Crowding Out

Small Classes Teach a Lesson in Unintended Consequences

California's massive effort to reduce the size of primary-grade classes--one of the largest educational reforms in U.S. history, costing up to $1.5 billion a year and involving 1.6 million students in the first two years alone--requires major adjustments to make sure the huge investment pays off, according to an early evaluation of the program.

On the positive side, California third graders in classes of 20 or fewer students have achieved slightly higher test scores than students in larger classes. The gains have been made equally by students of all ethnic, racial, economic, and linguistic backgrounds. Teachers report less time disciplining students and more time helping poor readers one-on-one and addressing students' personal concerns. Principals and superintendents report greater teacher enthusiasm for all reforms. And parents report more contact with teachers and greater satisfaction with their children's education.

On the negative side, the rapid implementation of the program, begun in 1996, has worsened the inequities among schools. The push to reduce class sizes has exacerbated space shortages at already overcrowded, inner-city schools, placing them at a triple disadvantage. First, it is tougher for overcrowded, land-restricted schools to add classrooms. Second, as a result, these schools earn fewer of the state dollars awarded for reducing class sizes. Third, to add classrooms and earn dollars, overcrowded schools are more likely than affluent schools to usurp space from libraries, computer labs, arts and music programs, child care, and special education.

Class-size reduction has also created 23,500 new teaching jobs--a 38 percent increase in the number of kindergarten through third-grade teachers in California. But many of the new teachers are minimally qualified, allowing the most qualified to pick and choose from the "most desirable" schools. Consequently, the better-qualified teachers often get "crowded out" of the inner-city, low-income, heavily minority, heavily immigrant schools where they are needed most.

"Overall, we're looking at class-size reduction with cautious optimism," says Brian Stecher, senior social scientist at RAND and director of the evaluation team, which includes staff from five research organizations. "We are finding some positive effects but also some troubling aspects that need to be addressed." To make the program work more equitably, as intended, the researchers propose several midcourse corrections:

  • Construct more classrooms.

  • Revise the formula by which the state reimburses school districts for reducing class sizes.

  • Allow schools greater flexibility in how they use these funds.

  • Improve teacher recruitment, preparation, and professional development.

  • Create incentives for good teachers to work at schools where their expertise is needed most.

  • California's experience has important implications for the nation as a whole. At least 19 states already have some kind of class-size reduction policy in place, and a proposed national program would face similar constraints as in California, according to a separate RAND study. Another related RAND study has concluded that the single most promising strategy to recruit and retain high-quality teachers in low-income, heavily minority areas is, simply, higher pay.

    Bricks and Teachers

    California schools have responded remarkably quickly to the incentives to reduce class sizes from an average of about 30 students to 20 or fewer students in kindergarten through third grade. The state legislature agreed to reimburse schools a flat rate of $650 for each child in a smaller class in 1996-97 and $800 per child in 1997-98. In just these two years, almost all first- and second-grade students and almost two-thirds of kindergarten and third-grade students moved into smaller classes.

    Even so, the reimbursements have not covered the costs in more than 40 percent of the school districts, especially in districts with larger proportions of students who are poor, Latino, black, or still learning English. These districts have either absorbed the extra costs themselves--often by taking money from other programs--or delayed implementation. With funding linked to implementation, proportionately more money has flowed to wealthier districts.

    The single greatest factor in the uneven implementation of class-size reduction has been the shortage of facilities for new classes. Despite state reimbursements and despite a state bond measure providing $700 million for construction and renovation for smaller class sizes, the shortfalls remain. Yet the need for space is not uniformly distributed. Not only did the districts start out with different space needs, but several districts have succeeded in getting their own local school construction bonds approved by voters. State officials need to get a better understanding of these variable space needs and target more money to add classroom space in overcrowded schools, according to the researchers. Otherwise, the students who are most in need academically will continue to be the least likely to attend smaller classes.

    Moreover, California should consider changing the flat-rate funding formula that now favors wealthier schools. Because the cost of reducing class sizes varies from school to school, state reimbursements could, for example, match the actual costs of implementation rather than meet just a flat rate. Alternatively, the state could shift to a formula that gives a disproportionate share of the money to districts with a disproportionate share of at-risk students. The existing formula has widened the resource gap between rich and poor districts. Reversing these unintended consequences, say the researchers, "may help ensure that the persistent achievement gap for many of the state's low-income and minority students will, as hoped, be narrowed rather than widened by this reform."

    Schools could also be given greater flexibility in how they use state funds to deal with classroom and teacher shortages. The state Legislative Analyst's Office has suggested that individual classes could be permitted to have as many as 22 students so long as the average class size in a district does not exceed 20.

    Another key factor in the uneven implementation of the program has been the quality of teaching. Traditionally, teachers working in low-income, inner-city areas have had somewhat lower qualifications than teachers in wealthier areas, but the gap widened tremendously between 1995 and 1998. Figure 1 shows that, prior to class-size reduction, schools with higher percentages of low-income students had slightly higher percentages of teachers without full credentials. But with class-size reduction, the gap increased almost tenfold, as schools with the highest percentages of low-income students hired much larger numbers of teachers without full credentials.

    By 1998, nearly 20 percent of teachers in the poorest areas lacked full credentials, compared with 4 percent of teachers in the wealthiest areas. Of the 23,500 new teachers hired in California between 1996 and 1998, about 10,100 of them--or nearly 43 percent--lacked full credentials. Some districts could reduce class sizes only if they hired teachers on emergency permits.

    To improve the quality of teaching, Governor Gray Davis has proposed a mentor teacher program and pressed for more slots for teacher preparation at California State University and the University of California. These efforts may be insufficient, say the researchers. It may be necessary to bolster local school-based training programs and to financially entice qualified teachers to the schools in greatest need. Experienced teachers probably gravitate toward schools that appear safer, serve students with fewer challenges, provide supportive parent involvement, and offer higher salaries. Thus, offering higher salaries and other benefits might be the only realistic way to lure teachers to cities and school districts with teacher shortages.

    Related RAND research reinforces this conclusion. Sheila Kirby, Scott Naftel, and Mark Berends examined teacher demand and supply in Texas in high-poverty districts that tend to have more students at risk of educational failure. The researchers found that teachers in these districts tend to display an overall greater sensitivity to pay and working conditions. "Raising teacher pay holds the most promise for reducing attrition," the researchers concluded. Raising salaries "would not only increase teacher supply in general, but may increase the supply of high-quality teachers," who are even more highly sensitive to pay and working conditions, given their competitive opportunities elsewhere.

    Dividends from Equity

    California's class-size reduction effort arose from several motivating factors, including financial opportunity, academic deficiency, and social equity. First, the state economy had gone from recession to boom by 1996, and a fixed percentage of the new surplus had to be spent on education, thanks to the voter-approved Proposition 98. Second, a decade-long decline in student achievement had reached the point of alarm in 1994, when the average reading scores of California fourth-graders tied for last place out of 39 states. Third, particular concern had been mounting over a persistent achievement gap: Black and Latino students in inner cities were performing at considerably lower levels than the rest of California students.

    Research had shown that smaller classes could help disadvantaged students the most. California officials pinned their hopes on the results of a class-size reduction experiment conducted in Tennessee from 1985 to 1990. That experiment produced relatively large achievement gains for all students based on standardized test scores. Even more impressive, low-income and black students posted gains almost twice as large as those of other students. These results made the Tennessee model especially attractive for California.

    But the programs in Tennessee and California have been vastly different. The Tennessee program was a small, strictly controlled experiment of fewer than 10,000 students, whereas the California program, with few guidelines, has involved 1.6 million. In Tennessee, schools could participate only if they had enough classrooms available. The project had no effect on teacher supply. Classes had only 15 students on average. And the diversity of students in Tennessee allows no comparison with the diversity of students in California.

    Nevertheless, California students in classes of reduced size have posted modest gains. Figure 2 shows that third graders in smaller classes scored somewhat higher in reading, mathematics, and language than third graders in larger classes. For example, 34 percent of students in smaller classes scored above the national median in reading last year, compared with 32 percent in larger classes. In mathematics, 38 percent in smaller classes exceeded the national average, compared with 35 percent in larger classes. And in language, 36 percent in smaller classes met the mark, compared with 33 percent in larger classes.

    Although these differences are small, they translate into an additional 6,000 students who now score above the national median in reading and 9,000 now above the national median in math, according to the research team. Furthermore, the achievement gains were made regardless of race, ethnicity, family income level, or English-language fluency. If anything, it is striking that the gains have been equally distributed across socioeconomic groups despite the unequal implementation across the state. Students in smaller classes in low-income areas have kept pace regardless of their relatively less-qualified teachers and larger cutbacks in other programs. If the inequities in funding and teacher quality can be eliminated, perhaps the achievement gains will ultimately prove to be the greatest among the most disadvantaged students, as in Tennessee.

    National Implications

    The California experience contains lessons for the nation as a whole. President Clinton has proposed reducing class sizes nationwide to 18 students in grades 1-3. Such a venture would require 100,000 new teachers and cost $5-6 billion a year, even without salary hikes or new classroom facilities, according to a RAND study led by Dominic Brewer. As in California, the rapid increase in the demand for teachers nationwide would reduce the average quality of teachers, thus reducing the overall benefits of the program.

    Rather than spending $5-6 billion a year for more teachers, the same amount of money could be used for many alternative purposes. The money would be enough, for example, to raise the salary of every existing teacher in grades 1-3, in every public school in America, by $10,000. Although the authors do not suggest such an across-the-board approach, they do argue that some combination of targeted salary hikes and improved professional development would most likely improve teaching in the nation's K-3 classrooms. Therefore, the researchers send a note of caution: "Policymakers may well wish to consider whether schoolchildren learn better in small classes with less-qualified teachers, or in larger classes with more-qualified teachers."

    Nonetheless, the California experiment with smaller classes has scored some clear successes, including modest improvements in student achievement despite widespread problems finding space and qualified teachers. The evaluators of the California program--at RAND, the American Institutes for Research, EdSource, Policy Analysis for California Education, and WestEd--will continue their evaluation for three more years. Meanwhile, there is no reason to hesitate fixing the problems already identified. The sooner that California can correct the flaws in the current program, the more likely that it will prove to be a source of celebration rather than consternation.

    Related Reading

    Class Size Reduction in California: Early Evaluation Findings, 1996-1998, George W. Bohrnstedt and Brian M. Stecher (eds.), Palo Alto, Calif.: American Institutes for Research, 1999. Also available as RAND/RP-803, no charge.

    "Supply and Demand of Minority Teachers in Texas: Problems and Prospects," Educational Evaluation and Policy Analysis, Vol. 21, No. 1, Spring 1999, pp. 47-66, Sheila Nataraj Kirby, Mark Berends, Scott Naftel.

    Staffing At-Risk Districts in Texas: Problems and Prospects, Sheila Nataraj Kirby, Scott Naftel, Mark Berends, RAND/MR-1083-EDU, 1999, 106 pp., $15.00.

    "Estimating the Cost of National Class Size Reductions Under Different Policy Alternatives," Educational Evaluation and Policy Analysis, Vol. 21, No. 2, Summer 1999, pp. 179-192, Dominic J. Brewer, Cathy Krop, Brian P. Gill, Robert Reichardt.


    Contents