The content of the ASA's official journals reflects the research work of sociologists and the peer review and editorial processes of those journals. One of the points of contention raised in connection with the broader issue of journal diversity has been the content of the ASR relative to the research interests of the ASA's membership. In the spring and summer of 2002 Erik Olin Wright and Mike Hout undertook an analysis of the articles that appeared in the American Sociological Review and the American Journal of Sociology during the 1990s. The AJS is not an official publication of the ASA, but it is the other major general sociology journal that also has been the recipient of critiques similar to those that resulted in the formation of the Task Force. This summary represents a "borrowing" of some of the materials used. As part of an on-going project and with some financial assistance from ASA, this data collection effort provided an opportunity to contribute to the mission of the Task Force.

The team of faculty and graduate students from the Wisconsin and Berkeley Sociology Departments coded the contents of all articles published in ASR and AJS into categories that correspond to the fields that ASA members check off when they pay their annual dues. They coded each article as being about at least one but no more than three topics. The average was 2.1 for ASR articles; the median was 2 with a few more 3s than 1s. For AJS the mean number of topics was about 10 percent higher -- 2.3 -- but again the median was 2 (3s exceeded 1s by a larger margin in AJS, thus the difference in means). To be coded as being "about" a topic the substantive discussion had to address an issue at some length. Coders ignored mere "control variables" and the limit of three codes per article enforced that in practice. Most multivariate quantitative articles had far more than three independent variables; having to decide which 2 or 3 areas got tallied forced coders to focus attention on the central topic(s) of each article. A few discrepancies were difficult to code (e.g., coding the field "race, class, and gender" without also coding "ethnic, racial, and minority relations," "sex and gender," and "stratification and mobility"). Further, since ASA members have more than one interest; most journal articles address more than one substantive audience. But individuals' substantive interests are more complex than the average journal article, so the sum of members' areas of interest (365%, implying that members average 3.65 areas of interest) exceeds the sum of areas across articles (210% for ASR and 231% for AJS implying 2.10 and 2.31 areas per article in the two major journals). This creates a methodological problem for anyone who wants to compare the distribution of member interests with the distributions of ASR and AJS article contents. Because the number of interests per member exceeds the number of interests per articles, the average article "under-represents" the interests of the membership in an odd way. To eliminate this mathematical artifact from the analysis, they renormed the data.

The detailed comparisons between members' interests and journal contents are shown in Table 2 on the following page (PDF format). Of 70 areas tabulated by ASA, 37 areas of interest garnered at least 1 percent of the choices (after adjustment for multiple choices). The top five areas of interest are sex and gender, race and ethnicity, theory, social psychology, and family with about 4 percent of choices going to each. The 33 subject areas that had less than 1 percent each and the catch-all "other" aggregate to 15 percent choices (that works out to about one-half of a percentage point each). In terms of ASR publications, the top two areas - gender and race - are over-represented at 6 and 9 percent, respectively. Stratification, political sociology, economic sociology, development, organizations, and public policy are also more likely to show up in ASR than a random draw from members' interests would predict. The aggregate of small fields are, together, under-represented although it is hard to assess that outcome. These areas are small or overlap other areas sufficiently that remediating the discrepancy might well be counter-productive. The single area with the largest gap between members' interest and articles is medical sociology. Theory, methods, and teaching are also under-represented as topics in the pages of ASR¹. It is interesting that medical sociology, theory, methods, and teaching are each covered by an official ASA journal². From this observation it is possible that the presence of an official ASA journal channels good work in these fields to the relevant specialty journal. However, it may also be the case that reviewers respond more harshly on issues of general interest in these areas given that there is a specialty journal to which the authors can send rejected articles. Social psychology and education break that pattern; ASA publishes Social Psychology Quarterly and Sociology of Education, but these two areas are, nonetheless, proportionally represented in ASR. Figure 6 on page 23 (PDF format) displays a comparison between areas of interest and ASR articles.

The analysis also shows similarity and difference between the ASR and the AJS. As the third column of Table 2 and Figure 7 below show, AJS publishes somewhat different content than ASR does. The top two fields - gender and race - are less represented in AJS. Economic sociology and organizations are over-represented by a larger margin in AJS than in ASR. Theory and medical sociology, under-represented in ASR are proportionally represented in AJS. Education, quantitative methods, and qualitative methods are under-represented in AJS. This may reinforce the hunch that other ASA publications draw articles in these areas away from the ASR. However, it may also represent different approaches and strategies in these journals. For example, in the case of medical sociology, the difference may simply represent the AJS editor's decision to have a special issue (in 1992) on this topic.

ASA members' areas of interest are one valid basis of comparison for the content of ASR. But some topics - most notably teaching - may be of professional interest but not necessarily a research interest (of course, note the emerging subfield of the scholarship of teaching and learning which may be changing this). In any case, the areas of interest do not invite ASA members to make that distinction. The researchers sought out some other benchmark to reflect the distribution of ASA members' research interests. Though the team wanted to code all articles and books written by sociologists, this task far exceeded the resources available to the Task Force. Instead, doctoral dissertations, closer to the point of research production than publications, are free of any publication biases that might exist on the part of journal editors or book publishers. On the other hand, they are not all of equal quality. Even this scaled-back project exceeded the available resources. However, Wright was coding Madison dissertations for a presentation on American sociology he was preparing. With some assistance from the two departments and ASA, the team also coded the dissertations written at Berkeley. The Task Force considered but was unable to find the resources to expand this arduous effort. While a convenience sample, this provides first analysis of the issue of representation in the journals for new scholars.

Figure 8 on the following page indicates that the dissertations at Berkeley and Madison confirm the centrality of race and gender in recent research. Of dissertations written in the 1990s, 20 percent addressed gender issues and about 10 percent address racial issues. Contemporary dissertations at Madison also reflect the rise of economic sociology - so well represented in ASR and AJS but less well reflected in members' interests. One-third of Madison dissertation in the 1990s and one-tenth of Berkeley dissertations were in economic sociology. Berkeley dissertations on culture are the one example of work that is reflected in members' interests but not in the journals. One-third of Berkeley dissertations in the 1990s were about culture; only 5 percent of ASR articles were³.

Summary. Overall the ASR is more reflective of sociologists' interests than some of its critics suggest. Gender and race are more prevalent as ASR topics than as members' areas of interest, even though they are #1 and #2 among members' interests. Articles about theory and about methods are under-represented relative to members' interests in these fields. But that may not be a problem as (a) ASA has official journals addressed to these subjects and (b) nearly all articles in ASR and AJS address theory and all empirical articles, by definition, apply at least one method. Some surprises include the prevalence economic sociology along side the more established fields of political sociology and stratification. Articles that address public issues are over-represented relative to members' choices about areas. While few ASA members are likely indifferent to the public implications of their research, the norms are such that our primary interest in academic so checking off public policy is rare. Further, medical sociology, theory, and methods are under-represented in ASR but not in the AJS. There is some possibility that the existence of official ASA journals in these areas limits submissions and acceptances at ASR. It may also be the case that having these specialty journals shapes reviewers' evaluations of these areas as specialty areas, putting them at higher risk for rejection on those grounds. Further, there is some caution that the difference between AJS and ASR on the representation of medical sociology may represent a one-time, special issue on the topic in the AJS. Interestingly, this was a strategy mentioned by the editors in both the formal and informal interviews to help increase diversity and representation of sociologists' interests.

¹ Obviously, the vast majority of ASR articles use theory or quantitative methods and most use both.
² There was some disagreement among members of the Task Force and the Publications Committee about whether qualitative methods is covered by Sociological Methodology.
³ The adjusted percentage is 2 percent of ASR articles were about culture but for this comparison, the raw percentage is more like the comparison figure.