Posted on 01/05/2003 8:29:30 PM PST by GailA
Eric Hanushek's "The Evidence on Class Size" is available in PDF format. To view this document, you must use Adobe Acrobat Reader 3.0. Click here to download this free software.
Below is a summary of the report. This summary is also available in PDF format, complete with charts and graphs.
The Evidence on Class Size Eric A. Hanushek University of Rochester
Executive Summary
A wave of enthusiasm for reducing class size is sweeping across the country. This move appears misguided. The primary conclusion of this paper is:
Existing evidence indicates that achievement for the typical student will be unaffected by instituting the types of class size reductions that have been recently proposed or undertaken. The most noticeable feature of policies to reduce overall class sizes will be a dramatic increase in the costs of schooling, an increase unaccompanied by achievement gains.
This conclusion is frequently greeted with surprise, but it should not be. Class sizes have been reduced over a long period of time with no evidence of overall achievement gains. Moreover, the effects of class size have been studied more intensively than any other aspect of schools, and this extensive research simply does not offer support for the types of policies to reduce class size that have been proposed. Broadly reducing class sizes is extraordinarily expensive and, based on years of research and experience, very ineffective.
There are powerful reasons for us as a Nation to expand and improve investment in human capital. The strength and vitality of our economy depends importantly on having a skilled workforce that can compete in the international economy. Acknowledging the need for investment does not, however, lead to unqualified support for any policies labeled “investment in our youth” or “school improvement.” Recent policy discussions have been laced with programs that fundamentally involve haphazard and ineffective spending on schools and that offer little hope for gains in achievement. The current set of class size proposals falls into this category.
People supporting broad class size reductions generally point to a few studies or a few experiences that suggest improved performance with smaller classes and then rely on policies to carry the day. A thorough review of the scientific evidence provides no support for broad programs of class size reduction.
1. We have extensive experience with class size reduction and it has NOT worked.
Between 1950 and 1995, pupil-teacher ratios fell by 35 percent. While we do not have information about student achievement for this entire period, the information that we have from 1970 for the National Assessment of Educational Progress (NAEP) indicates that our 17-year-olds were performing roughly the same in 1996 as in 1970. There are some differences by subject area, but the overall picture is one of stagnant performance.
The aggregate trends cannot be explained away by a worsening of students over time. While some family factors have worsened – increased child poverty and fewer two parent families, others have improved – more educated parents and smaller families. Nor does appeal to the mandated increases in budgets for special education explain the ineffectiveness of past reductions in class size and increases in spending. Eventhough special education is more expensive than regular education and even though it is an increasingly important issue, it simply is not large enough to rationalize past resource growth.
2. International experience suggests NO relationship between pupil-teacher ratios and student performance.
Dramatic differences in pupil-teacher ratios and in class sizes across countries are unrelated to measures of mathematics and science achievement. While there are many differences across countries that are difficult to adjust for in any analysis, there are also large differences in pupil-teacher ratios. These differences hold the possibility of understanding the effects of class size on performance, but – quite surprisingly – the international differences suggest a slight positive relationship between pupil-teacher ratios and student achievement.
3. Extensive econometric investigation show NO relationship between class size and student performance.
Extensive statistical investigation of the relationship between class size and student performance shows as many positive as negative estimates. With close to 300 separate estimates of the effect of class size, there is no reason to expect performance improvements from lowering class sizes. Moreover, because of the controversial nature of these conclusions, they have been carefully scrutinized – and the policy conclusions remain unaffected.
These studies are important because they provide detailed views of differences across classrooms – views that separate the influence of schools from that of family, peers, and other factors. As a group, they cover the influence of class size on a variety of student outcomes, on performance at different grades, and on achievement in different kinds of schools and different areas of the country. In sum, they provide broad and solid evidence.
4. Project STAR in Tennessee does NOT support overall reductions in class size except perhaps at kindergarten.
Much of the current enthusiasm for reductions in class size is supported by references to the results of a random-assignment experimental program in the State of Tennessee in the mid1980s. The common reference to this program, Project STAR, is an assertion that the positive results there justify a variety of overall reductions in class size.
The study is conceptually simple, even if some questions about its actual implementation remain. Students in the STAR experiment were randomly assigned to small classes (13-17 students) or large classes (21-25 students with or without aides). They were kept in these small or large classes from kindergarten through third grade, and their achievement was measured at the end of each year.
If smaller classes were valuable in each grade, the achievement gap would widen. It does not. In fact, the gap remains essentially unchanged through the sixth grade, even though the experimental students from the small classes return to larger classes for the fourth through sixth grades. The inescapable conclusion is that the smaller classes at best matter in kindergarten.
The STAR data suggest that perhaps achievement would improve if kindergarten classes were moved to sizes considerably below today’s average. The data do not suggest that improvements will result from class size reductions at later grades. Nor do they suggest that more modest reductions, say to 18 or 20 students per class, will yield achievement gains. The STAR evidence pertains to a one-third reduction in class sizes, a reduction approximately equal to the overall decline in pupil-teacher between 1950 and today.
5. The quality of the teacher is much more important than class size.
Considerable evidence shows that by far the largest differences in the impact of schools on student achievement relate to differences in the quality of teachers. Thus, whether or not large-scale reductions in class sizes help or hurt will depend mostly on whether or not any new teachers are better or worse than the existing teachers. Unfortunately, the current organization of schools and incentives to hire and retain teachers do little to ensure that the teacher force will improve. Simply grafting on different certification requirements are also unlikely to work. If we are to have a real impact on teaching, we must evaluate actual teaching performance and use such evaluations in school decisions. We cannot rely on requirements for entry, but must switch to using actual performance in the classroom.
6. While silver bullets do not exist, far superior approaches are available.
The states and federal government are in a unique position to initiate programs that promise true improvement in our schools. They are not programs that mandate or push local schools to adopt particular approaches – such as lowering overall class sizes or altering the certification of teachers. Instead they are programs that develop information about improved incentives in schools.
The largest impediment to any constructive change in schools is that nobody in today’s schools has much of an incentive to improve student performance. Careers simply are not made on the basis of student outcomes. The flow of resources is not related positively to performance – indeed it is more likely to be perversely related to performance. The unfortunate fact is, however, that we have little experience with alternative incentive structures.
A very productive use of state and federal funds would be to conduct a series of planned interventions that could be used to evaluate improvements. Minimally, instead of funding lowered class sizes everywhere, the states and federal government could team together to mandate more extensive random-assignment trials and evaluation of the benefits of lowered class sizes, à la Tennessee. More usefully, they could work to develop a series of experiments that investigates the construction and implementation of alternative incentive schemes – from merit pay to private contracting to wider choice of schools. Much of our knowledge about treatment therapies in medicine is directly related to prior experimentation. The last period of social experimentation by the federal government during the 1960s and 1970s produced many useful policy insights. A new program of trials with altered performance incentives could place an indelible positive stamp on the Nation’s future by committing to learning about how schools can be improved. Today we do not know enough to develop an effective program of improvement. Nor will continuation of past research programs help, because they must rely upon the existing structure of schools with the existing incentives (or lack of incentives).
Eric Hanushek is a Professor of Economics and Public Policy at the University of Rochester.
The Thomas B. Fordham Foundation 1627 K Street, NW, Suite 600 Washington, DC 20006 (202) 223-5452; (202) 223-9226 (fax) Questions? Comments? Email us: backtalk@edexcellence.net
To order publications see: http://www.edexcellence.net/fordham/foreports.html
The Foundation is neither connected with nor sponsored by Fordham University.
Slow local class size reductions have been shown to be able to be carried out with little negative effect - however rapid or large-scale reductions (Large school districts for example), create the effect of merely taking a portion of the class, and placing it in the care of a less-able teacher. This would seem obvious, given that there is not a huge surplus of wondrous teachers out there suitable for hire, and that one would assume that schools are already trying to hire the most able and qualified applicants within their criteria. This would suggest that the teachers brought in to decrease class size are quite likely to be less able.
It is all a scheme. They always find some so-called expert to twist the numbers and "prove" that class size must be smaller. Same with taxes, corporate welfare, crime & punishment. With government whores, facts never get in the way of their already decided answer.
Those nuns did one heck of a job. Even with a large amount of kids in the classroom, we learned how to read.
Amazing, but true.
Considerable evidence shows that by far the largest differences in the impact of schools on student achievement relate to differences in the quality of teachers. Thus, whether or not large-scale reductions in class sizes help or hurt will depend mostly on whether or not any new teachers are better or worse than the existing teachers. Unfortunately, the current organization of schools and incentives to hire and retain teachers do little to ensure that the teacher force will improve. Simply grafting on different certification requirements are also unlikely to work. If we are to have a real impact on teaching, we must evaluate actual teaching performance and use such evaluations in school decisions. We cannot rely on requirements for entry, but must switch to using actual performance in the classroom.
In New York, the worst schools are those serving predominantly black kids, and those are, not coincidentally, the ones with the worst discipline problems and the highest percentage of incompetent teachers. How do I know that? Because those schools' staffs ran off most or all of all the white teachers years ago, but at many of them, the majority of the teachers remaining have flunked the certification exam at least once. (IOW, the majority of their teaching staffs are uncertified.) And the cert exam gets dumbed down for black and Hispnic teachers on a yearly basis.
How do I know about the violence? In recent weeks in the New York media, waves of stories have described schools in The Bronx, Brooklyn, and Queens, in which students have graduated from daily assaults on classmates to daily assaults on teachers. The schools are predominantly black.
And I know of stories the media haven't told, for instance of an elementary school in which black kids as young as seven constantly break the bones of Asian classmates the same age. (Recently, a friend of mine broke the law, in helping to get a maimed Asian child transferred to another school. However, the school authorities are investigating -- not the attacks, but the transfer.) The black kids do as their parents teach them, and rarely does anything happen to them.
55 was typical in mine.
If a teacher signs a 5 year contract to work at an inner city school, the gov't will forgive a whole $5k in student loans. Big deal! Not enough incentive for me, I'm afraid.
Read the "Screwtape Letters" by C.S. Lewis, if you have not already read it. It is 60 years old but still (and always) very accurate. Yes, sometimes evil things do happen because someone wishes it to be so.
As homeschoolers, we've got two students per teacher. I'm seriously glad we decided to go that route...
Disclaimer: Opinions posted on Free Republic are those of the individual posters and do not necessarily represent the opinion of Free Republic or its management. All materials posted herein are protected by copyright law and the exemption for fair use of copyrighted works.