James Coleman, principal author of the landmark study, Equality of Educational Opportunity (1966), commonly referred to as the Coleman Report, is remembered as one of the twentieth century’s greatest research sociologists.1 Part of the reason this study brought him renown is that one of its findings—that black children learned somewhat more in predominantly white schools than comparable black children in predominantly black schools—was ideologically congenial to many American Sociological Association members and, more generally, to liberal proponents of busing as a means of improving the academic performance of black children. This finding became a rationale for busing black schoolchildren from the black neighborhoods, where they lived and attended predominantly black schools, to predominantly white neighborhoods, where white children would outnumber them.
Two other surprising findings of the study had at least as important implications for educational policy but received less media attention. The first was that school districts that spent larger amounts of money on each child did not have appreciably better educational outcomes than districts that spent less per child. Parental education and background seemingly had a greater effect on student achievement. The second finding was that the widespread assumption that black school districts were less well-financed than white school districts was mistaken; black and white districts received pretty much the same level of funding.
Coleman himself was politically liberal; he opposed racial segregation in public schools. Nevertheless, he conducted his sociological research to arrive at truthful descriptions of society, not to advance a political agenda. He supported busing because he thought that it was justified by the research of Equality of Educational Opportunity. In 1975, however, Coleman changed his mind about the usefulness of busing after his further studies showed that white families moved out of central city neighborhoods into the suburbs at an increased rate when busing was ordered. As a citizen, he still strongly opposed racially segregated schools, but as a scholar he now believed that busing led to more, not fewer, segregated schools.
As perhaps the most well-known research sociologist in the United States, Coleman became front-page news when he changed his view about the desirability of busing. Then, Alfred McClung Lee, president of the American Sociological Association (ASA), denounced Coleman to the governing council of the ASA, calling for the council to refer his recent revised view of busing to its Committee on Professional Ethics. The implication was that Coleman was expressing partisan prejudices, not drawing scientific conclusions based on his interpretation of objective data. This also became front-page news.
At that time I was a member of Social Democrats USA, the president of which was Bayard Rustin, who had been a close colleague of Martin Luther King Jr. and the organizer of the 1963 March on Washington to promote racial equality. I asked Rustin whether he would meet with Coleman with a view to issuing a news release defending the respectability of scientific research on the busing issue. Rustin agreed to do so, and Coleman agreed to meet with Rustin in New York before the ASA council vote took place. I was present at the cordial meeting that ensued. Rustin agreed to issue a statement, if it was needed, to deflect the impression that the civil rights community believed that Coleman was a racist. Rustin never needed to issue a statement, however, because Lee’s motion to ask the Committee on Professional Ethics to condemn Coleman received only two votes from members of the governing council: Lee’s and that of Pamela Roby, a specialist in feminist studies at the University of California, Santa Cruz. So the proposal died.
Now, in the year of its fiftieth anniversary, the Coleman Report continues to receive praise for its substantive findings and its methodological rigor—with one notable exception. This past spring Caroline Hoxby, a brilliant Stanford University econometrician, pointed to methodological errors in the report, specifically to “Coleman’s lack of differentiation between causality and correlation” in his findings, for example, that black children do better in integrated schools.2 Hoxby apparently thought that Coleman did not know that it takes more than correlation to establish a causal relationship between two variables. Having obtained her B.A. from Harvard in 1988, Hoxby was too young to have known Coleman personally or to know that for years he taught methodology courses to sociology graduate students at the University of Chicago; those courses undoubtedly stressed the need for additional evidence before inferring causation from correlation. Why then did he permit the report to suggest that integrated schools helped black youngsters achieve better academic results when they were in classes where most of the other students were white? For a simple reason: Researchers, politicians, and policy makers infer causation from correlation—provided they find a plausible explanation for the connection—because it would be too expensive and time-consuming to devise controlled experiments.
For example, the crime rate is higher in low-income neighborhoods than in wealthier suburban neighborhoods. From this finding criminologists and journalists infer that poverty somehow promotes criminality; no one demands a controlled experiment. In the case that Hoxby criticized, she faults Coleman for failing to consider another possibility: “While it is possible that the black children in integrated schools had higher achievement because their white classmates causally affected them, it is just as possible that the sort of black families who were motivated and able to live in integrated neighborhoods were advantaged in numerous hard-to-observe ways.”3 Hoxby is correct in her criticism; moreover, research methodology has advanced since Coleman’s time. However, she does not give Coleman credit for reversing his initial optimistic inference after new data showed that white flight undermined his belief that busing to promote integrated classrooms could improve the academic performance of black children. That reversal was what led Lee to accuse Coleman of racism.
In order to understand how and why the president of the ASA accused James Coleman of racism for concluding from his statistical studies that busing amplified white flight, one needs to know who Lee was, how he became ASA president, and how the ASA had evolved from an organization many of whose activist members believed that the central mission of sociology was not to describe how society works but to change it.
In the 1950s and early 1960s the ASA was a small professional organization with two classes of members: full members eligible to vote for officers, and associate members and graduate students in sociology, who could not. ASA annual meeting attendees went to plenary sessions at which distinguished sociologists such as James Coleman were invited to present their work to one or two hundred listeners. The ASA president was elected by the voting members, but candidates for the presidency were usually only those named by the Nominating Committee from among sociologists it considered distinguished, although nominations could be made by petition.
Alfred McClung Lee belonged to an activist group that described this system as elitist, undemocratic, and self-perpetuating. Its members recommended Lee to the Nominating Committee several times as a presidential candidate in ASA’s annual general election. The Nominating Committee never accepted these recommendations, probably because traditional scholarly sociologists regarded Lee as a heavyweight activist, not a heavyweight scholar. Lee and activist friends tried a different approach: nomination by petition. In 1974, Lee was put on the ballot along with two candidates offered by the Nominating Committee. The activists voted for him, and the traditional sociologists split their votes between the other candidates; Lee won the ASA presidency for the 1975–1976 academic year.
The election of Lee as president was one sign of the growing emphasis on activism in the ASA. Another was the amended ASA constitution permitting all members, not just established scholars, to vote. Finally, in the name of democratic reform, the annual meeting eradicated most of the plenary sessions people once attended in order to learn from professionals they admired. These were replaced by hundreds of short papers or poster sessions that very few people attended. The benefit of these hundreds of sessions was not intellectual—it was to enable graduate students and assistant professors to charge their expenses for attending the meeting to their universities.
In the latter part of the twentieth century, many young sociologists—and some, like Lee, who were not young—shared his assumption that racial prejudice underlay Coleman’s reversal of support for busing, so Lee had some support for attacking Coleman within the ASA. When Coleman was invited by the Program Committee of the Eastern Sociological Society to come to its 1975 meeting in New York, it became evident that the animus against him had spread and there was a plan to protest his appearance. I was at the meeting, and watched more than a dozen protesters standing in the back and side aisles holding large signs accusing Coleman of racism while he spoke. I do not remember whether any of them participated in the question and answer session.
Coleman was shaken by this experience as well as by the publicity produced by Lee’s recommendation to put Coleman’s reversal on busing before the ASA’s Committee on Professional Ethics. But these accusations did not succeed in driving Coleman out of the profession. Today they are largely forgotten. Coleman himself was elected president of the ASA in 1991. Lee died in 1992, just as Coleman was finishing his term. Coleman died shortly thereafter, in March 1995, prematurely succumbing to prostate cancer at age sixty-two. His research continues to be widely cited.
James Coleman was among the few twentieth-century research sociologists who dedicated large-scale research projects to describing how society is—not how they wanted it to be. Similarly, Coleman’s colleague, Harvard sociologist Samuel A. Stouffer, conducted surveys involving tens of thousands of research subjects to learn about the regularities of human interaction. For example, during World War II Stouffer directed surveys of half a million soldiers for the U.S. Army; after the war he distilled what he had learned about social behavior from these surveys in the two-volume monograph, The American Soldier, including the concept he developed to explain a baffling finding, “relative deprivation.”4 The title of his collected research papers, Social Research to Test Ideas (1962), explained what he and Coleman were trying to do: use empirical data scientifically.5
Coleman and Stouffer used enormous data sets, not just a few hundred cases or sometimes a mere handful. And they were meticulous in their research methods; if they used a sample to throw light on the behavior of much larger populations they weren’t satisfied with a response rate of 50 percent because they could not tell how and whether respondents who were queried and didn’t want to participate differed from those who agreed to participate. Contemporary social researchers who are trying to describe voting preferences or interest in a cereal or a hairspray may survey a large number of respondents, but they know that scientific rigor is not so crucial; less is at stake.
Few sociologists today embrace social research as zealously as Coleman and Stouffer did in the twentieth century, and more are attracted to progressive ideologies that seek to change the world. Sociological research scholars can still be found in the ranks of the ASA and the Sociological Research Association, a more traditional sister organization, but contemporary sociologists usually use more modest samples than Coleman and Stouffer did or base their research conclusions on data collected for different purposes and stored in university or governmental data banks. For example, using data from a large data bank, University of Texas at Austin sociologist Mark Regnerus conducted an interesting study of the social adjustment of the adult children of heterosexual and homosexual couples. This may be an excellent study, and Regnerus considered over three thousand responses—a goodly number—but it is unlike Coleman’s and Stouffer’s work because Regnerus did not collect the data himself; he used questionnaire items composed by other researchers that he cleverly drew from the data bank.
Paradoxically, although James Coleman is recognized as one of the twentieth century’s leading research sociologists, his greatest fame came from a finding that his later research led him to revise. He concluded in Equality of Educational Opportunity that the academic performance of black children was better in integrated classes than in all-black classes—and therefore supported busing to promote racially integrated schools. Coleman changed his mind when he found that busing, far from reducing racially segregated schools, made segregation worse. Coleman’s entire career is a testament to scholarly integrity.