A r c h i v e d  I n f o r m a t i o n

Class Size and Students At Risk: What is Known?...What is Next? - April 1998

Research Priorities:
Five Issues in Need of Further Research

The past decade of research on class size has opened exciting possibilities for improving the performance of students in the elementary grades and, hopefully, in later grades as well. Recognizing the potential of small classes, a number of states are beginning initiatives to reduce class size in some or all districts. At the same time, important questions remain unanswered, especially with respect to students at risk; the most pressing of these are outlined here.

Many of the issues raised in this research agenda can be partially addressed through the use of existing data. Using extant data bases offers unique opportunities and considerable economy. The process is not intrusive and shortens the time required to provide information substantially (Cooley & Bickel, 1986). In this instance, the data base that can be assembled from STAR and related studies is of unusual scope and quality. Mosteller (1995) noted:

Because a controlled education experiment (as distinct from a sample survey) of this quality, magnitude, and duration is a rarity, it is important that both educators and policy makers have access to its statistical information and understand its implications. Thought should be given to making sure that this information is preserved and well documented and that access to it is encouraged. (p. 126) As of this writing, the STAR/LBS data base consists of the original kindergarten through grade 3 data on approximately 7,100 children each year,20 including class placement, demographic information, and achievement scores obtained annually. The LBS follow-up data include achievement tests through grade 9, ratings on the Student Participation Questionnaire in grades 4 and 8, and student responses to the "Identification with School" scale, administered in grade 8. Other STAR data have been collected but are not computerized; these include teacher exit interviews, teacher and teacher-aide time logs, data on a matched sample of comparison schools that did not participate in the small-class experiment, and observations of a sample of grade 1 teachers conducted the year prior to teaching a small class and again during the small-class year. If STAR/LBS data were made available to the research community, their analysis could prove invaluable.

At the same time, new data may be preferable for answering some questions and may be the only way to obtain definitive answers to others. In order to obtain answers, it is important that districts and states undertaking small-class initiatives systematically collect information before, during and after implementation. Not only will problems and successes associated with small-class initiatives be documented, but there is much to be learned of interest to educators generally.

Issue 1: Short-Term and Long-Range Effects of Small Classes
for Students At Risk

Additional research on the effects of small class size, focused on at-risk students is needed to answer questions such as those discussed below.
What are the "true" immediate and continued effects of small classes on the achievement of students at risk?

Immediate outcomes. Project STAR found significant academic benefits for pupils enrolled in small classes. However, as many as 18 percent of the classes drifted out of the ranges defined as "small" or "regular" during the 4-year study when students transferred into or out of participating schools. The Burke County study and Success Starts Small found small-class effect sizes as large as 0.4sigma and greater. An examination of just those STAR classes that remained in-range may yield effects substantially larger than those in Table 1.

A focused analysis of STAR data also could ask whether small classes reduce the achievement gap between minority students or students from low-income homes, and their non-minority peers. Since some students were only in small classes for 1, 2, or 3 years, the reanalysis could also reveal the benefits of small classes to students who are more mobile than others--an issue of particular importance to students at risk.

Medium-term outcomes. The LBS documented a continued but diminishing impact of small classes over subsequent years (grades 4 through 9). Again, analyses have not focused on the at-risk population and did not examine the achievement gap between white and minority students.

Long-term outcomes. Further data are needed to address the effects on pupil performance through high school.

What are the effects of small classes on non-achievement outcomes among students at risk?

Several studies (e.g., Rand study, Head Start, Project High Scope) suggest that the benefits of some early interventions persevere through and beyond the school years. The LBS documented improved classroom behavior in grade 4 but went no further. Of the negative events experienced disproportionately by students at risk, it is important to ask whether small classes reduce the need for disciplinary action, for special education placement, for in-grade retention, and increase the likelihood of a student graduating from high school.

Issue 2: Teaching Practices to Maximize
the Effectiveness of Small Classes

Studies to date suggest that small classes create a more personalized environment for teacher and students and that small classes produce a time efficiency by reducing the need for discipline and classroom management and delivering effective instruction in a shorter amount of time. Additional research is needed to answer subsequent questions such as those discussed below.
How do the most effective teachers take advantage of a small class setting to deliver more individualized instruction to pupils? and How can other teachers be taught to use these strategies?

Some teachers may use techniques designed to increase the participation of each individual student in classroom interactions. This is important in light of some youngsters' tendency to withdraw from participation--a particularly debilitating strategy.21 Some teachers may be able to increase parents' involvement in their youngsters' schooling. And some may be available to provide extraordinary support (e.g., extra attention; after-school help) for students having difficulty with class material; these "extras" are often lacking in schools serving students at risk (Ralph, 1989).

How do the most effective teachers take advantage of the time-efficiency provided by small-class instruction?

How do teachers in small classes allocate their time to working with individual students, small groups, or the whole class? What kinds of activities can be undertaken when instruction is more efficient? For example, if course material is reinforced, are additional activities implemented to push the students beyond the usual content? Is more focused evaluation and feedback provided?22

Issue 3: School and Classroom Conditions
That Interact With Class Size

STAR findings showed a disproportionate impact on minority students in some achievement areas each year (kindergarten through grade 3). Further research should examine other characteristics of schools and programs that may interact with class size and address questions such as those discussed below.
Can small classes offset some of the disadvantages of attending a large school?

Past research has documented that attendance and participation in academic extracurricular activities are inversely related to school size, that is, larger schools have decreased student participation (Lindsay, 1982; Cockman, Bryson, & Achilles, 1989; Fowler, 1992). There is also a carryover effect: high participants in high school tend to participate actively in post-schooling cultural and community activities (Lindsay, 1984). Most of this research involved high-school students. The mechanisms that explain the association of school size with student participation have not been uncovered, but results indicate that smaller schools are seen as "warmer" and more supportive settings (Finn & Voelkl, 1993);23 that is, they provide a more personalized environment.

Given that large schools are ubiquitous, this research raises questions about the potential benefits of small classes. One study using STAR data (Nye, 1995) concluded that the negative correlation between school size and achievement disappears for students attending small classes. Other questions yet to be addressed include: Does attending a small class--even in the earlier grades--produce higher student attendance and involvement in later grades independently of the size of the school? If so, is this associated with improved student performance and increased likelihood of graduating from high school? Is there an interaction of class size and school size in the elementary grades as well? Is the increased engagement associated with small classes beneficial particularly to students at risk attending large, perhaps more impersonal, schools?

Do small classes accentuate and extend the benefits of other early childhood programs and practices?
To date, no analyses have examined the combined impact of small classes with federal, state, or local programs directed at students living in poverty or who are otherwise at risk for school failure (e.g., Title I). Preschool participation and attending full-day kindergarten may also.promote the development of children generally and students at risk in particular. Some states do not have state-mandated kindergarten and, in others, half-day kindergarten is common.24 An analysis of some of the STAR data (Achilles, Nye, & Bain, 1994 1995) indicated a significant "test score value" for children who attended kindergarten. Further work is needed to document the combined impacts of preschool participation, attending kindergarten, and being enrolled in a small class on students at risk. Both short-term and long-range outcomes should be examined.
Do small classes accentuate and extend the benefits of other classroom practices?

Several examples illustrate this research question:

Cooperative learning has been used to promote the achievement of all students but students at risk in particular. Are cooperative learning techniques less effective or equally effective if the class size is small, or are the benefits accentuated?

Heterogeneous small groups and heterogeneous classes have been found to be academically beneficial to at least some students. In a review of the problem of "stratification" in heterogeneous classrooms, Cohen and Lotan (1995) noted that, with appropriate intervention, higher rates of participation can be encouraged among low-status students. No investigation has examined the interaction of class size with class heterogeneity by racial-ethnic, socioeconomic, or primary language characteristics. There are many possible avenues to explore.

Teacher aides are a major education intervention (e.g., Title I; special education; some remedial programs). The academic value of teacher aides depends both on their qualifications to provide instruction and on how they are deployed (e.g., for order-keeping, for bookkeeping, or as a true teaching resource). Research should ask whether teacher aides can be utilized to further enhance the benefits of small classes, or whether judicious use of well-prepared teacher aides in regular-size classrooms can produce some of the same benefits as small classes, but at lower cost.

Issue 4: Small Classes and Positive "Long-Term Trajectories"
for Students At Risk

Educational risk may be described in terms of group status characteristics or in terms of a set of behaviors. If these "risk" behaviors manifest in negative ways, such as not attending to the teacher, not completing required work, and skipping school, they create impediments to learning. On the other hand, if a student exhibits positive behaviors, such as attending and participating (e.g., "engagement behaviors"), the behaviors may serve as "protective" mechanisms to improve the chances of school success in spite of group risk status. While status and behavioral risk factors are often found in the same individuals, risk behaviors may be amenable to influence by parents, school personnel, and school programs.

There is evidence that risk behavior in school and the classroom and its obverse, engagement, is developmental and begins in the early school grades (see Finn, 1989, 1993). Active participation in the early grades, accompanied by some degree of academic success, serves to perpetuate continued participation throughout the school years; this would be a "positive trajectory." When a young student does not participate in the classroom, this may begin a cycle that results in adverse consequences over time. Barriers to success multiply. Risk factors "cluster;" that is, multiple risk factors are likely to occur in the same individual--especially over time. And risk factors "track;" that is, they have early forms that evolve into fully developed forms over time that are increasingly difficult to alter. Thus it is essential that educators identify and understand forms of disengagement from school in the early grades and do all that is feasible to intervene at that point. The central question then is:

Can small classes in the early grades begin students on a positive trajectory that persists through the school years?

Three key issues should be explored further. First, we need to assess the short-run and long-run likelihood of adverse consequences of early risk behavior. The relationship of status and behavioral risk factors in the early grades with absenteeism, suspensions, retention in grade, loss of identification with school and dropping out, and even drug use and contacts with police in later years should be studied carefully. Patterns of tracking and clustering of risk factors should also be documented.

Second, we need to understand why some students at risk succeed academically in spite of the obstacles they may face because of group status characteristics. Such students have been termed "educationally resilient" (see Nettles & Pleck, 1994; Rutter, 1990). With respect to resilient students we should ask whether they exhibit positive engagement behaviors beginning in the early grades. What sorts of preschool and early school experiences did they participate in? What sorts of support for learning did they receive from their teachers, parents, peers, and others?

Finally, we need to ask whether small classes in the early grades interrupt patterns of disengagement, decrease the likelihood of adverse consequences, and increase the likelihood of positive behaviors (and achievement) over subsequent years.

Issue 5: Assessing the Costs of a Small-Class Initiative

Every school, district or state planning to undertake a class-size initiative confronts the budget question. However, the question is not as simple as asking "How much more will additional teachers and classrooms cost?" because associated benefits may produce savings and careful planning may be able to contain the expense. There are a number of related issues, elaborated below, about which there is a small but growing base of knowledge. Additional research is needed to address these questions more fully.
Do classes of 15-18 pupils really cost more if weighed against the benefits that accrue?

Researchers have not yet assessed the total impact of small classes, but research has demonstrated academic benefits in all subjects that persist into later grades, and improved learning behavior at least through grade 4. Related studies previously discussed have indicated fewer grade retentions and fewer disciplinary referrals. If, in the long run, the need for remedial and special education teachers is reduced, discipline problems and violence are reduced, and/or fewer students leave school without graduating, then there is a real gain on the output side of the equation. Most of these effects are well-documented while some require further research. It is clear, however, that small classes produce an array of academic and behavioral benefits that have cost-savings value.

How can the costs of implementing small classes be contained?

If hiring more teachers is the only strategy used to reduce class size, a small-class initiative undoubtedly will be expensive. Again, however, it may not be expensive in relation to the benefits that accrue or in comparison to other interventions with an equally broad array of outcomes. Although, at present, there are no prescribed solutions to the issue of cost, a number of districts have found ways to achieve small classes, even within the usual per-pupil expenditures. Some schools have experimented with creative scheduling plans. Others have redeployed staff in order to achieve smaller class sizes; for example, by assigning Title I teachers or specialty teachers to small classes, using supplemental state funds for additional teachers, or allocating part-time teacher aide funds to full-time teaching positions (see also Miles, 1995).

While reassignments such as these do challenge people's thinking about "business as usual," initial reactions from these sites indicate that both teachers and administrators are satisfied with the decisions. However, the experiences of these schools and districts must be systematically documented in order for us to obtain further answers to the question of how costs can be contained. Additional field-based research is needed urgently to build a broader knowledge base that educators can use for decision making. Further, a mechanism is needed for compiling the experiences of local sites into a central database that can be tapped by researchers and policymakers alike.

20 The total sample size (approximately 12,000) exceed the number of students for any given year, since it includes pupils who were in STAR classes for 1 or 2 years but not other years.

21 Brophy and Rohrkemper (1989) have produced one of the few lists of strategies to encourage participation by students who are shy and/or withdrawn.

22 Both observational and interview data are needed to address these questions. Since small classes are being implemented in many states and districts across the country, further observational data should be readily accessible.

23 Teachers and administrators may actually behave in warmer and more supportive ways in smaller schools; that is, there may be a factual basis to this perception. This has not been studied.

24 At the time Project STAR began, Tennessee did not require that children attend kindergarten.

-###-

[Instructional Practice and Student Behavior] [Table of Contents] [References]