThirdGrade	68.82	50.73	18.09*
Mean Change From Second grade Test to Third Grade	48.20	29.59	18.61*
Total
Mean Third grade Score	612.02	596.44	15.58*
Mean Change From First grade Pre-Test to Third Grade	110.72	92.77	17.95*
Mean Change From First grade Post-Test to Third Grade	55.71	47.55	8.16*
Mean Change From Second grade Test to Third Grade	35.87	24.06	11.81*
*significant at .05 level

African American andWhite Achievement. African American students scored lower than whitestudents on the first grade pre-test total scale score, as shown in Table 40. Thisresult is statistically significant for both SAGE and Comparison schools,though the gap between African Americans and whites is larger in the SAGEschools. The change from first grade post-test to the second grade test showsthat the SAGE African Americans kept pace with white students, but did notfurther close the achievement gap in second grade.

The change from the secondgrade test to the third grade test shows SAGE African American students gainingsignificantly more that SAGE white students, closing the achievement gapfurther. This is not true for Comparison students.

Table 40. African American and White Achievement on Total Scale

	First Grade Pre-Test	First Grade Post-Test	Second Grade Test	Third Grade Test	Change From First Grade Pre-Test to Third Grade	Change From First Grade Post-test to Third Grade	Change From Second Grade to Third Grade
SAGE
African American	503.65	559.37	577.17	612.02	110.72	55.71	35.87
White	533.43	582.67	606.99	632.58	99.58	50.93	25.88
Difference	-29.78*	-23.30*	-29.82*	-20.56*	11.14*	4.78	9.99*
Comparison
African American	509.36	552.54	573.45	596.44	92.77	47.55	24.66
White	527.89	578.00	600.37	630.00	94.88	47.12	24.06
Difference	-18.53*	-25.46*	-26.92*	-33.56*	-2.11	0.43	0.60
*significant at .05 level

Hierarchical Linear Modeling

Hierarchical linear modelswere used to evaluate the third grade achievement results using the same seriesof models used to assess first and second grade results. Three sets of analyseswere done. The first, shown in Table 41 used first grade pre-test as theinitial achievement level of the students. The second, shown in Table 42, usedfirst grade post-test as the initial achievement level of the students. Thethird, shown in Table 43 used the second grade test as the initial achievementlevel of the students.

HLMresults. Tables 42-44 provide asummary of the effects of each of the level-1 and level-2 variables for each ofthese analyses. Level-1 effects can be interpreted as the weighted average ofthe within-classroom effects of the level-1 variables. Level-2 effects can beinterpreted as the classroom effects of the level-2 variables.

Table 41. HLM Results for 1998-99 Third Grade Students: FirstGrade Pre-Test as Initial Achievement

Source	Total	Reading	Language Arts	Mathematics
Level 1
Pre-Test	0.671	0.510	0.363	0.432
SES	-5.257	-5.351	-6.843	-3.597
Attendance	-0.205	0.156	-0.053	-0.026
Level 2
· Class Size	-1.083*	-0.966*	-0.723*	-1.483*
B. Class Size	-1.261*	-1.362*	-1.140*	-1.339*
SAGE	-2.910	-6.478	-6.862	2.350
C. Class SES	-13.593*	-10.965*	-12.831*	-18.531*
Class Size	-0.979*	-0.882*	-0.646*	-1.346*
*significant at .05 level

Table 42. HLM Results for 1998-99 Third Grade Students – FirstGrade Post-Test as Initial Achievement

Source	Total	Reading	Language Arts	Mathematics
Level 1
Pre-Test	0.661	0.444	0.294	0.496
SES	-4.490	-6.514	-8.032	-4.247
Attendance	0.205	0.478	-0.111	-0.263
Level 2
A. Class Size	-1.113*	-0.988*	-0.713*	-1.513*
B. Class Size	-1.363*	-1.438*	-1.011	-1.346*
SAGE	-4.045	-7.257	-4.793	2.713
C. Class SES	-14.243*	-11.603*	-12.726*	-18.270*
Class Size	-1.007*	-0.917*	-0.635*	-1.382*
*significant at .05 level

Table 43. HLM Results for 1998-99 Third Grade Students – SecondGrade Test as Initial Achievement

Source	Total	Reading	Language Arts	Mathematics
Level 1
Pre-Test	0.707	0.757	0.433	0.537
SES	-2.369	-3.877	-5.398	-3.205
Attendance	0.178	0.343	-0.158	-0.107
Level 2
A. Class Size	-1.067*	-0.961*	-0.724*	-1.503*
B. Class Size	-1.284*	-1.371*	-1.128	-1.229*
SAGE	-3.601	-6.705	-6.611	4.479
C. Class SES	-14.608*	-11.381*	-13.579*	-18.392*
Class Size	-0.971*	-0.882*	-0.644*	-1.370*
*significant at .05 level

Model A. Class Size. Depending on the test, an increasein class size of one person can be expected to produce a .72 to 1.51 loss inaverage post-test performance. The results for all scores show this effect tobe significant. It is noteworthy that the results of Tables 41-43 for the ModelA are similar, indicating that the third grade class size has little effect onthe three year gain and second grade class size has little effect on the twoyear gain. In other words, the positive effect of reduced class size ismaintained, but not increased from the first grade post-test or second gradetest. As was discussed relative to first and second grade findings, HLM resultsindicate the classroom level effect of the SAGE program and are therefore morepowerful than the regression results in understanding of the program.

Model B. Class Size, SAGE. Combining class size andSAGE participation in a single analysis isolates the effects that SAGE mighthave beyond those produced by lower class size. However, due to the redundancyof these two variables in this case, this effect is not apparent and can beassumed to be zero.

It is instructive to view the results of model B in the context of model Aresults. Due to the multicollinearity of the SAGE and class size variable(i.e., they are very highly correlated), in model B the class size effectincreases from that in model A. Concomitantly, the SAGE effect is negative tocompensate for this effect. This is a statistical phenomenon, not aninterpretable result.

Model C. ClassSES, SAGE. As in the first grade analysis and second grade analyses, thismodel combines class SES and class size. The results indicate that class SEShas a significant effect on the class average test performance. Class sizestill has a significant effect on the test scores once SES has been accountedfor, although the effects are somewhat attenuated.

Effect Sizes

Because the sample sizesused in the analyses were very large, even small group differences will resultin statistically significant results. In order to better characterize theactual differences between groups, effect-size indicators were constructed. Twodifferent indicators were used. First an "unadjusted" effect size wascomputed by dividing the difference between the SAGE and comparison post-testmeans by their pooled standard deviation. Because these means are affected bypre-test, SES, and attendance differences, a second effect size measure was computedadjusting for these differences. This second measure used the raw scoreregression coefficient for the SAGE dummy variable in the regression analysisas an adjusted mean difference and divided this by the pooled standarddeviation. The results of these computations are presented in Table 44 andshould aid in the evaluation of the practical significance of the class-sizeadvantage. The scale score values for the CTB level 10, 11, 12 and 13 tests areprovided in Table 45 as an additional aid to the interpretation of thepractical implications of the SAGE results.

Table 44. 1998-99 Adjusted and Unadjusted Effect Sizes

	First Grade		Second Grade		Third Grade
	Adjusted	Unadjusted	Adjusted	Unadjusted	Adjusted	Unadjusted
Mathematics	.147	.049	.115	.147	.193	.163
Reading	.161	.099	.277	.283	.136	.189
Language Arts	.308	.199	.308	.404	.350	.328
Total Score	.236	.126	.258	.328	.213	.243

Table 45. Scale Score Descriptive Statistics

	Lowest Obtainable Scale Score	Highest Obtainable Scale Score
Reading
Level 10 (first grade pre-test)	355	626
Level 11 (first grade post-test)	407	701
Level 12 (second grade)	423	722
Level 13 (third grade)	427	750
Mathematics
Level 10 (first grade pre-test)	290	629
Level 11 (first grade post-test)	324	680
Level 12 (second grade)	347	720
Level 13 (third grade)	385	740
Language Arts
Level 10 (first grade pre-test)	325	620
Level 11 (first grade post-test)	400	680
Level 12 (second grade)	424	706
Level 13 (third grade)	455	730

Additional Analyses

The 1998-99 SAGE data baseprovided the opportunity to examine some specific factors that may be relatedto student performance within the SAGE program. Specifically, the followingresearch questions were addressed:

1. Is the number of years of SAGE program participation related to individual or class achievement gains in grades one, two and three?
2. Is the socio-economic status (as measured by participation in the school lunch program) of SAGE participants related to individual achievement gains in grades one, two and three?
3. Is the type of SAGE classroom configuration related to classroom achievement gains in grades one, two and three?

A description of the analytical method and a summary of the results of theseanalyses is presented below for each question.

Years of Participation.

The relationship between the number of years of participation in the SAGEprogram and academic achievement gains was examined at both the individualstudent level and the aggregate classroom level. At each grade level, thepost-test scores at that grade served as the dependent variable. Theindependent variables included the previous grade post-test (for first grade,this was the first grade pretest), individual attendance, SES (as measured byschool lunch participation), and number of years of SAGE participation. Thislatter variable is the factor of interest. Table 46 shows the number of casesanalyzed at each grade level by the number of years of SAGE participation.

Table 46. Number of Cases by Grade Level

Years Participation	1998-99 Grade One	1998-99 Grade Two	1998-99 Grade Three
One	382	275	276
Two	1155	404	213
Three	0	1277	1234
Total	1537	1906	1723

The results of thisanalysis showed no statistically significant relationships between years ofprogram experience and achievement gain in any of the content areas whenpretest, attendance and SES were controlled with one exception. This was forthe Total CTBS score for first graders where a significant negativerelationship existed. While all of the first grade relationships were negative,this was the only statistically significant result. All results at grades twoand three were in the expected positive direction, but none was statisticallysignificant. It appears that the "new" first grade students (thosewithout kindergarten experience in SAGE) in the 98-99 cohort outperformed thosewith this experience. It is not known how many of these students had preschoolexperience outside of the SAGE program, making this finding difficult tointerpret.

Analysis of years ofparticipation in the SAGE program was also examined at the classroom level. Forthese analyses, variables were computed to represent average classroom levelsfor each of the variables used in the individual analysis. That is, averagepretest, post-test, attendance, SES and years of SAGE experience were computedfor each of the SAGE classrooms. Analyses similar to those done at theindividual level were then done at the classroom level. For these analyses,there were 92 first grade classrooms, 79 second grade classrooms, and 85 thirdgrade classrooms. In first grade classrooms, no significant relationshipsemerged between average years of SAGE participation and classroom achievementgains on any of the CTBS scores. For grade two, significant results were foundfor the mathematics sub-test and the total score (which can probably beattributed to the mathematics portion). Here a strong positive relationshipexisted indicating that those classrooms with a higher proportion of SAGEexperienced students outperformed those classrooms with lower proportions ofSAGE experienced students. At the third grade level, significant results werefound only for the reading sub-test, however both the mathematics and languagesub-tests showed strong positive, although non-significant, relationships.

Socio-Economic Status.

The relationship between socio-economic status (as measured by the lunchparticipation variable) and academic achievement gains was examined at theindividual student level at each grade. Regression analyses were done for eachCTBS sub-test and the total score in order to address this question. Firstgrade post-test scores, second grade and third grade scores served as thedependent variables. The independent variables included the previous gradepost-test (for first grade, this was the first grade pretest), individualattendance, and SES (as measured by school lunch participation). This lattervariable is the factor of interest.

The results of theseanalyses are summarized in Table 47 which shows for each CTBS test and gradelevel where significant relationships were found. In all cases, the resultsindicate a negative relationship indicating that those with a lower SES index(higher actual SES) outperformed those with a higher SES index (lower actualSES).

Table 47. Significant Relationships for CTBS and SES by Grade Level

CTBS Subtest	98-99 Grade One	98-99 Grade Two	98-99 Grade Three
Reading	*	*	*
Language	*	*	*
Mathematics		*	*
Total		*	*
* = significant negative relationship found

Type of Classroom.

The implementation of the SAGE reduced class size feature has taken a numberof forms. However, there are primarily two configurations: "true"15:1 ratio classrooms where an individual teacher has 15 or fewer students and30:2 ratio classrooms where two (or more) teachers have been givenresponsibility for more than 15 students. As in the past, it was of interest todetermine if there are any achievement advantages associated with either ofthese basic configurations. These analyses were done at the classroom levelwith average post test performance serving as the dependent variable in eachcase. Independent variables included the appropriate average pretest score anda dichotomous variable indicating classroom type. Statistically significantresults were found in two areas of Grade Two as indicated in Table 48. In thesecases, the 15:1 ratio classrooms outperformed the 30:2 ratio classrooms.

Table 48. Statistically Significant Results between 15:1 and 30:2Classrooms

CTBS Subtest	1998-99 Grade One	1998-99 Grade Two	1998-99 Grade Three
Reading
Language		*
Mathematics		*
Total		*
* = significant relationship found favoring 15:1 over 30:2

Aggregated Results vs. Class Level Results

It is important to notethat all of the generalizations provided in the quantitative analyses are basedon aggregated results. By aggregated, we mean that the achievement scores from"small" classes are pooled and compared to those pooled from"large" classes. The result from these pooled scores is that onaverage one can expect a small class to gain .2 to .3 standard deviations (2 or3 months) more than a large class in early grades. A key term here, however, isthe phrase "on average." Small class size does not necessarilyguarantee better achievement in every class.

The data show that a majority of high achieving first grade classrooms wereSAGE classrooms, but with respect to 1997-98 and 1998-99 first grade SAGE testresults, a few large (Comparison group) classes outperformed the majority ofthe small (SAGE) classrooms. Nevertheless, when achievement data of the SAGEand Comparison first grade and second grade classrooms were examined over atwo-year period, classrooms associated with the SAGE program consistentlyformed the majority of top performing classrooms. For first grade, a total of84 classrooms were tracked (66 SAGE and 18 Comparison). Only one Comparisonclassroom ranked among the top 30 based upon student achievement across the twoyears. Fourteen of the lowest performing 33 classrooms were from Comparisonschools. For second grade, a total of 74 classrooms were tracked (60 SAGE and14 Comparison) over a two- year span. Of the top performing 30 classrooms overthis period, only four were from Comparison schools. Nine of the remaining 10Comparison classrooms fell among the 30 lowest performing classrooms over thissame time span.

ANALYSESOF SAGE CLASSROOMS AND SCHOOLS 1998-1999

In this section the effects of the SAGE program on teaching, curriculum,staff development, and lighted schoolhouse services are reported. Dataregarding the effect of reduced size classes on teaching were obtained fromcase studies and teacher and principal questionnaires. The teacher andprincipal questionnaires also provided data concerning curriculum, staffdevelopment, and lighted schoolhouse services.

Teaching

During 1996-97 and 1997-98 classroom events data were collected throughclassroom observations, teacher interviews, teacher logs, teacherquestionnaires, and principal interviews. These instruments revealed that thedominant characteristic of teaching in reduced size classes isindividualization. When classes become small, teachers provide for individualstudent needs through one-to-one tutoring, small group activities, and totalclass instruction where each child receives attention. This increased use ofindividualization is occasioned by less time spent on discipline and more timeavailable for instruction, greater knowledge of individual students, and moreenthusiasm for teaching. Individualization along with a slight increase in theuse of hands-on activities results in more and deeper content coverage which inturn, it is speculated, brings about greater achievement. The type ofindividualization that seems to occur, however, is more process than substance.Teachers basically use direct instruction methods to accomplish establishedgrade-level curriculum.

In an effort to confirm and extend these findings, case studies consistingof more intensive classroom observations and interviews were conducted inselected SAGE schools in 1998-99. Also, the Teacher Questionnaire was given toall SAGE teachers and the Principal Questionnaire was given to all SAGEprincipals. These instruments present data on the classroom teaching of allSAGE teachers, as well as on curriculum, staff development, and lightedschoolhouse programs.

Case Studies

Portraits of teaching in the three different SAGE schools follow. The namesof the schools, teachers, principals, and parents have been changed to protectthe privacy of those involved. Although each portrait was prepared by adifferent researcher and, therefore, differs in format and style, commonthreads emerge which will be addressed after presentation of the caseportraits.

The case studies were conducted in three SAGE schools each representing adifferent type of SAGE classroom configuration: 15:1 Regular, 15:1 SharedSpace, and 30:2 Team Taught. In each school one first grade, one second grade,and one third grade were identified to receive special attention. Over thecourse of the year these classrooms were observed during reading andmathematics instruction four times, formally, and additional times informally.The observations were guided by previous SAGE classroom events findings, butthe observers also searched for new events, practices, or themes. Also, overthe course of the year, the teachers of these classes were formally interviewedthree times and informally interviewed additional times. The formal interviewsrequired the teachers to describe their teaching philosophy, teachingpractices, changes brought about by SAGE, use of individualization,perspectives on previous SAGE findings regarding classroom events, and otherteaching topics. In addition to these two types of observations and interviews,observations were made of selected school events and the principal and parentswere formally interviewed. All interviews were tape recorded and transcribed,and all observation notes were transformed into expanded accounts.

Cleveland Avenue School: Shared RoomClasses (15:1)

School

Location and Community

Cleveland Avenue School is a Milwaukee Public School located in aneighborhood with both advantages and disadvantages. The principal describesthe neighborhood as "a very nice area." Certainly, it’s abustling location with successful businesses nearby and lots of people out andabout. The trees are old and tall, signifying the grace and age of thecommunity. There are several parks nearby; on a nice day the children can walkas far as Lake Park, a huge grassy recreational area. In addition there is agood deal of elder housing around the school allowing the children’s frequentparticipation in intergenerational activities.

The biggest problem withthe area is the lack of children living there. The housing near the schoolconsists mostly of apartments and elder housing. Consequently, most of thisschool’s children are bused in from other neighborhoods. The principalcommented that, as a result, it’s difficult for many parents to get to theschool, even though, she said, most parents would like to be involved in theirchildren’s education.

School Characteristics

Cleveland Avenue School hasbeen a part of this neighborhood for 150 years. The school itself is a very oldbrick building that has been painted white. It is surrounded by a fenced andpaved playground that seems adequately sized for the number of children. Insidethe school, all the classrooms and hallways are light and bright. Everythinglooks freshly painted and the halls act as mammoth bulletin boards, coveredwith posters and student work.

The classrooms are smallcompared to contemporary classrooms, and the SAGE classrooms seem especiallycrowded because they are divided between two teachers, each with 15 students.With tables, chairs, teacher desks, bookshelves the preponderance of learningmaterials (including computer stations) in each room, as well as the constantcomings and goings of volunteers, student helpers, and other professionals,entering and exiting these classrooms is no easy matter.

The most obvious spacialconstraint is seen in the library, located in the basement of the school. Thisis a typical, original basement with cement floors, small doors, and low-slungheating ducts. The library may be adequately equipped, but it is not muchlarger than a regular classroom and did not appear to be used often.

School Population

There are 16 regularteachers K-8 with the addition of 21 special teachers and 10 otherprofessionals and para-professionals who deal with children. The school alsohas 2 administrators, a secretary, an engineer, a cook, and a nurse. The schoolhas an ethnic mix consisting of African American, Hispanic, Caucasian, Asian andNative American students. The principal said, "It’s a verymulticultural-based student population which gives us a lot of opportunity tobe more diverse in what we teach."

School Philosophy

The principal and herstaff, believing this was an "outstanding" school, gottogether two years ago to set a specific direction for the school. They wrote agrant to allow them to become an accelerated school. That plan, the principalsaid, was based on something she and the staff believed, that every child is giftedand talented. Their goal was to make Cleveland Avenue School the first choiceof all students in the city and to retain their students until they completethe eighth grade. The principal said she wanted parents to say, "Yes,it’s an accelerated school; yes, it’s a working school; yes, it’s anaccomplishing school and I want my child to go there."

The principal shared herviews about SAGE participation, some of which are later reflected in theteacher interviews and observations. SAGE forces teachers to share their space,which is difficult. On the other hand, she said, it allows students to havemore personal attention; in return, those students seem to have fewerdiscipline problems. In addition, SAGE brings teachers together. The principalsaid they can no longer go in their rooms and shut their doors; now they worktogether and collaborate.

Reduced Class Size Effects

Interviews with threeteachers and observations of their classes in Cleveland Avenue School revealedfour characteristics of small size classes:

1. high levels of classroom efficiency;
2. a positive classroom atmosphere;
3. expansive learning opportunities; and
4. enthusiasm and achievement among both students and teachers.

High efficiency in theclassroom was the first noticeable effect of small classes. That efficiency wasevident in teachers having broad knowledge of students, both academically andsocially; in teachers having "easy" control of their classrooms; andin teachers’ ability to work on several tasks simultaneously.

Classroom efficiencyappeared to improve classroom atmosphere, evident in levels of compliance andcomfort. The comfort level appeared to be true for students as well asteachers. Besides comfort, students appeared to be voluntarily compliant, afactor which was evident both in students’ ability to remain focusedacademically for remarkably long periods of time and in their willingness andeven eagerness to help each other learn.

Efficient classrooms withpositive atmospheres appeared to provide students with expanded opportunitiesfor learning. Many of those opportunities related to individualization that wasapparent in one-on-one instruction and in small group situations. Teachers’conversations and classroom observations revealed a good deal of hands-oninstruction that occurred with high levels of frequency, variety andavailability. Content was reported to be increased both in quantity andquality, especially with emphasis on higher level thinking skills. The needs ofspecial students were more likely to be met in these classrooms.

Efficient classrooms withpositive atmospheres and expanded learning opportunities led to obviousenthusiasm and achievement. Not surprisingly, everyone appeared to beenthusiastic, teachers about teaching and students about learning. Both alsoappeared to achieve at higher levels. Teachers relayed remarkable achievementstories about students and also referred to their own professional growthresulting from teaching fewer students.

Classroom Efficiency

Knowledge of students. Teachers believed their teaching had changedin smaller classes because they were able to get to know their students better."I...take advantage of the opportunity to work with each student, find outmore about them and they find out more about [me]." Knowing studentsacademically translated into learning advantages as another teacher explained."So every child I know a lot better because [I] can sit down and talkto them and get down on their level. [I’m] able to pinpoint the problems withstudents in their learning and work on just that one goal with them or twogoals that they each need."

Knowing students alsoincludes understanding their social needs. "You know their homebackgrounds, you know how they feel and what they think about things."Another teacher gave an example. "I had one [student who] had a reallyawful home life and was taken out of the home and I think she’s now adopted.But when she came, she wouldn’t talk. She wouldn’t say anything. She couldn’twrite at all....These children are really, really needy." She went onto describe how she helps these children. "They get a hug every daywhen they get on the bus, and in the morning when I pick them up, they have togive me a hug or I give them a hug....They really respond to that, even asmile, you know any type of affection." The teacher explained thateventually the young girl improved. "Now she’s reading on level.Writing is not quite there yet, but she’s sounding words out...[She’s] writingas well as my top child at the end of the year when I had thirty in a class."

Another student in theclass also had home problems and cried a lot. The observer noticed him sobbingin the classroom, the teacher explained, "This happens every day. Wejust ignore it until it stops. [To the child] Brian, go out in the hall."By the end of the school year, Brian’s crying had stopped entirely.

"Easy" control. Teachers reported and observations revealedthat control in a small class was relatively easy. This seemed to be the casefor several reasons. Kids were more focused academically, a small group ofstudents and one teacher could easily agree on what behavior is appropriate,there was tolerance for a wider range of behaviors, and teacher proximity actedas an indirect control.

One teacher explainedmisbehavior and class control this way. "When kids disrupt, it is mostlikely because they’re behind and not understanding what’s going on so theyhave nothing else to do but bug everyone else around them. So [in a class of 30students] you’re always redirecting, redirecting, redirecting, spending most ofyour time redirecting and disciplining kids where you’re not getting as muchinstructional time in." Besides being more on-task academically, shesaid a small group can more easily agree on what behavior is acceptable. "Iget them adapted to my procedures right away, stopping any difficulties. I’mright here so they know very quickly how I run the classroom and eventually wewill run the classroom together. Their style will change it."

Acceptable behavior in asmall class appeared to cover a wider range of activity than one would expectin an average classroom. Asked why this was so, one teacher explained, "Basically,when you’ve got 30 in a room, they stay in their seats unless they raise theirhand and have to sharpen a pencil. With 15, I have reasons for them to get upand move around."

Observations supported theidea that teachers were very tolerant of unexpected student spontaneity. Inmany instances, students were encouraged to shout out answers, and sometimesteachers were able to respond to each student. During a delightful hands-onactivity with M&Ms, first graders fully enjoyed themselves. Throughout theactivity, students talked out loud, commented to themselves, babbled, andchanted little rhymes. One boy was full of rhythm, hands in the air, feettapping, bouncing in his chair. Even when he began tapping his fists on thedesk, no one seemed to be disrupted, or for that matter, even noticed.

Summarizing a readingactivity in the same room, the observer noted that students in the room areallowed every opportunity to act naturally and be themselves. When they feelthe need to walk around a little, they are up and about. Maybe they read for awhile at the windowsill, standing up. When one student wanted more privacy, shepulled her desk away from the others. In fact, every day in this class, thedesks seem to be arranged differently. Students are also free to read withanyone they please. Once the teacher asked a student to pick someone new whenit was time to read, but otherwise students were free to associate and readwith whomever they chose. These children appeared to be more receptive tolearning than if they had been confined to their seats in a class of 30students.

In small classes, thereappeared to be very little direct control of students. One teacher said, "[Students]feel more control of their room rather than me controlling them." Infact control is so subtle, it is hard to identify. All the teachers describedcontrol similarly. One teacher said, "And the discipline, obviously,you’re right there with all of them. They aren’t going to have a chance to actout." Another said, "Before [I] took a lot of my teaching timewith discipline problems. Having 15, I’m so close to them. Generally, I don’thave to say a thing; I just look at them and they shape up and get back towork...So I don’t spend a lot of time with discipline anymore."

Teacher organization. Efficiency in the classroom, besides stemmingfrom teachers’ knowledge of students and "easy" control of theclassroom, was evident in teachers being able to accomplish a number of taskssimultaneously. One teacher explained, "I try to respond when they needhelp; I try to respond right away when I can. So sometimes I’m doing two orthree things at once." Consequently, teachers felt they were moreorganized, students spent more time on task, and students got more help fromteachers.

Organization was evident insituations such as initiating a project. Contrasting the mechanics of doing anart project with a regular class to doing that project with her 15 students,one teacher explained, "[In a large class], if you passed out all thematerials right away, you would have 20 kids acting out over there. [Here]...Ican set it up as I’m explaining and as we’re going along and we’re all doingeverything together."

In other instances, exceptduring direct instruction, teachers were doing several tasks at one time. Oneteacher agreed when the observer commented that she was doing all kinds ofthings at one time. Kids were coming and going while she collected field tripmoney and also marked down the names of kids who did not get stickers on theirpapers. Furthermore, it did not seem confusing to her. Another teacher made ahabit of pulling up a chair next to any student who needed help, all the whilefielding questions from the class while she performed some other task such ascounting the lunch money. It was common procedure in another class for studentsto show the teacher their papers while she was also conducting a reading groupwith a different set of students. When asked to comment on her ability tomanage multiple student projects occurring in her room simultaneously, oneteacher responded, "I don’t know if I’d necessarily be able to do thatall the time with a group of 30 to 32 kids like regular classes have. I’dprobably be exhausted. And I don’t know how much you’d get accomplished."

All in all, there wasmarked efficiency in these small size classes. That resulted in teachers beingmore organized, more relaxed and energized to teach. In the same respect,children were understood by their teachers and seemed to thrive in anenvironment without direct control where they were comfortable and be able tobe themselves.

Positive Classroom Atmosphere

Without question, increasedclassroom efficiency contributed to a positive classroom atmosphere. In a roomwhere students were well understood by their teachers, where control was"easy," and teachers were organized, everyone seemed verycomfortable. In addition students seemed to be voluntarily compliant. Studentsdemonstrated voluntary compliance in two ways: first, they appeared to beindependently focused academically, and secondly, they were willing, almosteager, to help their fellow students.

Comfort level. "With 13 kids, [school] is a lot moremanageable and I’m a lot more comfortable." The comfort level ofstudents and teachers was a dominant feature of SAGE classrooms. Teachersreferred to it as both an interpersonal characteristic and also a physicalcharacteristic. "Basically, the children...feel comfortable coming tome and ask me how to do things...I take my time to work with them."Comfort also meant proximity with students.

"When I’m sitting down with them at theirlevel, they don’t see me standing over the top of them. Some kids getintimidated by that, so I always have an extra chair at each table and I sitdown with them at their level. I think they feel more comfortable and they openup a lot more...They like...being seen as an equal....They still see you as anadult, and they respect you as an adult, but they see you as...moreunderstanding....I feel it’s more respectful because I know I wouldn’t want tobe stood above....If I don’t want it, I don’t want to do it to my kids."

The comfort level ofteachers influenced their role in the classroom. Explaining that role, oneteacher commented, "I walk around and say, ‘Now, does this look right?’I’m more of a guider." Observations also showed how comfortableteachers create happy learning experiences. In one instance, the teacher wasdirecting a reading lesson with students reading individually when she stoppedthe class and made a comment in a way that made the students giggle. She beganlaughing with them. When the students resumed reading, the laughter started upagain and the teacher laughed along with the students. Seizing on theopportunity to have some fun, the teacher pulled out a reference book andshowed pictures which made the students laugh even harder. Later, when asked tocomment on the incident, the teacher acted like it happened all the time andattributed everything to the fact that the book was funny.

Students’ VoluntaryCompliance. In small classes, students appear to be voluntarily compliant,a quality which probably results from being well understood by their teachersand from learning in an organized, non-authoritarian classroom where everyoneis comfortable. Compliance was noticeable in two different kinds of situations.First, students appeared to be very independent learners who could remainfocused academically for remarkably long periods of time. Second, students werewilling, and in many cases very eager, to help their classmates learn.

Teachers say that in largeclasses, kids tend to loose focus; however, in small classes, teachers are ableto train students to stay on task. "I’ve taught them to be independentworkers and not just come to me as a resource." She impressed uponthem that they are a team and it’s just like work. If one person messes up orgets out of line, the entire group suffers. That one mistake makes a problemfor the entire organization or project.

Two specific examplespointed this out. In a second grade classroom, students had been working anentire morning of academic work. Yet, students remained very focused. Evenafter several hours, students acted as if they not only liked to work, it neveroccurred to them not to work. All morning they were on task and obviouslylearning. The teacher commented later keeping 30 students focused for thatlength of time would have been impossible. She said that only in small groupsare second graders are able to work like this.

Similarly, in a third gradeclass, students worked steadily on a long project without any prodding orreminders. Usually one student in each of five groups read while othersfollowed along. As they read, they asked each other questions and completedparts of their worksheets. When asked why the project went so smoothly, theteacher said that no one disturbs or bothers anyone else. The students are 100%comfortable with each other. The teacher went on to say that she could leavethe room for a long period of time and these groups would continue working ontheir own. Overall, these students were lively, happy and moving around most ofthe time, yet they always appeared to be on task.

The first grade teachersaid student compliance was also evident in homework and attendance.

"[My students] are all doing their homework at night [and] bringingit back. When I had a large class, I had one or two or three that would bringit back." [SAGE also improves] attendance because the kids lovebeing at school. In fact, they’re always saying to me, ‘Do we have schooltomorrow?’ ‘No, it’s Saturday.’ ‘I don’t wanna be home. I wanna come toschool.’...They really like school because they get the extra attention."

Students also appeared tobe kind to each other and liked to help each other out. Teachers describedcircumstances where children wanted to be sure their friends understood thematerial. "I have one little girl; she’ll feel bad when her frienddoesn’t know. She’ll show her how to do it." Another teachersummarized students’ care with each other comparing her class to a family.

"These guys are so much of a family andare so nice to each other and work so well together and accept it if someone isa little bit lower than them. They help them out. They’re just so loving. They’rea wonderful group of kids. So, socially, they’ve really learned, not only in myclassroom but outside. If we go walking on a field trip...the people will[comment], ‘They’re so good; they’re so nice together, so nice to each other.’It’s wonderful."

Expanded Opportunities

Small class size led to more learning opportunities for students. Theseopportunities came in the way of individualization, hands-on activities, thequantity and quality of content and in the meeting of needs for specialstudents.

Individualization. Teachers were unanimous in believing thatindividualization was the main effect of small class size. One said itaccounted for 80% of the difference. But individualization referred to at leasttwo different things in teachers’ conversation. Either it meant workingone-on-one with students or meant small group work, both of which allowed forimproved levels of student participation.

Regarding workingone-on-one with students, teachers saw it as an effective means forremediation. One said it was the big thing helping her right now,

"I can get around to each child, even thechildren who are real quiet. Usually [in a large class], you have problemsgetting to them. These children, I mean, I get to see them two or three timeseach day and work with them individually...That’s the big difference....I haveseveral [children who] did not go to Kindergarten....and I try to read witheach one of the four at least a few minutes every day by themselves....and Ichoose two or three of the other students each day [too]. Another teacher said,"I don’t want to leave without them mastering the unit...They weren’tgetting it with the whole group. I still have two that I have to further workwith…but I just keep giving them individual attention."

Others saw one-on-oneindividualization as a way to break away from old habits of direct instruction.

"I get them all up close a lot of the timeor I’m right down there with them at their level...Before SAGE I did a lot ofdirect instruction to make sure I could always see all their little eyes…Now Ido my five-minute spiel up there at the board and right away they’re all intoit and I roam. All the time....[since SAGE] I can break away and get hands-onwith the kids and have fun with it."

Teachers feltthat the break from direct instruction was important for some children.

"This little girl over here that I wasmentioning doesn’t learn well from direct instruction so you have to resort tohitting every learning ability and [each child’s] way of learning in your classroom.Direct instruction isn’t always the way to get at that. I know I don’t learnthat way. I’m very bad. If someone’s doing direct instruction,...I don’t learnthat way....A teacher needs to know that not every child is going to learn thatway. We have to hit all the different types of learning."

One-on-one instruction insmall classrooms also provided some important attention to learning detailwhich might not occur in a larger class. One teacher said

"Each kid gets an opportunity torespond." For example, in a second grade classroom, it was time forstudents to tell about their pictures. Each student rose from the chair andpushed it under the desk. Then, standing and holding their drawings, studentstold their stories. When one girl said, over and over, "she got....,shegot...,she got...," the teacher corrected her very softly until the girlwas saying "she has..." each time. When the next student rose to tellher story, she said, "she got..." and another student quietly said,"she has..."

While all this washappening, the teacher was also taking notes on each speaker. She explainedlater that the notes she was taking were questions she was going to ask thestudents later about their pictures. She said with 30 students, that individualattention would not have been possible.

In another instance, twostudents disagreed about the cause of the major problem in the first threechapters of their text. The teacher gathered the two around her and asked eachgirl to state her opinion ("without mumbling, please") andthen insisted each defend her idea. Overall, one-on-one instruction andattention was an important theme in all three classrooms.

Individualization alsomeant utilization of small groups. Teachers were adamant about the importanceof being able to divide children into groups for instruction. They alsodefended the practice of ability grouping for reading.

"We have all the kids in groups so they’rebeing challenged. They feel comfortable with their group because they know thatwe all have problems and...they fell comfortable reading or expressingthemselves. They don’t have to worry about someone seeing them mess up. [Kidswho are little below reading level do not have to] worry about more advancedkids looking at them, tired that they are taking so long. [Those kids will be]reading ahead and turning the page. So these kids are comfortable with theirgroup, doing their work, but still, they...are being challenged."

In other situations,teachers described grouping children according to ability so that achieverscould help non-achievers. "One boy in my [third grade] room hasKindergarten skills. I have him work with one of the most advanced studentsbecause he keeps the other boy from falling behind." In addition,children were grouped according to personalities and particular skills,depending on the tasks to be performed by the group.

In all cases, the size ofthe class determined that teachers were able to group children.

"If I had a class of 30, I wouldn’t evenbe able to move those groups apart like that....Even when I did my first yearof teaching [with] 30 fourth graders, I couldn’t do two groups The kids thatdid know the work...would have to [be monitored] so much, I wouldn’t get achance to teach that [other] group. [Today] I told the kids to go in the backand play the game and they were able to do that."

Hands-on activities. Teachers felt that hands-on activities wereimportant to children’s learning. Their description of hands-on included alllearning which did not take place with books and papers at students’ desks.They were proud of the variety and frequency of these activities.

"Usually I [usehands on] two or three times a week or more," commented one teacher.Another said, "We use manipulatives every single day." Thefirst grade teacher said she used manipulatives to start each unit. "WheneverI start a unit, I have an overhead and I [use] the overhead and they work withtheirs at their seat, with their manipulatives. She mentioned manipulativessuch as cubes, wooden blocks, cereals, vegetables, and money."

Other examples included theuse of lot of maps, taking a full year to plan and act out a play, going downto the library researching, getting on the computer, finding something on theinternet.

Teachers emphasized thatmuch of that activity would not have been possible with a larger class. With aclass of 30, "I wouldn’t even have enough counters! Furthermore, with alarge class, you might be able to do that but it would be much slower."Another said, "In a large classroom, you had to watch those children,especially first graders, because the hands-on stuff would be all over theroom."

Giving an example, oneteacher described a typical small classroom situation, "So if a kidasks a question all of a sudden, I just go click it in and it’s right thereright up on the TV screen so they can all see it…With 30 kids in the classroom,they wouldn’t sit quietly while I typed it up and got it up onto the TVscreen."

Content Quantity. Teachers did not hesitate to mention thepossibilities of getting more deeply into content as a result of smaller classsizes. They explained that students have more opportunities because they canresearch while the teacher is working elsewhere in the classroom. They can usethe computer or the library, both of which are more available with fewerstudents. In addition, there are more opportunities because students have morefreedom to move throughout the room and the school. Teachers said they finishedup things quickly, at a faster rate, so they could plan additional activities,especially fun, hands-on learning and reinforcement.

In one instance, secondgrade students were reading and one student sounded out the word appendicitis.Quite spontaneously, the teacher pulled out a reference book to show themwhere the appendix is. All the students stood up and leaned to the center ofthe table as the teacher explained the food process through the body and wherethe appendix was in that picture.

Content quality "They’re doing work to learn". Teachersdescribed improved curricular content in terms of the number of complexprojects they were able to initiate. One teacher described special projects onethnicity she was able to do with her small class, six in all, far more thanshe would be able to do with a larger class. During a reading class, a firstgrade teacher had each of her students read word lists individually. She hadeach student read an entire list of words individually. The teacher explained thatin a normal class that would not be possible. "It would take me two orthree times that amount of time to get through the word list. The childrenloved it and when it was over, they wanted to do it again."

Another teacher describedhow she broke down each novel into activities for groups in her class. "Ihave one group who are story reporters who...[do] a story map, a story web onthose three chapters....Then we have the chapter performers [who] perform two orthree pages out of that chapter....I have the novel enrichers [who] have achoice of things they can do...researching....We have the vocabulary finders[who] go through and find words that they don’t know...define the word,[etc]....So we do those kinds of activities throughout the year on chapterbooks." She added that none of the worksheets were book-specific,rather they were designed to be used with any chapter of any book. Thegroupings taught kids to be independent, provided beneficial pairings betweenstudents, and added advanced skill development for third graders.

Another teacher said herclass was going to put on a play. "With 29 kids, it’s hard to put on aplay with your class. Now with 13, we’re going to perform a play towards theend of the school year; we’re going to start now...We aren’t going to have kidsleft out and not get a part in the play. It’s easier to manage the 13 kids whoare going to put it on [at different times and in different places throughoutthe school]."

Improved curricular qualitywas evident in the emphasis on higher level thinking skills throughout the workin the three classes. The first grade teacher explained that when she had alarge class, she taught the whole group with direct instruction most of thetime. "But then I had my reading groups. [My other students] did a lotof writing from the board, filling in blanks, that type of thing...basically, Ihad to keep them busy while I was teaching the groups...Here...they’re reallynot doing any busy work; they’re doing work to learn."

Higher level thinkingskills were also evident in the questions teachers asked. They were open-endedquestions asking children to really think hard. "Now what is theproblem? What has happened so far in the story? Is that a big event or a smallevent? What is the problem of the book? So all the big events surround that bigproblem?" In another class, the teacher asked, "What kind ofslogan could be used to advertise that character? What is the biggest problemdeveloping in these three chapters?" Teachers were inclined to askchallenging questions and were patient waiting for each child to think throughthe questions and respond.

Beyond challenging theclass as a whole, teachers worked to address each student’s ability level."Since the first day of having a SAGE class, I’ve always advanced theadvanced kids and tried to pull up the others." This was a new conceptfor her when she began teaching a small class. "I didn’t have time with29...It’s sad to say and I don’t like to admit it, but... some [did] get lost."

Teachers describedadjustments they made in their teaching so that all students’ needs were met."It was easier to find stuff for the lower kids, but it’s hard tofigure out how to challenge the higher kids. That was my weaker point. [Thestudent’s] mother and I sat down and we talked about it and I’ve been able tolet him build off that. He enjoys writing books so if he finishes he can go andresearch and write a book." Another emphasized the extra effort thatwas necessary. "It’s double planning, but it’s O.K. because otherwiseyou’re sitting there watching some of the kids who are so bored. So it’schallenging but it’s also rewarding.

Special Students. Several types of special students appeared tobenefit from small class size. Children from disadvantaged backgrounds weregiven special help, and other children who showed special education tendencieswere spared from being labeled at an early age. Even children who had alreadybeen given labels were able to spend most of their time mainstreamed in SAGEclasses. All three types of children appeared to achieve at a high level inSAGE classrooms.

One type of special studentconsisted of children who had not experienced Kindergarten. The first gradeteacher referred to several of those students who were new to school. She saidthey had made tremendous progress. "I’m really pleased with them....Ihave more of a chance to help each child....I try to read with each of the fourat least a few minutes every day by themselves." [One of those studentsis] really gaining speed and should be close to grade level by the time sheleaves here."

Other children seemeddestined for exceptional education programs but were withheld from testing andplacement because of their small class size situations. Teachers gave severalexamples. "I did have one kid where the parent thought that at anotherschool her child should have gone LD. But I didn’t see it in that child....Thisyear he’s on level and reading....His mother noticed he was doing much better[and] she was really happy."

In the second example, ateacher talked about a child who could be labeled emotionally disturbed.

"But if we did label her she would have toleave our school because [she would be] put in the emotionally disturbedroom...and we don’t want to hurt her socially. But because of my smallclassroom, I’m able to keep her in my room, give her the structure that sheneeds, and help guide her along. Now if I had a group of 29, there’d be no way.We’d test her, get her the help that she needs in a different classroom because[she] would disrupt my class....I just keep going back to the ED teacher andasking her for advice and working with her alone in my classroom."

Another teacher described achild with severe ADHD and mild ED tendencies. "We haven’t labeled herbecause small class size has allowed her not to be labeled because she does gether focus and attention. We know when she gets up to fourth grade, there’sgoing to have to be a separate IEP made out for her because she is going to belost. We know she’s going to be lost. [If she could stay in a small class]she’d be fine."

Teachers also told aboutchildren who had been identified as having exceptional educational needs butwere being especially well-served in SAGE classrooms.

"SAGE does allow for exceptional educationkids to be reached and you see a greater growth in exceptional educationkids...Even emotionally disturbed children tend to act out less in the smallerclass size than in a regular classroom because ...there’s more guidance fromthe teacher, less opportunity for other kids to act out and set off thatchild."

Teachers went on to saythat LD children normally are a year and a half behind. The three LD childrenin third grade were almost in SAGE classes at grade level or even above gradelevel. "Even your so-called ‘label’ children show in theirre-evaluations that they have made a greater leap because of the small group,one-on-one contact."

Enthusiasm and Achievement

The result of havingefficient classrooms with a positive atmosphere and more opportunities forlearning was evident in the enthusiasm and achievement on the part of teachersand students.

Enthusiasm. The spontaniety in the classrooms was obvious from theobservations. For example, in a first grade class, students clapped when theirteacher finished reading a book. In another situation, students leaped fromtheir desks almost before the teacher finished giving a math assignment.Students ran to their partners, grabbed meter sticks and began hustling around theroom doing their activities. Everyone was involved and active, talking,thinking, debating and problem solving. Students were learning and were veryexcited about it, expressing their enthusiasm without inhibition.

Enthusiasm on the part ofteachers also seemed to be a product of small class sizes. One teacherexplained the importance of that.

"If the teacher isn’t excited, the kidsaren’t going to be excited and they aren’t going to be learning." Besideshelping kids, enthusiasm helped teachers. "When you’re a first gradeteacher…it’s so exciting, especially in the spring. When you set back andlisten to your kids read and watch them do the math, you just get goose bumpsall over. Having the 15 makes a tremendous amount of difference."

Another teacher told howclass size affected her even when school was over for the day.

"I used to go home and I’d be sofrustrated because there’d be one or two children who needed my help or neededto talk to me and I didn’t get around to them until I was driving home and I’dremember that I didn’t find out what they wanted. And so I’d be really, reallyupset with myself. It’s really stressful, not necessarily during the day, butthinking about what I didn’t do...Now I get around to all the children and Ifeel more relaxed when I leave here and I feel like I’m doing a much better joband the children are learning much, much better."

Achievement. Teachers relayed many stories of SAGE students’achievement. They said their students generally finished the year in reading booksabove grade level, and they seemed to be ahead of non-SAGE children in otherskill areas, such as writing and work skills.

All teachers reported thattheir children were reading above grade level. At the end of the school year,for example, a third grade teacher said that she had only two students readingat grade level, and because of SAGE, everyone else in the class was readingabove and beyond the fifth and sixth grade level.

Teachers also describedcharacteristics of SAGE students not related to reading. One teacher said thatthe difference between SAGE kids and non-SAGE kids in the classroom is thequality of the work. "I could go on and on about some of the [other]kids. They are not on reading level. They don’t have really good workhabits." She also said that attendance is not a problem with SAGEkids, but kids who come from other schools, even if their ability level ishigh, have poor attendance and poor work habits.

A first grade teacher wasimpressed with SAGE children’s ability to write. "I mean...you can’tbelieve how adept [SAGE kids] are in writing. I just got a new student in fromanother city in Wisconsin; he’s supposed to be like their top student....Hedoesn’t know his vowel sounds...My children are far ahead of him."

Non-SAGE students needed alot of individual work to keep up. "I keep giving them individualattention...[They] don’t have the necessary skills to...begin with. So I haveto keep going back and redoing."

Generally, teachers believeSAGE students can and will surpass other students as long as they are in SAGEclasses. "I tell parents if you have a child in a SAGE classroom, youcan expect a lot from that teacher." Teachers believe those students"should never be behind again...Hopefully, they will continue to bechallenged and the teachers can find time to continue to challenge them."

Teachers were interested inthe effects of SAGE and felt they grew personally because of their involvementin the project. When presented with a model of the classroom process andeffects based on results of the first two years of SAGE, teachers spent a gooddeal of time studying the model and interpreting it. Overall, they wereimpressed with its accuracy. One teacher exclaimed, "That’s it. That’sexactly what we’re doing."

Teachers also believed that they had achieved professionally as a result ofsmaller classes in becoming better teachers. They said they saw professionalgrowth in themselves, first in learning to individualize and then in usingthose skills at a classroom level. All agreed that they grew gradually in theirability to individualize and that over time, it became easier. "I think[my individualization] started right away. I noticed that some kids needed more,and...by the second year...I knew I could challenge them more....This year...Iknew exactly what I should be doing with them." Teachers alsoreflected on the outcome of their personal growth. Determining that she was nowable to challenge the needs of high students and low students simultaneously,one teacher saw all the students achieving and said to herself, "Wow,last year at this time I wasn’t doing this with any kids!" One teachersaid she noticed gradual development in her ability to individualize, that itbecame easier for her, especially with lower students.

One unexpected outcome ofSAGE classrooms was teachers’ freedom to leave for a day or two forprofessional development. One teacher said she determined that her spellingcurriculum was not effective because it was not showing up in students’writing. So, she attended an inservice to learn a whole different spellingcurriculum. She said that she was comfortable leaving her students with asubstitute teacher, unlike when she had 29 students. "I hated leavingthem because you always came back and heard about how awful they were. Thisdidn’t get done and this happened." But with 13 students, she was ableto leave for a couple days. As a result, she will initiate a new spellingprogram. "It’s not going to be hard…with only 13 [students]."

Parent View

Janet Simms, a parent of a17 year old and a 9 year old, talked enthusiastically about SAGE and ClevelandAvenue School. Janet is involved in the school as a board member and avolunteer mother. The 17-year-old came to Cleveland in the fifth grade. Theteachers brought his grades up and his self-esteem up so that he eventuallybecame an honor roll student. His whole attitude about life changed, and shebelieved it was the staff of teachers who made him feel that way. His teacherdid have 30 students in the class, but she was very strict.

Janet’s younger child isnine and has been in Cleveland Avenue School since Kindergarten. He’s in thirdgrade now in the SAGE program. She loves it because the teachers give thechildren the message that they are important. With the SAGE program, with aratio of 15:1 or less, there is a family atmosphere.

She said one advantage ofSAGE is that children can speak what they feel and they don’t have to wait solong to speak. There are many advantages of "little people" beingable to speak more often. Sometimes, they come to school feeling bad aboutproblems at home. Maybe a parent is stressed out, maybe there’s an illness inthe family, maybe a divorce. A lot of things can go on in the family. Sometimesthose children come to school very sad, but if they are in a small class wherethere is someone to listen, they will probably walk out of school at the end ofthe day with a smile on their faces. Because of the SAGE program, the teacherhas time to listen. That’s important for little people. This experience whenthey are very young will have an impact on their lives. Those first few yearsof life makes people who they are; it’s a time to bring out the most valuableparts of the inner self.

Janet’s son had learningproblems because he didn’t hear for the first year of life. He has madeenormous progress at Cleveland Avenue. He started at a very low reading level,at least one grade behind. Now he’s in third grade and he’s gone beyond thirdgrade, almost a year ahead. They say he has knowledge at an eighth grade level.Janet believes that happened because teachers were willing to listen and hearhim talk. She said teachers have a lot of time for parent contacts. She getslots of calls because her kids are doing something very well in school.Teachers seem to have a lot of time to get to know parents and also kids seetheir teachers as confidants.

There are moreopportunities for these classes, too. As a parent volunteer, Janet’s gone toSt. John’s with classes, to the museum, to Discovery World, and to IMAX. Shesaid they take walks to the lake so they can see the beauty. A lot of kidsdon’t take walks at home because their neighborhoods are not safe or theirparents don’t walk. But here they can do that. Their third graders were able tosee what it’s like to go to college. This kind of activity contrasts with herother son’s experience in another school where they didn’t take any fieldtrips.

She said it will bedifficult for her child to adjust to a large class size, but he will carry onwith relationships and skills he developed in SAGE classrooms. He probably willgo back to his former teachers because he knew he could talk to them at anytime. He learned interpersonal skills in the early years that he probably willbe able to use in the upper grades.

Janet stresses that sheowes so much to SAGE. A really good school and small class size, she said,makes little people good people. She suggests "SAGE kids" will bedifferent as adults. "Watch them in the future; there won’t be as muchviolence in their lives. They learned how to care for each other, how toexpress themselves, how to control anger. These things can all be taught in asmall class."

Meadow View School: Team TaughtClasses (30:2)

School

Location and Community Description

Meadow View is partof the Milwaukee Public School system (MPS). The housing on one side of theschool appears to be predominantly older, lower middle class single familyhomes. Two churches, one of the churches with a school, are found nearby. A newsubdivision consisting of attractive, midsize single family homes is beingdeveloped. The residential community with spacious yards to the north and westof Meadow View does not give the impression of what one expects of an urbanschool environment. Indeed, across the street from the school’s main entrance,a hayfield extends a rather spacious non-urban view. Behind the school and tothe south of the school, the view changes to one of older homes with smalleryards, duplexes, and other apartment-type homes. In general, this area of thecommunity exudes a lack of care or homeownership when compared to thesingle-family homes that dominate the north and west side of Meadow View.

School Appearance, Architecture, and Condition

The main part of the school building is constructed of lannon stone withsome additions to the original school structure consisting of yellow brick. Alarge lawn covers the front and the north side of the property. To the south ofthe building is a parking lot. The playground is behind the building. Overallthe school appears to be an older structure which is confirmed when enteringthe building, walking down the halls, and visiting classrooms (probably 50’s).The building is an older construction, but no immediate disrepair or neglect isnoted. As indicated in the MPS School Selection Guide. the school is wheelchairaccessible and offers services for the orthopedically impaired. Someclassrooms, the library, cafeteria, and the gymnasium are located in thebasement. The halls appear crowded with chairs and tables and abound withcolorful student work which includes posters, artwork, writing, scienceprojects, and other projects. Groups of students and their educationalassistants work at tables in the hall on a regular basis. Space is at a premiumat Meadow View. The school does not have a music room or an art room. Music andart are taught in regular, grade level classrooms. Teachers and their equipmenttravel to the individual classrooms when specials are scheduled. The teacher’slounge is small, and staff meetings are held in the gymnasium. Crowdedconditions are also noticeable in classrooms. Teachers speak of congestioncaused by a high number of students, two teachers, equipment, and materialsmaking their home in a classroom too small in size.

Student Population

Meadow View’s total enrollment is approximately 520 students ingrades K4-6. About 85% of the students are considered minority students.The principal estimates a student mobility rate of up to 45% in the courseof a school year. About 70% of the students are bused to Meadow Viewaccording to the principal. Parent SES is assumed to be low as to be incompliance with SAGE SES requirements.

Type of SAGE School and Staff

Meadow View is now in its third year as a SAGE school and is considered a30:2 team taught type of SAGE school. For Meadow View’s teachers, SAGE hastherefore become synonymous with team teaching. In the MPS School SelectionGuide, Meadow View is described as a regular elementary school with atraditional basic skills curriculum, a modified Program for the AcademicallyTalented and a focus on preparing children to become lifelong learners. MeadowView’s teaching staff consists of 29 teachers in self-contained classrooms,four special education teachers, two administrators, a social worker, apsychologist, a guidance counselor, a half-time librarian, a Title 1 resourceteacher, and three teachers in the specialty areas of art, music, and physicaleducation. Educational assistants and volunteers regularly assist teachers. Avisitor to Meadow View cannot help but notice the large number of volunteersworking with students in the halls and classrooms at any given time. When askedhow he gets volunteers for Meadow View, the principal stated that all he has todo is ask. Retired Meadow View teachers return on a regular basis to help outand so do community members, parents, and students (middle school and highschool) who are involved in the Responsive Schools Project. One teacher’smother, for example, is a regular helper in the third grade classroom. Grandma,as the teacher’s mother is generally referred to, enjoys volunteer work withchildren. Volunteering at Meadow View is particularly enjoyable, she states,because the principal is so accessible and easy to talk to. She feels that heappreciates volunteers which has not always been her experience at other MPSschools.

Principal Interview

The principal has been in this leadership role at Meadow View for sevenyears. He tells his teachers that he has high expectations and high standardsfor students. And although Meadow View’s students’ test scores do not reflecthigh achievement at this point in time, it is his goal to have Meadow View’schildren as well prepared as any child attending any school, private or public.During an interview, the principal made the following comments about the SAGEprogram at Meadow View:

"The SAGE program itself has changed a lothere at school, it has people work more as partners, more as team members,working together as opposed to individuals. I think that is very important. TheSAGE program has also helped us to support and maintain our community …withbefore and after school programs. Because of the SAGE program there are manymore things that we can do; it services the parents and also services thechildren and that is due to the program. Overall is has affected the studentsand students’ achievement; our students are doing better since we have thesmall ratio of student to teachers. And I think too that the program will be alot better if we could deal with one of the problems we have here at MeadowView. We have, I could say a negative …, and that is our mobility rate. We arenot showing a growth that I would like to see, and I think that is because ofthe high mobility rate. If it was a little bit lower, I’m sure that we wouldhave much more substantial growth."

The high rate of student turnover, the principal points out, has a negativeimpact on the SAGE program because the school often does not "get achance to get a child from kindergarten that goes all the way to third grade."He does not anticipate any major future changes as a result of the SAGE programbut would like to see SAGE extended to include all of the grades at the school.SAGE has had a very positive impact on Meadow View, and some parents, theprincipal notes, who had left the district, are coming back because of the SAGEprogram.

Teaching

Three classroom teams at the school were studied since Meadow View has a30:2 team taught type of SAGE program. A first grade, second grade, and thirdgrade team were selected using SAGE evaluation data from previous years andinput from the school’s principal. The first grade teacher team, Team One,consists of two male teachers now in their third year of team teaching in theSAGE program at Meadow View. The second grade team, Team Two, has a femaleteacher who had previously taught a first grade SAGE classroom at the schooland a former male fourth grade teacher with no prior experience in the SAGEprogram. The third grade team, Team Three, is made up of two female teachersnew to both SAGE and Meadow View.

Teaching experience for the teachers on Team One, includes eight years ofteaching and a recent Master’s Degree for one of the teachers and 14 years ofteaching experience for the other team member. For Team Two, one member has 14years of teaching experience and two Master’s Degrees; the other member has 17years of experience and is currently working on a Master’s Degree. Team Threeconsists of a teacher with 22 years of teaching experience and three Master’sDegrees and a teacher with 9 years of experience who is also working on aMasters’ Degree. All teachers have been involved in continuing education coursework and inservices. However, none of the course work or inservices relatedspecifically to small class size teaching.

Team One’s philosophy is guided by high expectations for all students. TeamOne teachers believe that all students can learn and having fun while learningis important. The philosophy that guides one of the team members of Team Twocan best be described as life-long learning and to have fun with learning. Theother member believes that all students can learn; it is just a matter offiguring out the key to each child. Team Three’s philosophy emphasizes highexpectations for students and parents. Recurrent themes evident in all threeteams are high expectations and the belief that all students can learn.

Changes Made and Anticipated.

When asked to describe major changes in their teaching as a result of theSAGE program, teachers of Team One noted that they have more time to preparebetter for their students, to do more hands-on activities, and to teach insmall groups. One of the teachers observes:

"I mean before we were really unable to dothat at all and I think a small group is, to teach a small group is a lotdifferent than to teach in a large group. Uhm, a lot more hands on things, youknow, more activities because there is always that extra person around to helpwith kids who are having difficulties."

A teacher on Team Two notes "The biggest change for me that I cansee right now that I’m going to have lots more energy, lots more energy. A lotmore energy to work with students on a one to one basis." For TeamThree, teaming is a new experience, and change for them means also adjusting toteam teaching. Both teachers also observe that they have more energy and can domore because they do not "get as personally drained" now thattwo teachers are in the classroom. Generally, teachers agree that they get toknow their students better, can assist students more frequently, and are ableto cover more content. During an interview, a teacher on Team Two makes thefollowing comments:

"Well, I think, the major change hasreally been that we are able to reach students. I noticed that before we hadSAGE, I think, I taught well then, but it took me a while to know each student.Maybe a month before I probably got to know how they reacted, what triggeredthem, what turned them on educationally. Now because there are two of us in theclassroom, we are able to get around to each student for more personal time,and so we are learning the students’ personality and their goals much quickerthan before we had the SAGE program….we always had a pretty strong curriculum.I just think that we are doing a better job at, ah, with the curriculum. We canactually do more than we were doing before. …students are working morerigorously now, and where we weren’t able to meet some goals, we are gettingcloser to that now."

Teachers did not indicate any major changes in teaching style as a result ofthe SAGE program. One teacher noted that having a teaching partner has made hima little less rigid and more open to working with small groups. Anotherteacher, now in the third year of SAGE, described the teaching style as "…prettymuch what I’ve been doing all along, it is just a matter of working with yourpartner and being flexible and team teaching. I’ve been doing it for threeyears now; I’m pretty much in the mode of it." Two teachers indicatedno change in teaching style. When asked specifically about the use of directinstruction in their classrooms, the teachers generally indicated that theystill use direct instruction but less of it. They report that they can do moreexperimenting with activities and have more flexibility to do other things likesmall group instruction. "Well, I still do use direct instructionmyself too like I have in the past. I just think it’s probably changed in thesense that maybe I don’t have to do as much direct instruction …"observed one of the third grade teachers.

In sum, what can beconsidered the most significant change as a result of the SAGE program toclassroom life is described by one of the experienced SAGE teachers as betterlearning opportunities. A Team One teacher states "…so I think it hasopened up a lot of opportunities for us and the students to learn better."Better learning opportunities in a SAGE classroom are described by anotherMeadow View teacher as opportunities for active and engaged learning ratherthan busy work:

"I think that one of us is, you know,talking or engaging the kids, or the kids are talking, engaged in activelearning of the lesson more than here you have this, do this 45 minutes orwhatever …. one of us, it seems, is always talking or saying something, youknow, so and the kids, I think, then they are more alert, and they are moreengaged that way."

Reduced Class Size Effect

Better or increased learning opportunities in SAGE classrooms may beattributed to better knowledge of students, more individualization, fewerdiscipline problems, and more content coverage and hands on activities. Theseperceived effects of the SAGE program were noted frequently by the teachers inthis case study.

Knowledge of Students.

Teachers generally report that they do know their students better and thatit takes less time to get to know them. While knowledge of the students’personal life is important to the teachers and more opportunities are noted forgaining such knowledge, teachers seem to place an emphasis on knowing anindividual student’s academic performance level. One teacher states that "…beforeI had SAGE, it took me a while to really figure where someone actually wasbecause you have so many children. And by having two of us, …you are able tosee who needs help with maybe some skill of reading …" Another teachercomments "So I think having two teachers in the room really helps youknow how well each child is doing." Teachers seem to use the more indepth knowledge of students’ academic levels to provide more individualizedlearning opportunities.

Individualization.

Case study teachers institute individualized learning opportunities indifferent ways. Team One, for example, uses small groups and educationalassistants to individualize or "tailor" instruction to the individualacademic needs of students. Three educational assistants work with small groups(1-5 students) of students for one hour on a daily basis. The teachers describehow they individualize:

"I don’t know the word individualize isexactly what we’ve got. We started like you said with more of a small group. Iguess "tailor" is a better word for us. We’re tailoring the way we’reinstructing to meet the needs of certain pockets, I guess that we see. Uhm, andright now we’re going to focus more on that middle 50 per cent, that’s wherewe’re gonna devote most of our energies because they’ve already kind of got agood base going. And those are the children we feel are going to learn the mosteven from the individualized instruction. Uhm, so they’re going to get themajority of that individualized time. The children who are in the very bottomare also going to get time, but it’s going to be supplemented between the aides– we’re having them, teaching assistants, now – the teaching assistants aregoing to be working on the early success program with them…. I guess, probablythe way we individualize more than anything is probably with the work of aidesand maybe even some peer tutoring where we will have some individual studentsgoing out and reading with certain kids who need more attention or lessattention".

Team Two also uses educational assistants and the reading resource teacherto help with groups of students at varying reading levels. Two educationalassistants help for about 50 minutes in the classroom and another aide comesfor about 30 minutes a day. A special education teacher pulls out the only LDstudent for one hour daily. Additionally, the teachers on this team alsoindividualize homework assignments and classroom activity packets so thatstudents are "able to have extra work at their level whether it’schallenging or whether it’s reviewing, whatever they happen to needindividually." In another classroom, students receive individualattention from a volunteer who helps in the classroom approximately four days aweek. An educational assistant takes a small group of students for about 45minutes on a daily basis. Teachers state that better knowledge of theirstudents earlier in the year allows them to begin individualization, especiallysmall group individualization with educational assistants, at the beginning ofthe school year. With gaining experience in SAGE classrooms, one team notes,modifications occur over the years, and small groups may be expanded to includemathematics and not just reading.

Predominantly, individualization is perceived as assisting students in smallgroups based on their academic needs and grade level objectives rather thanstudents’ interests. Team One makes the following comments aboutindividualization in their classroom:

"We try to bring in their individualinterests and things through the stories, maybe that we read and tap them thatway, but we don’t necessarily say, you know what, what animal would you like tostudy about, things like that. Uhm, so I think it’s individual in a group sortof sense. You know that – we can kind of tailor to the four or five that needmaybe certain kids additional reading help. So we can, uhm, based on theirlevels individualize in smaller group maybe about four or five that will helpthem.

I think at our grade level, individualizationalso means just more practice. The kids get more practice. You know, they, whenwe break up into small groups, you know, there are, if there is five kids, theyget more opportunities to answer questions, they get more opportunities topractice the words, they get more opportunities to, you know, to do a lot ofthings that in the large group, you know, it would be cut back so much or theywouldn‘t get any practice in some cases. "

Team Two also states that they individualize their teaching and lessonsaccording to the child’s strengths and weaknesses. While the teachers regardindividual students’ interests of importance, they give priority to reachinggrade level objectives first and then would allow for more freedom in studentchoices.

For Meadow View,individualization needs to be considered in conjunction with team teaching andworking out a routine in the classroom that involves two teachers. Classroomobservations provide snapshots of how teachers have developed a routine of"lead and support." It appears that one teacher generallyleads the lesson or activity while the other teacher supports what is takingplace in the classroom by monitoring students’ progress, helping individualstudents or small groups, testing students, and reinforcing appropriatebehavior. Field notes contain examples of these practices.

During one school visit, the following observations were noted in fieldnotes. While one teacher leads the reading activity with the whole class, theother teacher is calling on individual students. The students come to theteacher’s desk, and she reviews projects with individual students. The topic ofthe project is "Inventions that Take Us into the Millenium." Earlierin the week, the students had created the actual invention/machine with mathmanipulatives and other materials in class. The teachers took photographs ofeach student and his/her invention. The teacher shows these pictures to thestudents as they come up to her desk. The students have also written a storyabout their invention. An Educational Assistant is helping with this activity;it appears that she sorted out papers that need corrections or are notfinished. The teacher reviews the writing with each student. Then the studentand the teacher make some corrections together. Examples of corrections arerun-on sentences and capitalization errors. The teacher asks questions, pointsout errors, asks the student for explanations, and then explains corrections tostudent. Some corrections pertain to the format and content of the writing. Forexample, the writing needs to have a title, the inventor’s name, and adescription of what the invention is or how it works. When all of thecorrections have been made, the teacher staples the photograph to the writing.The work will be displayed in the hall, where other invention projects arealready on display.

The following observation was made on another day. It is approximately 10:00A.M. outside the first grade classroom. Five students and one of the firstgrade teachers are seated at a table out in the hall by the classroom door.These five students and the teacher are working on reading skills. In themeantime in the classroom, the remaining twenty students and the second teacherare gearing up to review the answers to the morning work. On another occasionthe following scene of teaming and individualization was observed during avisit to the school. One teacher is seated at a table out in the hall workingon reading with one student. The teacher explained later that the schooldistrict requires a reading assessment of each student in the class. While theteacher in the hall works on reading assessment with one student, the otherteacher instruct the whole class. Students are working on what is called themorning activity. The work is on the board and students copy it on loose-leafpaper and put the answers/corrections on their paper.

Discipline

Teachers state that having two teachers in the room has impacted disciplinein two significant ways. First, discipline issues can be handled in a moreconstructive and positive way. As the teachers observe, you have more optionsbesides sending the student out of the classroom. One of the team members candeal with the situation immediately, and a lot of frustration for students andteacher can be avoided. Secondly, not only can teachers attend to disciplineissues immediately, they also do not lose their "momentum" in thelesson. Being able to continue with instruction, teachers note, is a positiveeffect of having two teachers in the classroom. One teacher makes the followingcomment about discipline:

"…if there is a discipline problem, beingable to deal with it immediately, keeps all the rest of the kids on task, andit prevents them from going off task also. To me what I’ve seen and experiencedin this, I mean, we spend so much more time teaching, one just picks up forother one if they are dealing with something. And there’s not - like a beatisn’t missed. And if you were alone, you know, that would not behappening."

Teachers agree that fewer discipline problems or being able to have oneteacher deal with the problem, has a positive effect on instructional time.What teachers describe as "not missing a beat" or being "ableto continue with instruction" translates to better and increasedlearning opportunities for students in SAGE classrooms.

Content Coverage

Teachers feel that they are more likely to cover the curriculum required bygrade level and at times are able to go beyond the required curriculum. Thefollowing teacher comments illustrate the teachers’ perceptions about contentcoverage:

"…We always had a pretty strongcurriculum. I just think that we are doing a better job at, ahh, with thecurriculum. We can actually do more than we were doing before. Students areworking more rigorously now, and where we weren’t able to meet some goals, weare getting closer to that now."

"Ithink we are accomplishing more, covering more things."

"We’vecovered more things and deeper …"

"…youreally dig deep into a subject."

Teachers seem to also see a connection between more and deeper contentcoverage and opportunities for more hands-on activities.

Hands-on Activities

Finding the time for hands-on activities and making the activity aworthwhile learning experience are of importance to the teachers. One teacherat the school states:

"…a lot of times when I had things to passout, it was almost time to collect them. So by having two adults there, we’reable to do more of the hands-on. Plus we are able to actually observe what theyare doing with the hands-on, if they are using the correct manner not just tobe playing. So they actually learn from them."

Team taught SAGE classrooms seem to make hands-on activities possible atMeadow View. A trip to the pumpkin farm, for example, was developed intosubsequent math and science activities such as cutting pumpkins open, pickingseeds out of the pulp, and counting the seeds. The teachers said that if theyhad been by themselves, meaning one teacher with 30 students, it would havebeen a giant mess and impossible to keep students focused on what they weresupposed to do. The following two teacher comments illustrate teacherperceptions about hands-on activities in team taught classrooms:

"I am really finding that you can get to alot more material and more quickly than before. And we are able to expand moreWe’re able to do hands-on; it’s a great way to learn. With 30:1 it’s reallydifficult to get all the pieces set up and … without losing your mind. We’reable to do a lot more hands-on things and …ah, to teach a lot more than we hadbeen in the same amount of time.

I think it leads to more hands-on thingsbecause even in science you can give kids opportunities to do things and thenthere are two people who are both monitoring and helping them with problems andthings like that. And so it keeps them more on task and if they are more ontask, they’re learning more. And with the two of us, I know, I’m more willingto try certain things because there is always a second person to be able tohelp out."

Teachers frequently mention math manipulatives such as cubes, shapes, andcounters when talking about hands-on activities. Team One estimates that theyuse hands-on activities 2-3 times a week. Team Two states more than once aweek. And teachers generally feel that they are more likely to engage inhands-on activities now that two teachers are in the room.

SAGE Difference

Teachers were asked to reflect on the extent children who were in a SAGEclassroom last year are different from children who were not in a SAGEclassroom previously. Two teachers perceive no significant differences, academicallyor socially. But both note that in general the kids seem to be more dependenton the teacher for attention or direction than they were used to. One of theseteachers, a first year SAGE teacher, coins this observation "learneddependency on the teacher" and speculates that having two teachers, asin team taught SAGE classrooms, may have contributed to this dependency.Another teacher observes that groups of children differ from year to year, andin some years he noticed a greater academic readiness than in other years. Oneteacher observes that the school’s high mobility rate often makes it difficultto determine a student’s prior SAGE experience. SAGE classrooms are naturallyoccurring classrooms, and students from other SAGE and non-SAGE schools move inand out of Meadow View’s classrooms throughout the year. Another teacher whohas been a SAGE teacher at the school since the inception of SAGE makes thefollowing observations:

"I don’t know what they were like beforethey came to me. But I know, I think the ones who were in SAGE last yearbecause they were all together last year have some kind of a family unity bond.So that was kind of special. And then those who came in new … [inaudible]…became part of the class. But I really don’t know what they were like beforethey came so I really wouldn’t be able to see the difference. Academically, Ithink, uhm, our kids that were in SAGE were a little bit stronger, you know, onaverage. Some would be low, but uhm, that’s about all I can really say."

This view, however, is not shared by her colleague who states that he couldpredict who had been in a SAGE classroom. To this teacher, the child who hadbeen in a SAGE classroom seemed to have more confidence about the second gradeenvironment which students new to the SAGE experience of a reduced-sizeclassroom seemed to be lacking.

Accuracy of Model

In general, teachers did not recommend adding or subtracting elements fromthe Model of Teaching, which is based on the results of the first and secondyear of SAGE. One team perceived individualization to be the main effect ofsmaller class sizes, another team stated that student achievement was the maineffect, and the third team thought that students’ needs were met better. Twoteams felt that the model is lacking a parent component. One teacher makes thisstatement:

"Just one thing to consider about it ishow - and this is a hard thing to determine, but how involved are parents. Youknow. Where do parents fit into the model? Because I think a lot of the kids,you see the big improvements where the parents are really working with them athome and that tends to be, I think, it’s a strong support for what the kids aredoing. And we can tell them something for 6 hours a day but if they’re gettingsomething else for the other …"

Only one team engaged in speculation about redrawing a few arrows or addinga few arrows. Noteworthy is a suggestion of more time for preparation in teamtaught classrooms. And two teachers note a relationship between "moreteacher enthusiasm" and team work. However, the model does not accountfor issues relating to team taught situations. One team states that they workwell together but were teamed by chance without knowing each other prior tocoming to the third grade team situation at Meadow View. Both taught at otherschools prior to this year. They feel that good working relationships of teamsare not the case for all teams at the school, and team member combinationschange from year to year. The teachers state that teachers at the school havethe option to team teach the class or to split up the group of students. Thesetwo teachers plan to team teach again next year.

Parent Perspective

A parent who is also an educational assistant at Meadow View providesinsight into her and her children’s experience with the SAGE program. Theparent shares the following observations:

"I’ve worked here for about 8 years. And Isee the difference with the SAGE program, with the two teachers in theclassroom than the one teacher in the classroom. You know if a kid is having aproblem, he has two teachers in the classroom. You have one teacher that couldtake that kid out while the other teacher teaches. You know the other kids, youknow what I mean, I love the program. I really do. And I mean, I’m going backto school and I’m gonna be teaching. And I hope that when I do finish my degreethat I end up being in the SAGE program because I love the team teaching thing.It’s wonderful. It can’t get no better than that. Two teachers in the classroomand then you have an assistant that comes in and helps. But like I said itcan’t get no better than that.

What I have noticed is my daughter, she iscoming home, uhm, she’s talking about how much she’s learned in school. She’stalking about the different things that she learns. Ah, she’s coming homesaying big words. [Laughter] "Prediction" and "Discover"and you know and she, um, she came home one day, and she was telling me aboutthe different groups that she goes in. I noticed that the teachers have twoseparate groups. And by the assistant being in the classroom, then they havethree. So the kids get to learn more. You know, like the group … that are …like one teacher would have 15 kids and the other teacher would have the other15. But you know the kids are really learning because the teacher don’t haveone teacher is not trying to teach 30 kids and is distracted when one kid actsup, and he can’t pay attention to the other rest of the other 29. So I reallylike it. I love the program. And I wish they would keep it because my daughterhas learned a lot with the SAGE program. And she's gonna be going to the secondgrade. And they have SAGE in second grade so I know she’ll learn a lot more.You know because with the one teacher in the classroom, I know it’s hard. It’svery hard. And the SAGE program works a whole lot for my daughter."

Oakdale School: Regular Reduced Size Classes(15:1)

School

Oakdale Elementary School islocated in an economically middle class integrated neighborhood. However, fourbuses are bringing mostly African American children from a predominatelylow-income neighborhood to make up approximately 200 of the 330 studentsattending the school. About 64% of the student population qualifies for free orreduced lunch.

The building itselfconsists of two stories arranged in a semi-circular fashion. As one wandersthrough the building, it is not the classrooms or support areas that are mostnoticeable; it is instead the pervasive feeling of peace, pride, and priority.There are signs of the efforts to create peace over every classroom door, onbulletin boards, and in the busy sounds of learning that waft down the halls.

Particularly impressive to me were the rules listed in the library and mediacenter (LMC) that were clearly created with student input and followed throughon the theme of PEACE which is pervasive throughout the school. While I wasthere, I heard no raised voices or disgruntled sounds. It really was peaceful!

According to Jane Peters, the principal of Oakdale, the staff consist ofover 50 professionals, numerous additional teacher aides and volunteers whocome into the classroom from two to three times each week. The volunteerprogram is very well developed. In large part it consists of grandparentvolunteers who are retired University of Wisconsin professors. Support comes inthe form of different reading support teachers; music, art, physical educationteachers; computer teachers and signers for students who are hearing impaired;teachers with experience in working with students who have a variety oflearning or emotional disabilities; and a very active LMC program director. Twounusual support features at Oakdale are the Book Room and the PARRoom. The Book Room is a room that contains copies of all books used in theschool. In this room there are listening centers, Big Book centers,Anthologies, Plays, Book Boxes, Leveled Books, professional resources forteachers, TV/VCRs with closed captioning, and a Publishing Center for studentsand staff use. The PAR Room (Positive Attitude Room) is located off by itselfand is staffed by a full-time teacher. There are small study spaces forcompletion of work, individual help, or simple separation areas for problemchildren. Although the teachers acknowledge that some students do not like tocome to the PAR Room, most students do not look at it as somethingembarrassing. Rather, they look at it as something they need for a short time.This room did not feel like a time out room or a negative place, but clearlythere were rules, and it is considered a work space, not just a punishmentplace. The atmosphere was in keeping with the whole school’s emphasis onbecoming peacemakers.

The type of SAGE classroom practiced at Oakdale is fifteen students with oneteacher in a separate classroom. Working hand-in-hand with SAGE, at Oakdale, isan active Title I program which has led the staff to using "bestpractices," especially for reading. They are in the process of planningand working on the same kind of program for mathematics. There areapproximately three classes of every grade. Kindergarten, second, and thirdgrade classes were all 15 or under. A source of dissension is the fact that allthree of the first grade classrooms had 18 or 19 students and the teachers feltthis was not good. One first grade teacher described the situation in this way:

"I did have 18 (students) and we were notpleased with that. They did take care of the problem of being over 15-1 in 2ndand 3rd grade but (voice raised) Jane Peters did not listen to us because wewere asking for another classroom or something that we could do toalleviate...because you know if you took three out of each classroom, out offour classrooms, you’d almost have a full classroom...another full class. Shedidn’t do anything. She said, "Well, you’ve had it good in the past".We weren’t real pleased with that and I did call DPI (Department of PublicInstruction) on it. I sometimes wonder if the test is valid because we areover! We did ask our reading support person to go to the principal (JanePeters) and talk about getting assistance. So, in the past week she has come upwith some assistant time which should alleviate getting to some of thekids."

During the course of the school year, this problem was solved by hiring twofull time teacher aides for first grade.

It appeared that the case of the teacher with 18 students might not beworrying about student learning as much as the power struggle between the firstgrade teachers and the principal. However, later in the school year, the firstgrade teachers seemed to be content with the situation as it was. Surprising tothe teachers, their class numbers stayed at 18 or 19 all year with less studentmovement than they had experienced in the past.

SAGE is not limited to class size at Oakdale. Lighted Schoolhouse is part ofthe SAGE plan. Conferences are done in three different ways to encourageattendance. They do some outreach for parent education both in the area of parentingand some technology training. Ms. Peters, the principal, refers to Oakdale as afamily school that focuses on the area of family literacy as the road forproviding a variety of ways for people to be connected to the school. They havechild care available, clubs, Retired Seniors Volunteers (RSVP), and a varietyof educational opportunities for teachers and parents. Not only are educationalopportunities available, but leadership roles in sharing knowledge with otherprofessionals are part of the expectations for the teaching staff of Oakdale.They have a tradition of studying, learning, planning, and then sharing withothers in their field.

A Picture of Teachers and otherProfessionals at Oakdale

Oakdale has been a SAGE school for three years. The principal, Jane Peters,has a BSE in elementary education and a Ph.D. in education.

When asked about herphilosophy, Jane Peters talked about student learning and teaching tochildren’s strengths. She mentioned that all people are life-long learners andthat the real philosophical dilemma is to find the best way to modelcollaboration, trust, and good learning as teachers work toward understandingthat wherever a student is academically is OK; it is just not OK to be thereforever. Jane sees her role very clearly:

"I think the other part of my job isreally finding the right questions to ask. And I used to think my job wasworking with teams and doing the stuff of the school. I came from acoordinator’s background and that’s what you do, but I have learned, at leastfor this building, is finding the right questions to ask, paying closeattention in meetings and such so that when teachers talk in meetings and so onto understand what that talk means. You know, what is it, what kinds of thingsdo I need now that I have made the assessment? What is it that I need to helpthem or to provide for them to get to this next level, or move on? So, it’s alot about the questions you ask and I came to that realization later ratherthan earlier (laughs at herself) But I do think it’s how you ask them, and whatyou ask and then seeing what comes of it. You can only work from theirstrengths just like the kids, what they can almost do. So, there is a wholelaundry list of things we don’t do that probably need to be done, but we haveonly so much energy and we look at what is the biggest payoff to us in terms ofstudent learning. I guess that is another part of my job, this constancy ofpurpose. Providing the lead to say when we do this we are really doing theseother things. Being able to say OK that really is a need but before we can getthere, we need to work here. We use the lessons we learn from our strengths tohelp us to accelerate our progress through our not as strong areas, but we mustwork from what we can do before tackling what we can almost do."

Jean John has been a teacher at Oakdale for 14 years and has a BSE inelementary education.

Jean described herphilosophy of education in terms of priorities. She is very experienced inteaching first grade but still makes student learning a top priority. She likesto give hands-on experiences that students can relate to. She sees this as anopportunity to see things from a child’s perspective. She likes to incorporateproven strategies such as the Reading Recovery model into her own teaching. Amajor goal is to get the children to be independent learners and use strategiesthat will enable them to work independently and do their problem solving ontheir own. Ms. John commented that flexibility is a necessity for working successfullywith other professionals at Oakdale, but that most people place the good of thechild first. She admonishes everyone to place children as a top priority"cause that’s why we’re here!".

Cooperative groups, out ofseat activities, center work, manipulatives, interactive writing, whole group,and small group activities are methods Jean uses to teach the strategieschildren need to move from what they know to what they can do next.

I noticed that Jean often referred to what she had learned as a result ofthe Title I and literacy training that the whole school participated in. Itseemed that she was very confident and comfortable speaking to me. She made apoint of letting me know that knowledge and training brings change to the wayshe approaches her class. That was part of what she meant when she talked aboutflexibility.

Carla Berg has a BSE and has been at Oakdale for eight years.

Ms. Berg has eighteen yearsof teaching experience. Her philosophy toward her students is one of a personaltype classroom. She places great emphasis on knowing her students asindividuals and maintains a quiet, but deeply interested persona in theclassroom. Her own words describe her philosophy best:

"I get to know the students in a morepersonal way. It is almost like I think of them as my family and so I likeknowing them better an’ particularly goin’ out every day to help them asindividuals or in small groups. I am able to get up from my reading table androve through the room to make sure that the children are doing the right thing,keep them on task. So, I don’t have any behavior problems. With smaller groupsI can meet their individual needs and receive an immediate response from thechildren. I know each child’s particular style of learning. We are able to do alot of role playing when children have conflicts; stop right then and workthings out. Now they don’t have to sit in their seats so much and it’s muchmore fun. I find it is professionally rewarding too."

Ms. Berg is a soft-spoken African American woman. Although she was verynervous about being interviewed, when she was in front of her class, there wereno signs of any nervousness. She seems to be very student centered in herteaching with an instinct for student success in what the principal describedas a challenging class. Her gentle manner holds true in her dealings with allof her students.

Jackie Park has a BS in Finance and accumulated 20 years in banking/financebefore returning to school to earn a BSE.

Jackie came from a 20 year business background where she freely admitted themoney was better, but the work became mindless. She returned to school to get adegree in education because she felt called to this profession. She brings agroup of unusual experiences to her classroom and a powerful dedication to herstudents and her professional life.

Jackie is a very articulate woman with whom it is easy to develop animmediate rapport. Professional discussions flowed and frequently were long.She made a point of putting the students first in every conversation we had.From observation, it appears that she follows through with her beliefs. As shefunctions throughout the day, she seamlessly takes over the role of schoolleader and willing mentor for others. She was an outstanding spokes woman forSAGE. Her words about what she believes are clear:

"The goal I have as a teacher is to helpthe child become responsible for their learning, not to tell but to facilitate,to take a child from whatever point they are at when they come to this schooland lead them to the next point, whatever that is ...as far as we can get them.I set high expectations. We work through frustrations and in our school we seea lot of frustration from the kids. We see a lot of turnover. So, my goal isfor them to become responsible and to love it. To just love learning and becurious about everything around them. That’s my goal.

I make certain that I interact with every kideach day. The kids in my class are actively learning all day long. I do a lotof one-on-one and hands-on learning. We have the "best practices"that we have been working on as a school and we have also adopted a literacyprogram or literacy philosophy. We have gone from not clearly understandingwhere each child is at to understanding where each child is and taking them tothe next step. Assessment drives the learning and makes clear the teachingpoint. So, I use the specific Teaching, Learning Cycle, with different tools,all to help the student find the answers they need to move on with theireducation.

The role of the teacher takes a lot of energybecause students move through material faster [than before SAGE]. But it isalso very exciting to see students moving forward with such deep learning. Theteacher has to model enthusiasm and love of learning so that the children canalso learn to become life-long learners. SAGE experience has changed myteaching life. It took me a while to adjust. I had my first six years here,learning to work what I call crowd control, learning to be effective in theclassroom, at least as effective as I could be with a larger class. With SAGE I had to adjust from the group mode to theindividual mode. The questions about how to keep children busy and active as Iwas working with other parts of the group gave me practice in centers and"busy" work. Now almost everything we do is as a small group and I gobetween those small groups and interact with every group, every day, whether itis reading, science, math or personal conference. I know my kids and I canconcentrate on guiding the child to the next step".

Teaching

Individualization.

Every SAGE teacherinterviewed felt that individualization was a strong point of growth.Particularly effective was the response of one African American parent speakingto me at a small PTO meeting. When asked if she could describe anychanges from the SAGE classroom when it comes to your child’ learningactivities or attitude toward school anything that is maybe a result of havinga smaller class size, she responded.

"I have a daughter in the special needsand having a smaller class, she gets more of the teacher’s attention so sheperforms better...not only performs better, but she finishes her things. Theteacher is able to check and see, oh, she did finish the project. In fact,where we were there were 23 kids in the class and this wasn’t always the case.I had to ask the teacher to send home one copy of all of her work so I knowwhat Cindy didn’t finish and that is not the case here (Oakdale). She is sohappy to have finished it (assignments) in class that her self-esteem hasimproved as she finishes and she is able to get the sticker or the stamp thesame as the other kids.The smaller class also makes it possible for theteachers to get to know the kids a lot faster, so they can assess theirstrengths and weaknesses right away and start working from those points rightaway."

Teachers at Oakdale commendSAGE for making more individualization possible. However, the idea ofindividualization comes in many forms. One common thread among all descriptionsof individualization is that the students’ reading program is planned from theplace the student begins and changes according to the students’ needs. Becauseof the high level of volunteerism and the classroom use of support personnel,students rarely work in groups larger than three, and those groups often changeaccording to the specific learning needs at any given time. Although one firstgrade classroom teacher spoke of working with groups of four or five, groupslarger than three were never observed once she had the use of a teachingassistant and volunteers. When teachers speak of whole group, they are oftenreferring to preparation for either individual work or very small groups. Allteachers cite the movement away from text as the guiding force and worksheetsas the individualization. There are many learning centers located in allclassrooms and they seem to be utilized on a daily basis. Because writing isbasically an individual process, it is important to note that teachers feelthat students do more writing for strategic purposes. The assessments ofstudents’ reading are generally done on an individual basis, using runningrecords and anecdotal notes. Most teachers considered this an important part ofthe individualization process because of the constant need to evaluate whereeach student is to go with the next phase of their reading program. Jackie Parktalked about individualization in the following manner:

"More individualization, definitely, thatseems to be the whole key. As I said before, for me, it took a while to learnhow to do that, and do less whole group and now I hardly ever do any wholegroup instruction. What I’ve learned to do now, when I do math and science [inintegrated for], I give them a problem to work on, then I take some of the kidsto observe and dialog a little bit, then I send them back to work on theproblem and dialog with the other group a little. Yes, you can get in much morecontent, and the discussions incorporate more of what the kids have to say.Plus, they all participate because it is more intimate. The children often areguided to using higher level thinking skills during the discussions, I thinkbecause we have more time....no, we don’t really have more time, our time isbetter spent."

It seems to me that many times, when an Oakdale teacher speaks aboutindividualization, they are stating that the students’ learning plan iscentered around where they are academically (in reading, for instance). Thisdoes not necessarily mean that the student is taught one-on-one for much of thetime. In fact, very little one-on-one was observed at Oakdale. Most of all, Isaw small groups or even whole groups of fifteen, with constant individualcontact from the teacher were prevalent. Many classes used the support staffand grandparent volunteers to make their groups smaller. The classroom teacherroamed around between groups speaking to individual students or in some cases,the students came up to the teacher. When it was working, it was like a well-orchestratedmusical with everyone in the room playing his or her part to make the chordsall fit together. One math class was particularly effective in individualizinga program for each student by using a whole group approach. I had followed Ms.Berg’s class to Math while she met with a parent. The following is an accountof Ms. Berg’s math class where an individualized program was presented throughwhole group instruction.

Outside the classroom a student and adult were working on reading at a deskin the hall. The teacher was seated at her desk, with a short line of studentswaiting to speak to her while classical music quietly played. Out of the sevenstudents in the room, those who were not in line worked in the"areas" of the classroom. Some worked on math at the back table usingmanipulatives, some finished their math and went to either their own desk orcarpet to read, and some worked at checking out work individually with theteacher. This was clearly a normal routine. Some students hummed with themusic, others talked quietly, but none was uninvolved in what was obviously awork time.

Each child had a personal math book for January and February, made by theteacher. The teacher was running what looked to be a mini-economy book storewhere students had to trade in certain math points for pretend money topurchase items from their classroom store. Students did conversions of pointsinto money with the teacher’s help.

The following conversation occurred:

Student: What does threequarters equal?

Teacher: What is 20 plus20?

Student: 30, 40.

Teacher: What is 5 plus 5?

Student: 10.

Teacher: What is 40 plus10?

Student: 50.

Teachers: How may quarters isthat?

Student: 2.

Teacher: What is 50 plus20?

Student: 60, 70.

Teacher: What is 70 plus 5?

Student:  75.

Teacher: Then what does 3quarters equal?

Student: 75 (Happilysmiles).

Teacher: What are you goingto do with that information?

Student: Use it!

In this impressive, individualized lesson, the student who had the questiongained a way to solve problems, but the quiet patience with which otherstudents waited was also impressive. Perhaps they knew their time for questionsor answers would come also. This illustrated a real advantage of a smallerclass.

Knowledge of Students/Discipline

Most of the teachers interviewed felt that knowledge of the student andimmediate attention leads to an easier time with discipline. Ms. Berg describedthe connection very clearly:

"SAGE has been wonderful because I can getto the students and it’s more of a personal type of classroom. It’s almost likeI think of them as (pause) my family and so I like knowing them better, andparticularly going out every day to help them as individuals or in smallgroups. I think this helps children to feel better about themselvesindividually, and I think it helps with the discipline.

The type of things I try to do are to praisethe children by going around and giving them the personal touch or a tap on theshoulder or stickers. I think it has created a closer relationship. When achild has a particular interest or problem, we can give them more of a personaltouch. It keeps the children on task and then I don’t have behavior problems.The groups are much smaller that I can teach and I can more meet theirindividual needs. I also receive an immediate response from the children

I have an individual in my room who had a badsituation at home and from knowing him so well, I knew that I had to give himspecial attention that day and not stress him out by giving him extra work, andgiving a lot of extra hugs to get him through the day. The payoff for this boyis I think he may be more successful in life, especially coming from a homewhere he has to do whatever he is doing on his own. So SAGE gives him a lot ofsupport and some experience with how to function successfully as a member of aclassroom society. I hope that transfers to society at large."

As SAGE teachers at Oakdalefound that they spend less time on discipline, they also discovered that thereseems to be more instructional time and according to most, the time is spentinteracting student to teacher and student to student, which creates moreexcited enthusiasm from students and teachers. They all suggested that theyfelt they were able to delve deeper into subject matter and still offerindividual growth because they are not spending time disciplining students.

Content

Teachers seemed to feelthat they were able to go deeper into subject areas often by using integratedactivities, such as science and math together, and more hands-on activities.They all mentioned the use of manipulatives. However, there was almost an airof apology when they discussed their math program because they said it was beingdeveloped. Having smaller classes enabled the scheduling of blocks of time forreading and for math. Most teachers felt that longer blocks of time would leadto greater content coverage in math, but this was the first year.

Perceived Effects

To a person, all agreed that the main effects of SAGE were in greaterstudent success in school. More individualization seems to be the key as far asthe professional staff, parents, and volunteers are concerned. Students,generally are seen as having more enthusiasm for their own learning becausethey are so completely involved in a personal kind of learning. Assessmentsthat guide the teachers seem to be better directed at individual student’slearning. The model of classroom events from the first two years of SAGEexplained what staff thought was going on. Jane Peters was particularlyarticulate in explaining the vision of this model:

"I think the major success this year andevery other year of SAGE has been the increased achievement. The children areachieving so much more than they have in the past with the old class size. Aswe have developed our assessment strategies and become much more involved inlooking at assessment as part of the learning cycle, it’s become much moreapparent to us how much learning goes on when we reduce the class sizes. Forexample, we are right now, at or above the district average in our writingassessment. When you consider the neediness of the building, that’s a prettydramatic kind of thing and I think some of that is a result of SAGE.

However,although there is a lot of truth in the model, my experience would say that alot of this is certainly right on to what we have seen, but we have to becareful when we talk about class size being totally responsible or even in amajor way responsible because I think it’s the small class size combined withsomething else, the staff development that goes along with what you want peopleto do with those smaller class sizes. People have to teach in different waysfor the class size to give you the payoff that you really want it to have.

Certainly there will be less discipline justbecause there are fewer bodies in the class but to make this really count ischanging. How do you help children to learn about being more self-disciplined?How do you set up structures within the classroom to build a sense ofcommunity? Those kinds of goals come through teaching teachers and having anexpectation in the building (for use of the teachings learned in staffdevelopment). You know you have an idea of how you want the climate to be and Ithink class size is a big part of that but it isn’t either what you start withor what you end with, it’s part of the mix, if you will.

I certainly see more teacher enthusiasm. Ithink that some of this enthusiasm comes from better assessment and being ableto document and watch developing growth. Not to have to infer growth orsurmise, but to be able to see it contribute to more teacher enthusiasm becauseteachers see how much they are able to do. And class size makes it easier to dothose best practices...I think the small class size does just that."

The theme of staff development, working hand-in-hand with SAGE, appeared tobe consistent. It was a little ironic that no one mentioned that staffdevelopment is a part of the SAGE plan. Even Jane Peters seemed to think of themany different areas of staff development that have taken place at Oakdale asan add-on to SAGE rather than an integral part of learning how to make maximumgains from a small class.

Even though there are other aspects of SAGE, most of the people at Oakdalecentered their discussions around the differences that a small class could makein a student’s learning. Given a lead to talk about community or the importanceof continual development of teachers still seemed rarely to lead to integrationof SAGE with other sectors of school life. Clearly SAGE is visualized as onlylimiting class size and other developmental programs are completely separatefrom SAGE.

The SAGE program seems innately to strike a positive chord with everyoneconnected to Oakdale. In many different words, people told that reducing classsize is a "no brainer". It also makes sense that the strong staffdevelopment component will make the whole experience of smaller classes asuccessful one for students.

Case Study Summary

The three case studies confirm and clarify SAGE findings about classroomevents from 1996-97 and 1997-98. The conclusion that the main effect onteaching of reduced class size is increased use of individualization isstrengthened. All of the teachers involved voiced this opinion and extensiveindividualization was witnessed during every visit to the three schools.

The individualization that occurs is procedural rather than substantive. Itis, as one teacher remarked, "tailored instruction." Studentsare neither permitted to pursue their own interests nor are they provided witha personalized curriculum that varies from the established curriculum. Allstudents learn the same content and skills, but they may learn them at a differentpace and in a different way. As the Cleveland teachers reported, instruction isbased on the individual's current level of proficiency. It builds on what eachstudent presently knows or knows how to do.

One-to-one tutoring takes place in reduced size classes both for shortperiods as the teacher monitors an activity and for longer periods when aproblem is encountered. But, the dominant mode of individualization is fluid,homogeneous, small groups led by the teacher. The reading and mathematics lessonsare usually taught according to a common format including a total classoverview, orientation, or directions and then arrangement of the class intogroups monitored or taught by the teacher as well as teacher aides orvolunteers. With groups of 3 or 4, each student can actively participate. He orshe can raise questions, make comments, show work, receive feedback, andrethink and revise.

What appears to happen to teachers when they teach reduced size classes isthat they develop a different mind set. Instead of viewing their pedagogicalworld as one class of 25 students, they view it more as 15 classes of onestudent. Given the speed with which teachers adapt to teaching reduced sizeclasses, it appears that this state of mind has been present all along but madedormant by the stress of large classes. Once this state of mind is activated,it pervades all facets of reduced class size teaching. Even when the class of15 students is taught as a whole, each student is heard and each receives a"tailored" response.

How individualization as practiced in reduced size classes leads toincreased learning over students in larger classes might be explained withreference to constructivistic theory. Constuctivism suggests that humansconstruct or create their own understandings based on their prior knowledge.When they experience new ideas, they either fit them into their presentknowledge structures, revise their existing knowledge structures to accommodatethem, or develop new knowledge structures. For this process to take place,articulation and critique are essential. For knowledge to grow, we must givevoice to our present understandings. We must say what we believe or show whatwe know how to do. Others can then offer feedback, advice, or challenge, whichcauses us to rethink our understandings. It also causes the person giving thecritique to rethink the conceptions that he or she holds. The opportunity forarticulation and critique to occur increases dramatically as classes becomesmaller. More students can share their understandings more often and teachersand others can offer more personalized feedback more frequently. As studentsdisplay their knowledge, teachers can target their assistance. The resultshould be more and deeper understanding or learning.

1998-99 case study data also support the conclusion drawn from previous SAGEdata that individualization is related to increased knowledge of students, toreduced discipline, and subsequently, more instructional time, and to greaterenthusiasm. Teachers reported that they know their students both academicallyand socially much better which results in a more personal, relaxed, family-likeclassroom atmosphere. They also said, and observations substantiated, thatclassroom time was totally devoted to teaching because disciplinary behaviorwas generally not needed. It is interesting to note that discipline problemsdisappear in reduced size classes not only because inappropriate behavior isinstantly recognized in a small class and can be given a response with no delayand because teacher-student proximity reduces its occurrence, but because in areduced size class inappropriate behavior is redefined. As several teachersrevealed, in a small class students are given more freedom. Many behaviors nottolerated in a large class, because of the problems they create, such aswalking around the room, are acceptable in a small class. Further, there isalso less misbehavior because students' greater understanding in small classescauses them to be less confused and, subsequently, better behaved. In relationto teacher enthusiasm, the case study teachers provided similar responses asSAGE teachers from prior years. They indicated that reduced class size and thestudent progress that they see energized them and caused a great deal ofsatisfaction and excitement regarding teaching.

The dominant mode of interacting with students that emerged from the casestudies is direct instruction but growing use of hands-on activities was alsonoted. Increased use of hands-on activities such as manipulatives inmathematics, drama in reading, and other non-text book, non-worksheetactivities occurs in reduced size classes because normal impediments to theiruse are not present in reduced size classes. In reduced size classes fewermaterials are needed, teachers have more time to prepare, misbehavior is lesslikely, and basic curricula have been covered.

Consistent with previous findings, the effects of reduced size classes andthe individualization it fosters as reported by the case study teachers isincreased learning. Teachers reported that more content, including contentdesignated for the next grade level, and deeper content were acquired bystudents. Other outcomes or effects were also reported, however. These includecritical thinking, independence, and social responsibility, as well asenthusiasm for school and improved attendance. The personal and social effectsoccur because of the family-like environment and individual attention thatexists in the reduced size class. Students are freer to express themselvesverbally and physically, they help each other as they observe the teacherhelping them, and they see their ideas as having worth when the ideas receiveattention from others. A teaching team new to SAGE at Meadow View School reportedthat they thought team-taught, reduced-size classes caused students to becomemore dependent. This could be seen as a sign of the success ofindividualization rather than an undesirable outcome. That is, reduced classsize students learn to speak up and ask for help when they have a problem orshare an achievement about which they are excited rather than to remain silent.Their actions could actually signal that they have become more independent, notmore dependent.

Most case study teachers see reduced class size as benefiting all students,but they commented on its particular benefits for special education students.They suggest that the individualization that it causes may prevent future needfor special education for some students, spare early labeling for others, and,for those already diagnosed, increase the time they spend in the regularclassroom. Simply stated, the needs of all students can be met in the reducedsize class, the teachers believe.

Some differences across classes were seen in type of SAGE, but differencesbetween reading and mathematics and differences across the three grade levelswere not observed. In terms of subject and grade level the pattern of teachingin which individualization is prominent does not appear to vary. In terms of typeof SAGE classes, team taught classes share the general profile of all reducedsize classes, but they achieve it in different ways to some extent. With twoteachers, they are able to do many things simultaneously, such as teaching,monitoring, disciplining, preparing, and evaluating. That reduced class sizebrings about a change in classroom events and has an impact on student learningis not in doubt. It is, however, as the principal at Oakdale School remarked,only "part of the mix." Individualization may not always bedone well.

All Teachers

Self reported data from of all SAGE teachers regarding their teaching arecontained in Table 49. These data obtained from the Teacher Questionnaire againshow that individualization is the major teaching product of reduced classsize. Table 49 shows that for the total group of SAGE teachers, the teacherbehaviors that received the highest ratings and rankings in addition toindividualization are engaging students in discussion, using hands-onactivities, teaching rather than disciplining, covering more content, andenthusiasm for teaching. Those behaviors receiving comparatively lower ratingsand rankings are the more student centered behaviors of using students' priorknowledge, using cooperative groups, giving students choices in learningactivities, and integrating content from several subjects. Problem solvingactivities and in depth content have ratings and rankings between the high andlow groups of behaviors.

Teacher behavior by grade levels is reported in Table 51, and teacherbehavior by type of SAGE classroom is reported in Table 52. It can be seen thatthe general pattern is also descriptive of each grade level and each type ofclassroom. The major teacher behaviors are individualization , student engagement,and to a slightly lesser extent, hands-on activities, teaching rather thandisciplining, and covering content. A difference regarding grade level is thatteachers at each succeeding grade are less uniform in their responses. Firstgrade teachers have the strongest agreement about the effects of reduced sizeclass on teaching and third grade teachers have the weakest.

The Teacher Questionnaire also reports data about teaching in terms ofstudent participation. At least 75% of the teachers see their students as beingmore attentive, participating more in class, asking for help more often, beingmore enthusiastic about tasks, and displaying more self-direction.

The principal estimates of the reduced class size effect on teachingpresents the same picture of teaching revealed by the case studies and theTeacher Questionnaire. Individualization, diagnosis, and treatment of learningproblems, assessment of progress, immediate feedback, and an environment orhuman relationships conclusive to learning were frequently mentioned. Oneprincipal commented in the following way:

"Establishing a classroom student-teacherratio of 15:1 or smaller, has allowed the teachers more individualizedinstruction. This has allowed the teachers a clearer picture of each child'sneeds. The curriculum is being met and students are progressing because of thefriendly small group environment. The social needs of the students are beingcared for which allows more and better time for instruction. Since lessdiscipline occurs, more time can be spent on the learning environment."

Table 49. Total Questionnaire Results, Grades K-3 (Percentages)

ITEM	Strongly Disagree	Disagree	Neutral	Agree	Strongly Agree	Ranking*
1. More time teaching	1	1	4	49	40	16.74%
2. Covered more content	1	0	6	46	41	6.68%
3. Integrated content	0	0	4	53	37	2.85%
4. More depth	0	0	6	53	36	5.34%
5. Individualization	0	0	0	34	59	23.33%
6. More engaging	0	0	2	40	52	12.91%
7. More Hands-on	0	0	2	44	48	9.08%
8. Student’s knowledge	0	0	6	55	32	3.47%
9. Problem solving	0	0	2	54	38	5.43%
10. Cooperative groups	0	1	12	46	33	3.03%
11. More opportunities	0	0	9	57	27	3.83%
12. Teacher enthusiasm	0	0	3	35	55	7.30%

N=417

*Average ranking of top three choices from the list of 12 items

Table 50. Questionnaire Results for First Grade, Second Grade, andThird Grade (Percentages)

	FIRST GRADE (N=110)					SECOND GRADE (N=101)					THIRD GRADE (N=106)
ITEM	SD	D	N	A	SA	SD	D	N	A	SA	SD	D	N	A	SA
1	0	2	4	48	44	2	0	3	50	41	0	0	4	53	36
2	0	1	5	43	48	2	0	6	47	41	0	0	9	51	32
3	0	0	1	55	42	1	0	3	55	37	0	0	10	53	29
4	0	0	8	45	45	1	0	4	58	33	0	0	10	56	26
5	0	0	2	25	72	1	0	0	34	60	0	0	0	42	50
6	0	0	4	35	60	1	0	3	41	51	0	0	1	44	47
7	0	1	3	44	51	1	0	3	48	45	0	0	4	50	39
8	0	0	4	55	39	1	0	9	49	37	0	0	7	56	30
9	0	0	5	55	39	1	0	1	55	39	0	0	2	55	36
10	0	2	14	42	40	1	2	14	44	36	0	3	15	41	33
11	0	1	5	58	35	1	0	12	60	21	0	1	13	56	22
12	0	0	2	29	66	1	0	6	33	56	0	2	2	43	44

Key

SD = Strongly Disagree

D = Disagree

N = Neutral

A = Agree

SA = Strongly Agree

Table 51. Questionnaire Results for Different Types of SAGEClassrooms (Percentages)

	15:1 Reg Ratings (N=86)					15:1 SS Ratings (N=9)					30:2 TT Ratings (N=41)					30:2 FT Ratings (N=5)
Item #	SD	D	N	A	SA	S D	D	N	A	SA	SD	D	N	A	SA	S D	D	N	A	SA
1	1	0	7	48	44	0	0	11	22	67	0	0	5	39	56	0	20	20	40		20
2	0	1	7	38	54	0	0	11	22	67	0	0	0	49	51	0	0	0	60		40
3	0	0	9	48	43	0	0	11	44	44	0	0	15	61	24	0	0	40	60		0
4	0	1	9	45	44	0	0	11	33	56	0	0	7	46	46	0	0	0	100		0
5	0	0	1	24	74	0	0	0	33	67	0	0	5	34	61	0	0	0	60		40
6	0	0	0	41	59	0	0	0	44	56	0	0	2	46	51	0	0	0	80		20
7	0	0	5	40	56	0	0	22	11	67	0	2	2	39	56	0	0	20	60		20
8	0	0	20	47	34	0	0	11	44	44	0	2	2	56	39	0	0	20	60		20
9	0	0	4	52	44	0	0	11	11	78	2	2	7	51	37	0	0	20	80		0
10	0	4	12	48	37	0	0	33	33	33	0	7	27	27	39	0	0	40	40		20
11	0	2	12	51	35	0	11	22	11	56	0	5	17	51	27	0	0	20	60		20
12	0	2	4	38	55	0	12	0	20	68	0	0	0	51	49	2	0	8	22	68

Key

SD = Strongly Disagree

D = Disagree

N = Neutral

A = Agree

SA = Strongly Agree

A Model of Reduced Class Size Teaching

The kind of teaching andlearning that reduced class size produces is displayed in

Figure 1. The case studies as well as the data from the TeacherQuestionnaire and the Principal Questionnaire reaffirm that individualizationbrought about by more instructional time, more knowledge of students, and moreteacher enthusiasm plus an increase in hands-on activities results in more andsomewhat deeper content, and, in turn, more learning. The case studies alsosuggest that student critical thinking, self-confidence, and socialresponsibility are outcomes of reduced class size.

Figure 1. A Model of Reduced Class Size Teaching and Learning

---

---

---

Rigorous Curriculum

For the purposes of the SAGE evaluation, rigorous curriculum has beendefined as curriculum that is consistent with national standards in reading,language arts and mathematics as proposed by professional associations. Table52 reports the extent to which the curriculum in the areas of reading, languagearts and mathematics in SAGE schools is consistent with these standards. Thesedata, derived from teacher perceptions on the Teacher Questionnaire, showoverall agreement with the standards in both curriculum areas. In reading andlanguage arts, the areas of greatest agreement are a) students are encouragedto choose books of personal interest, b) the names of parts of books aretaught, c) students are taught to apply a variety of decoding strategies, and d)students are introduced to literary forms. The areas of least agreement are a)students are taught to critique non-print media, b) students are taught howlanguage can be adjusted for different audiences, and c) students areintroduced to texts representing a range of historical periods.

In mathematics, the areas of greatest agreement are a) students have theopportunity to connect mathematics to everyday situations, b) students learnthe enumeration system through concrete experiences, c) students have theopportunity to connect mathematics with other subject areas, and d) basic factinstruction emphasizes thinking strategies. The areas of least agreement are a)use of calculators in appropriate situations, b) the concept of chance isexplored through actual events, c) perimeter and related areas are developedintuitively, and d) students create their own mathematics problems.

Teacher perceptions concerning rigorous curriculum are very similar to prioryears, with greater agreement in reading and language arts than in mathematics.The areas of greatest and least agreement within both curricular areas arenearly identical to 1996-97 and 1997-98.

Table 52. Rigorous Curriculum, Grades K-2 (N, Mean, and StandardDeviation)

	N	Mean*	SD
Reading/Language Arts
Students introduced to texts: represent range of genres	410	4.16	.64
Students introduced to texts: represent range of historical	414	3.57	2.57
Students introduced to texts: deal with topics relevant to real world	413	4.11	.59
Students introduced to texts: variety of ethnic, culture contexts	413	4.07	.72
Students taught to apply variety of decoding strategies	415	4.63	.54
Students introduced to variety of interpretative strategies	410	3.87	.82
Students taught names for parts of books	414	4.70	.51
Students introduced to literature terminology	416	4.43	.78
Students taught to categorize texts: fiction or non-fiction	415	4.10	.93
Students taught to categorize texts: topic or theme	412	3.91	.79
Students taught to categorize texts: author	415	3.97	2.15
Students taught to make associations among texts	412	3.77	.73
Student taught aware of how language can be purpose adjusted	416	3.68	.79
Students taught aware of how language can be audience adjusted	414	3.41	.83
Students encouraged to choose books interested in reading	414	4.73	.50
Students apply lang/conventions: critique/discuss print texts	407	3.64	.98
Students apply lang/conventions: critique/discuss non-print media	406	3.20	.99
Students apply lang/conventions: writing to develop interests	410	4.14	.79
Students apply lang/conventions: speaking to develop interests	409	3.89	.82
Mathematics
Students write own mathematics problem about real or imaginary	412	3.20	.84
Students encouraged to develop own strategy for solving problems	412	3.94	.76
Opportunity to investigate open problems have more than one sol.	413	3.61	.82
Write in math class to reflect and demonstrate understanding	412	3.22	.96
Mathematics language and symbols introduced in context of expl.	412	4.07	.67
Opportunities to make connections between mathematics and other	414	4.06	.61
Opportunities to make connections between math & everyday	412	4.25	.59
Estimation when working with quantities, measurement, comput.	414	3.77	.74
Opportunity to explore and use estimation strategies in real sit.	413	3.60	.75
Learn enumeration through concrete experiences	412	4.26	.68
Discuss, model, draw, write about their understanding	411	3.73	.80
Instruction of facts emphasize development of thinking strategies	409	4.13	.68
Develop own computation strategies and algorithms	404	3.43	1.06
Calculators used in appropriate situations	408	2.59	1.21
Instruction includes concrete experiences with metric units	412	3.20	1.00
Concepts of perimeter, area, volume are developed	410	3.21	.92
Opportunity to explore geometric shapes through concrete exp.	414	3.78	.75
Opportunity to work with 3-dimensional figures	413	3.49	.86
Formulate & solve problems involving collecting & analyzing data	413	3.60	.82
Make predictions, inferences, decisions from data	415	3.77	.75
Concept of chance explored by collection of data and other events	414	3.10	.86
Concrete and real experience to develop fraction concepts	410	3.63	.91
Recognize, describe, extend patterns	413	4.15	.73
Create patterns using materials and discuss patterns	413	4.00	.79

*Mean score using five point Likert Scale

The Principal Questionnaire results support the finding that the reading andlanguage arts curriculum and the mathematics curriculum generally areconsistent with national standards. About 80% of the SAGE principals see theircurriculum as being mostly or completely compliant in these areas, as seen inTable 53.

Table 53. Principal’s Perceptions of Rigorous Academic Curriculum(Percentages)

	Not Implemented	Somewhat Implemented	Mostly Implemented	Completely Implemented
Area
Reading/ Language Arts	0%	20%	53%	27%
Mathematics	0%	20%	57%	30%

Professional Development

Results concerning general and personal professional development asperceived by teachers and principals are contained in Tables 54, 55 and 56.Principals' views of the professional development program in their schools arereported in Table 54. The results show that new teacher transitions,collaborative planning, professional development, and staff evaluation programsgenerally are being implemented in SAGE schools.

Table 54. Principal’s Perceptions of Staff Professional DevelopmentPrograms (Percentages)

	Somewhat Implemented	Mostly Implemented	Completely Implemented
New teacher transition program	20%	50%	30%
Collaborative planning	23%	30%	47%
Professional development plans	23%	40%	37%
Staff evaluation program	10%	35%	55%

Table 55, which reportsboth the context and process of professional development in SAGE schools, showsthat professional development is a prominent feature of SAGE schools. In termsof context, most teachers agree that in their school it is ongoing, isadequately funded and supported, and brings about a change in classroompractices. In terms of process, most teachers agree that in their school, theteacher is seen as a learner, a collaborative climate exists regarding staffdevelopment, teachers and principals are knowledgeable, teachers use effectiveapproaches and strategies, and student achievement and performance is the goal.Areas of professional development in which there is some disagreement byteachers are the use of study groups to learn about change and innovations;out-of-school collaborative learning; learning about innovations prior todeciding about their use; assessing teachers based on student learning and developmentactivities that include theory as well as practice.

Table 55. Teachers’ Perceptions of Professional Development GradesK-3 (Percentages)

ITEM	STRONGLY DISAGREE	DISAGREE	NEUTRAL	AGREE	STRONGLY AGREE
1. Ongoing & Regular	1	4	6	46	42
2. Changes in Practice	1	6	20	56	17
3. Adequate Funding	5	13	23	42	16
4. Widespread Support	1	9	23	48	19
5. Joint Learning	9	40	26	17	7
6. Study Groups	8	27	27	29	10
7. Improvement Plan	2	6	22	54	14
8. "Teacher as Learner"	1	5	26	48	18
9. Staff Development	1	5	19	57	18
10. Precede Decisions	4	15	37	37	7
11. Program Evaluation	2	11	36	42	8
12. Staff Development Activities	3	18	32	39	8
13. Teachers Knowledgeable	1	3	9	55	33
14. Ensure Quality	0	3	5	49	43
15. Effective Approaches	0	1	12	58	29
16. Strategies	0	1	11	52	36
17. Focus on Goals & Curriculum	1	3	21	54	21
18. Performance Assessments	1	7	19	57	16
19. Staff Development	5	21	31	35	8

N=417

Teacher views of their own professional development as reported in Table 56show that over half of the teachers have a personal, written development planand, in almost all cases, it is developed by the teachers themselves or inconsultation with a school administrator. The results also show that mostteachers collaborate in planning activities, delivering lessons, evaluatingstudents, and in school-wide instructional initiatives. Further, they attendconferences and take improvement courses. Few teachers have attended classes onteaching small classes, probably because classes with this focus may not beavailable

Table 56. Teachers’ Perceptions of Their Personal ProfessionalDevelopment (N=417)

Question #20 Over the past year, I have…	Percentage of responses
Engaged in a mentoring relationship with another teacher.	43.4
Participated in joint planning activities with other SAGE teachers.	93.8
Collaborated with other teachers in delivering lessons.	87.3
Collaborated with other teachers in evaluating student progress.	88.2
Participated in a study group or on-line network.	34.5
Collaborated in school-wide instructional initiatives or themes.	66.7
Collaborated with other schools or institutions.	32.9
Conducted research connected to my teaching.	32.4
Attended a professional conference or skill-building workshop.	81.8
Attended a workshop, seminar or retreat focused on diversity or human relations training.	26.9
Attended a workshop, seminar or retreat focused on teaching smaller classes.	12.5
Taken a course for graduate of CEU credit.	57.1
Question 21 Do you have a personal formal, written professional development plan?
Yes	42.9
No	54.7
Question 22 Which of the following statements most accurately reflects the content of your professional development plan?
It was determined primarily by me	32.4
It was determined in consultation with school administrators.	11.0
It was determined in consultations with district administrators.	3.1
It was determined primarily by school and /or district administrators	4.6

Family Involvement and LightedSchoolhouse

The extent to which SAGE school parents are involved in education of theirchildren is reported in Table 57. The results of the Teacher Questionnaire showthat teacher-parent contacts occur mostly through teacher notes, teacher andparent conversations, and telephone calls. The use of weekly progress reportsis increasing, however. Use of weekly progress reports in SAGE schools has morethan doubled compared to reported use in 1996-97 and in 1997-98.

Table 57. Teacher Questionnaire Results for Family Involvement(Percentages)

ITEM	1996-97 (N=212)	1997-98 (N=315)	1998-99 (N=417)
Class Newsletter	71	62	62
Weekly progress report - requiring parent signature	24	28	54
Weekly progress report - not requiring parent signature	11	12	50
Notes sent home	98	93	95
Conversations with parents	95	94	95
Parental visits to school	74	76	71
Telephone calls	92	89	91
Home visits	10	14	12

School-wide opportunitiesfor family involvement reported by principals are shown in Table 58. The1998-99 data on lighted schoolhouse activities were collected directly from theschool principals rather than through the Department of Public Instruction, ashad been the case in previous years. Data were reported for all 30 SAGEschools; however, since the same person is the principal of two schools only 29completed questionnaires provided data.

Table 58. SAGE Schools’ Lighted Schoolhouse Participation (N=29)

Activity	Number of Schools Reporting the Activity	Range of Participants in Each Activity
Child Care	11	15-80
Health Care	4	25-250
Breakfast	24	All students, daily
Tutoring	25	8-400
Homework Help	13	20-300
Extended Library	13	40-500
Adult Recreation	18	10-750
Girl and Boy Scouts	24	6-300
Music Lessons	14	3-100
Summer Reading	16	20-400
Head Start	5	40-80
Family Resource Center	6	100-200
Technology Education	6	10-354
GED Preparation	2	10-20
PTA/PTO	23	8-800
Family Literacy	8	15-500
Parent Advisory	12	4-300

Principals also reported a number of additional activities well attended bySAGE families, such as meal activities which varied from Sunday brunches, chilidinners, spaghetti dinners, holiday dinners to inviting SAGE parents once amonth for the routine student breakfasts; home visits for spring report cardconferences; and a variety of special activities such as family fun days,reading nights, career exploration days, and theatre productions.

DISCUSSION: MAJOR FINDINGS, LIMITATIONS, AND FUTURE REPORTS

TheStudent Achievement Guarantee in Education (SAGE) program is a statewide effortto increase the academic achievement of children living in poverty by reducingthe student-teacher ratio in kindergarten through third grade to 15:1. Schoolsparticipating in the SAGE program are also required to implement a rigorousacademic curriculum, provide before- and after-school activities for studentsand community members, and implement professional development andaccountability plans. The SAGE evaluation is being conducted under contractwith the Department of Public Instruction by the School of Education at theUniversity of Wisconsin–Milwaukee.

During the 1996–97 schoolyear SAGE was implemented in 30 schools located in 21 school districts. Itencompassed 84 kindergarten classrooms, 96 first grade classrooms, and 5 mixedgrade classrooms enrolling 1,715 kindergarten and 1,899 first grade students.In 1997-98 the SAGE evaluation added 113 second grade classrooms in theoriginal 30 SAGE schools. In 1998-99 the SAGE evaluation was made up of 131kindergarten, 143 first grade, 143 second grade and 139 third grade classroomsenrolling 2,303 kindergarten, 2,508 first grade, 2,493 second grade and 2,572third grade students.

To measure academicachievement, first-grade students in SAGE schools and in a group of comparisonschools were tested in October 1998 and again in May 1999 using theComprehensive Test of Basic Skills (CTBS) Complete Battery, Terra Nova edition,Level 10 (Fall) and Level 11 (Spring). Second-grade students were administeredLevel 12 in May 1999, and third grade students were administered Level 13 inMay 1999. Standardized tests are not administered to kindergarten students aspart of the SAGE evaluation because in the judgement of the evaluation teamstandardized tests at the kindergarten level are not an appropriate evaluationmeasure.

Major Findings

The Achievement Effect of ClassSize Reduction

First Grade

• As was found in 1997-98, test scores of SAGE and Comparison schools show statistically higher performance of SAGE students in language arts, math and total scores on the post-test in 1998-99 (Table 13).
• In 1997-98 and again in 1998-99, African American SAGE students scored lower on the CTBS pre-test than African American Comparison school students, but made significantly larger gains than Comparison school students from pre-test to post-test, surpassing African American Comparison school students on the post-test (Table 16).
• In both 1997-98 and 1998-99 African American students scored significantly lower than white students on the pre-test total scale score for both SAGE and Comparison schools. African American SAGE students achieved greater gains on the total scale score than white SAGE students from pre- to post-test, closing the achievement gap. At the same time, African Americans in Comparison schools achieved lesser gains, and the achievement gap with their white Comparison school classmates widened (Table 17).

Second Grade

• African American SAGE students scored significantly higher than African American Comparison school students on every sub-test and total scores in 1998-99 (Table 27).
• When looking at gains made fin 1998-99 from the first grade post-test to the second grade test, SAGE African American students made the same significant gains that the white SAGE students did, and did close the achievement gap between African American and white SAGE students although the relative gain was not significant (Table 28).

Third Grade

• SAGE students scored significantly higher in reading, language arts, math, and total score than Comparison students on the third grade test in 1998-99 (Table 34).
• In 1998-99, test results suggest that statistically significant positive effects of SAGE occurred in first grade, were maintained in second and third grade (Table 35).
• In 1998-99, African American SAGE students performed significantly higher on every sub-test and total score over African American Comparison students on the third grade test (Table 39).
• When second grade is used for a baseline score, African American SAGE students outperform African American Comparison students in reading, math, and total in 1998-99 (Table 39).
• African American students gained significantly more then SAGE white students in third grade, closing the achievement gap. Comparison school African American students did not gain significantly from second to third grade. In Comparison schools the gap between the performance of white and African American students widened (Table 40).

Additional Analyses

• Although they are not statistically significant, there are positive relationships between years of SAGE experience and student performance. This suggests that years of participation in SAGE may have a positive influence on achievement, although further research is necessary in this area (pp. 57-8).
• In all cases, classrooms with more affluent children from outperformed classrooms with children from poorer families (pp. 58-9).
• Classrooms with 30:2 student-teacher ratios achieved just as well as classrooms with 15:1 student-teacher ratios with the exception of language arts and mathematics sub-tests in second grade (pp. 59-60).
• In 1998-99 at the class level of analysis, smaller classrooms tended to score significantly higher in language arts, mathematics and reading, as well as total score after adjusting for individual pre-test results, socio-economic status and attendance. In other words, classrooms with fewer students are more likely to have higher class average achievement scores (Tables 18, 29, 30, 41, 42, 43).

• Twenty nine of the thirty top performing classrooms for which two years of data are available are SAGE classrooms (p. 61).

Reduced Class Size and Life inSAGE Classrooms

Data collected from case studies conducted in three selected SAGE schoolsduring 1998-99, from Teacher Questionnaires administered to all SAGE teachers,and from Principal Questionnaires completed by all SAGE principals confirm,clarify, and extend SAGE results from 1996-97 and 1997-98 regarding classroomevents. The major findings related to teaching a reduced size class are thefollowing:

• Individualization dominates SAGE classroom life.
• Individualization is made possible because having fewer students enables teachers to know students better, it reduces the need for discipline which results in more time for instruction, and it increases teacher enthusiasm for teaching.
• The individualization that occurs is process individualization rather than substance individualization. Teachers vary their instruction for individuals, but the curriculum is the same for all students.
• The effect of individualization is increased articulation and critique. In the reduced size class students more frequently reveal their present understandings and teachers more frequently correct misunderstandings.
• The main type of instruction is teacher centered (e.g., students listen, answer, practice, etc.) rather than student centered (e.g., problem solving, creative projects, etc.), but the use of hands-on activities is growing in frequency.
• A product of individualization in reduced size classes in addition to academic development is student independence, thinking, and responsibility.

Other SAGE Interventions

• The reading-language arts and mathematics curricula at all three grade levels are in compliance with national standards. The compliance, however, is slightly greater in reading-language arts than in mathematics.
• Each SAGE school has an established staff development program that produces changes in classroom practices. The professional development program is mostly or completely implemented in over 75% of the schools.
• Lighted schoolhouse activities are present in each SAGE school.

Limitations

When considering the results of the 1998-99 SAGE evaluation several factorsshould be kept in mind:

• The number of schools in the Comparison group pool was reduced from 17 in 1996-97 to 14 in 1997-98; one school converted from a Comparison school to a SAGE school, while two additional schools withdrew for other reasons. This problem was addressed for the 1998-99 academic year with the addition of three new Comparison schools. As a result, the composition of the Comparison group has become less demographically similar from the SAGE group.

• Third grade results may have been influenced by two factors. First, in 1996-97 a considerable number of SAGE first graders achieved perfect scores on the spring post-test. This had the potential effect of placing a "ceiling" on the gains reported for SAGE first grade students. Secondly, what was a "ceiling" in 1996-97 became a "floor" for the scores of this group of SAGE students in second grade. It is not possible to know to what extent this phenomenon had an impact on the 1998-99 SAGE third grade achievement results.

• The longitudinal results reported are based on regression equations using individual level data. Since SAGE students do not move from grade to grade as part of an intact cohort it was not possible to use HLM analyses to compare the performance of SAGE 1998-99 classrooms to that of SAGE classrooms preceding years.

Future SAGE Evaluation Reports

• Smaller classes in the SAGE program have a significant effect on student achievement in the first grade. The data from the 1996-97, 1997-98 and 1998-99 evaluations indicate that this finding is robust. The problems created by using the CTBS Complete Battery, Terra Nova edition, Level 10 as a first grade post-test during the 1996-97 school year made interpretation of the achievement results of that cohort of SAGE first grade students problematic as they moved through second and third grade. Much clearer results should emerge as the SAGE 1997-98 and 1998-99 first grade cohorts move through the grades.

• The impact on class achievement results of non-SAGE students entering second and third grade for the first time will be considered.

• Student achievement is consistently higher in some SAGE classrooms are more effective in increasing student achievement than others. An examination of the most successful SAGE classrooms will be conducted to determine instructional and other factors related to achievement.

References

Allen, R. (1997). Effectsof the revenue caps on Wisconsin’s school districts. 1996-97 school year fourthyear study. Madison, WI: Wisconsin Education Association Council.

Bingham, S. C. (1993). White-minority achievement gap reduction and smallclass size: A research and literature review. Nashville, TN: Center ofExcellence for Research and Policy on Basic Skills.

Bryk, A. & Raudenbush, S. (1992). Hierarchical Linear Models.Newbury Park, CA: SAGE Publications, Inc.

CTB/McGraw-Hill (1991). Comprehensive Test of Basic Skills 4^thEdition. CTB Macmillan/McGraw-Hill.

Glass, G. and Smith, M. (1978). Meta-Analysis of Research on Relationship ofClass-Size and Achievement. San Francisco, CA: Far West Laboratory forEducational Research and Development.

Mosteller, F. (1995). The Tennessee study of class size in the early schoolgrades. The Future of Children, 5, 113-127.

Pardini, P. (1998). Class wars. Middle Ground, 1(4), 22-24.

Pate-Bain, H., Achilles, C. M., Boyd-Zaharias, J., & McKenna, B. (1992).Class size does make a difference. Phi Delta Kappan, 74(3), 253-256.

Viadero, D. (1998). Small classes: Popular, but still unproven. EducationWeek. Available: http://www.edweek.org/ew/vol-17/23class.h17.

Wenglinsky, H. (1997). When money matters. Princeton, NJ: EducationalTesting Service.