ESSAYS ON ISSUES THE FEDERAL RESERVE BANK APRIL 2007 OF CHICAGO NUMBER 237a

Chicago Fed Letter

Chicago Workshop on Black–White Inequality: A summary by Derek A. Neal, professor of , , and consultant, Federal Reserve Bank of Chicago The Chicago Workshop on Black–White Inequality, funded by the Searle Freedom Trust, meets on a semiannual basis to explore the causes and consequences of economic inequality between blacks and whites in the U.S. On December 15, 2006, the second meeting of the workshop was hosted by the Federal Reserve Bank of Chicago.

During most of the twentieth century, looked at the effects of several govern- each successive generation of black ment policies on skill development in Americans came closer than their pre- children, as well as the measurement decessors to matching the educational challenges faced by social scientists who achievement and economic success of seek to quantify changes in the black– their white peers. However, the conver- white skill gap. gence in skills among children and in labor market success among adults stalled Education policy around 1990. The Chicago Workshop Education policy is at the forefront of Materials presented at the on Black–White Inequality is an effort many discussions of inequality in the to explore the reasons for the recent United States. The ongoing debate sur- December 15, 2006, meeting lack of progress for blacks relative to rounding the No Child Left Behind Act of the workshop are available whites. Workshop meetings focus par- of 2001 (NCLB) highlights the public at http://economics.uchicago. ticular attention on the black–white perception that disadvantaged commu- edu/Inequality_Workshop/ gaps in reading, math, and other basic nities, especially disadvantaged minority skills that appear to play such a large communities, receive poor service from Schedule_dec15.shtml. role in sustaining economic inequality. their public schools. Much recent re- Because the economy of the twenty- search on skill differences between black first century places a high premium on and white youth seeks to understand cognitive skills, black–white economic the relative importance of black–white inequality will persist as long as these differences in family resources and skill gaps persist. school quality as contributing factors to the skill gap. A first step in this re- Workshop meetings involve presentations, search is understanding how this skill prepared discussant remarks, and gen- gap changes as black and white children eral discussion among leading move through school. Numerous studies and other social scientists from around document that the black–white skill gap the country, as well as a core working is quite large when children enter school. group of six scholars, of whom I am one. If the gap widens as children age, we The other five are: , cannot be sure that schools are playing Steven Levitt, and Kerwin Charles from a large role because differences in family the University of Chicago; Greg Duncan resources may exacerbate the gap even from Northwestern University; and as black and white children progress Roland Fryer from . through the same school. However, if In this Chicago Fed Letter, I summarize the gap is stable as children progress the workshop’s second meeting, which through school, then differences in school quality are neither closing nor scores from each particular scale are the size of the black–white skill gap and widening the skill gap as children age. transformed into standardized units how it changes as students mature, but it While such a result would not imply that such that the mean score is zero and the also tackled how current policy decisions improving schools in minority neigh- overall standard deviation is one in each may affect the skills of disadvantaged chil- borhoods would do little to close the year. These results highlight the need dren, especially disadvantaged children black–white skill gap, it would imply that for better ways to describe the relative in minority communities. Congress will the current gap is not primarily the progress of black children. Because test take up the reauthorization of NCLB in result of any black–white disparities in scores have no natural units, Reardon’s 2007, and this debate is at the center of public school quality. presentation demonstrated that some current policy discussions concerning claims about the gap between black and strategies for improving the performance Measuring progress white children in terms of basic skills of public schools in disadvantaged com- One might assume that documenting could rest on arbitrary choices of the munities. Two cornerstones of NCLB are the facts about the relative progress of scale used to report the test scores. He that schools should be judged by student black children during school is a fairly did not have access to either the Texas performance on standardized tests and straightforward task. Black children or North Carolina data and thus could that students should have the option to either fall further behind their white not be sure that scaling choices played choose a different public school if the test peers as they progress through school a role in creating different results for the scores in their current school are consis- or not. However, the existing literature two states. However, his analyses raise tently below state proficiency standards. on this topic does not provide clear a broad methodological challenge for Two presentations examined the likely answers, and the first session of the education research. Researchers must consequences of these policy priorities. workshop meeting demonstrated that be able to discuss the size of skill gaps The first by Justine Hastings, Yale Uni- this issue is more perplexing than some in terms of a meaningful metric. In re- versity, and Douglas Staiger, Dartmouth might imagine. Scholars made presen- cent research, working group member College, used data from a public school tations based on administrative data James Heckman has argued for methods choice experiment in North Carolina from different states. Eric Hanushek, that seek to express skill gaps in terms to examine the determinants of school Stanford University, and Steven Rivkin, of the expected difference in potential choice decisions. They reported that public school choice programs are often most valuable for families that are white Because test scores have no natural units, it has been demon- and economically advantaged because strated that some claims about the gap between black and the best public schools tend to be pre- white children in terms of basic skills could rest on arbitrary dominately white and geographically distant from disadvantaged neighbor- choices of the scale used to report test scores. hoods. During the initial years of NCLB, there have been persistent problems Amherst College, used administrative adult earnings available to people of with students not being able to exercise data to show that in Texas public schools, different skill levels. In my capacity as their right to public school choice be- the standardized achievement gap be- workshop director, I have argued that cause the best public schools often face tween black and white students grows social scientists should focus more on excess demand. The work presented by substantially during elementary school. measures of skill gaps that describe how Hastings and Staiger suggests that, even In contrast, Charles Clotfelter, Duke members of one group rank in the skill if the best public schools are forced to University, presented results using data distribution of the other group. The expand to take in students from failing from North Carolina that suggested amount of overlap between two distri- schools, these high-performing schools the achievement gap between black butions of scores focuses on how a set will often be located far away from dis- and white children is large but constant of test scores ranks students from the advantaged communities, making it dif- between third grade and eighth grade. most to least proficient. Descriptions of ficult for students from such communities relative ranks and the overlap between to change schools. This feature of NCLB The subsequent discussion of the two the distributions of scores for two groups stands in stark contrast to voucher plans presentations highlighted many factors provide important information concern- or proposals to expand the number of that complicate these calculations, but ing how two different groups perform charter schools that are allowed. Both Sean Reardon, Stanford University, made with respect to a particular skill, and voucher plans and the expansion of the most telling point. Reardon dem- this information does not rely on any charters provide mechanisms for new onstrated that conclusions concerning arbitrary notion of scale for the scores. schools to open in communities with whether or not black children are fall- schools that are currently failing to meet ing further behind white children as they Recommendations for future policy the NCLB standards for test score perfor- progress through school can be greatly The workshop’s second meeting dealt mance. Nonetheless, Hanushek correctly affected by the scale used to measure not only with basic issues concerning pointed out that NCLB does expand test scores. Further, this is true even when resources available to disadvantaged However, once this announcement is of such a system will not be easy because, students because the act only man- made, failing schools have the incen- as Reardon pointed out earlier, measures dates public school choice for students tive and opportunity to coach students of how much student achievement is who attend persistently failing schools. in ways that may inflate their scores on growing over time rely on somewhat these particular assessments. Once a arbitrary choices concerning the scale Next, Diane Whitmore Schanzenbach, situation is created where some schools used to report test scores. Still, the bot- University of Chicago, and I described coach their students more than others tom line is that if the reauthorization specific aspects of the NCLB testing for the state assessments, the results of of NCLB does not include provisions program that may be problematic. Using that reward schools for all the progress that students make, the law will continue to provide strong incentives over the The children of women who work more and earn more as part short term for some schools to ignore of their participation in welfare-to-work experiments enjoy note- their most disadvantaged students. worthy gains in academic achievement, compared with the Impact of work incentive programs children of women who were randomly denied access to the The December 15, 2006, gathering was same experimental programs. the second meeting of the Chicago Workshop on Black–White Inequality, and it involved one presentation that data from a period of high stakes testing these assessments will provide less valu- echoed a theme from the inaugural in Chicago Public Schools that predates able information concerning the actual meeting on April 21, 2006. This theme the NCLB, we showed that the relative relative performance of students from was the impact of the earned income performance of elementary school stu- various schools. tax credit (EITC) and certain welfare- dents on the tests used to generate ac- Further, Schanzenbach and I also dis- to-work programs that offer single moth- countability measures greatly overstates cussed concerns about the NCLB pro- ers the opportunity to increase their their future relative performance in high visions for scoring schools based on the family incomes if they work more. At our school. However, this was not true in test results of their students. NCLB first meeting, Gordon Dahl, University the period before the tests became part dictates that schools are judged solely of California, San Diego, and Lance of the accountability system. When by the fractions of students in different Lochner, University of Western Ontario, these assessments became high stakes groups who achieve a certain proficiency discussed their finding that increases in exams, failing schools faced strong in- level. A given school receives no imme- EITC benefits appear to generate more centives to coach students for the exams; diate reward for greatly improving the thus, efforts to game the system may have achievement of students who are far resulted in test scores that provided less below grade level unless it brings them reliable information concerning the Michael H. Moskow, President; Charles L. Evans, all the way up to the state’s standard for Senior Vice President and Director of Research; Douglas relative performance of schools. proficiency. Thus, in the short term, Evanoff, Vice President, financial studies; Jonas Fisher, Economic Advisor and Team Leader, macroeconomic NCLB outlines three roles for standard- schools may have incentive to simply policy research; Richard Porter, Vice President, payment ized tests. First, tests should be aligned ignore students who are truly disadvan- studies; Daniel Sullivan, Vice President, microeconomic policy research; William Testa, Vice President, regional with state-specific curriculum standards taged. The Chicago Public Schools programs and Economics Editor; Helen O’D. Koshy, so that tests cover what teachers are sup- adopted a testing system in the 1990s Kathryn Moran, and Han Y. Choi, Editors; Rita posed to be teaching. Second, schools that also had this feature in some grades Molloy and Julia Baker, Production Editors. should be held accountable for the test with respect to reading. In our research, Chicago Fed Letter is published monthly by the Research Department of the Federal Reserve results of their students, and principals we have found that during the 1990s, Bank of Chicago. The views expressed are the and teachers in failing schools should in Chicago, students at the bottom of authors’ and are not necessarily those of the Federal Reserve Bank of Chicago or the Federal improve or be replaced. Third, states the achievement distribution did not Reserve System. should publish report cards that provide experience the type of gains that other © 2007 Federal Reserve Bank of Chicago public information about the perfor- students experienced, and there is some Chicago Fed Letter articles may be reproduced in mance of their schools on standardized evidence that they may have done worse whole or in part, provided the articles are not reproduced or distributed for commercial gain assessments. At some level, it may be too after the reform began. NCLB is likely and provided the source is appropriately credited. much to rely on one set of exams for to serve the most disadvantaged students Prior written permission must be obtained for any other reproduction, distribution, republica- all three of these purposes. The desire poorly unless the reauthorization in- tion, or creation of derivative works of Chicago Fed for alignment with curriculum standards cludes a provision that rewards schools Letter articles. To request permission, please contact leads states to announce that specific for improving the performance of each Helen Koshy, senior editor, at 312-322-5830 or email [email protected]. Chicago Fed tests designed by a specific company child, even if these students do not make Letter and other Bank publications are available or group of researchers will be the it all the way to their state’s proficiency on the Bank’s website at www.chicagofed.org. standard for assessment in each year. standard. However, the implementation ISSN 0895-0164 work, higher family incomes, and higher factors. In any case, the results suggest evidence that black children begin school achievement among children. At our that when disadvantaged parents ben- well behind their white peers and that second meeting in December, Greg efit from stronger incentives to earn the home environments of black and Duncan presented work with several income, their children also often ben- white children differ greatly in many co-authors that found similar patterns efit. Duncan’s work is part of a grow- dimensions that affect children’s cogni- in welfare-to-work experiments. The ing literature that highlights the role tive development. Future workshop children of women who work more and of early childhood environments as meetings will devote more attention to earn more as part of their participation determinants of skill development. the links between black–white differences in welfare-to-work experiments enjoy in economic resources and black–white noteworthy gains in academic achieve- Conclusion differences in home environments for ment, compared with the children of The discussions at the most recent meet- children. Existing studies show that black women who were randomly denied ac- ing of the Chicago Workshop on Black– children live in home environments that cess to the same experimental programs. White Inequality confirmed the need are quite different from those of their At this point, it is not clear what accounts for further research into differences in white peers, and these differences appear for this result. We are not sure whether achievement between black and white to matter for cognitive development. it reflects the effects of higher family students, because the skill gap undoubt- However, more research is needed be- income per se, changes in family atti- edly contributes to income inequality lat- fore we can fully understand how the tudes and aspirations associated with a er in life. While there is no consensus economic prospects available to black mother working, changes in child care concerning whether or not black–white adults shape family structures, parenting arrangements associated with a mother differences in school quality contribute styles, and the allocation of resources working, or some combination of these greatly to skill differences, there is clear to children’s development.