Expanding the Scope of Research Productivity Indices: a Commentary on Dixon, Reed, Smith, Belisle, and Jackson (2015)

Abstract

Dixon et al. (Behavior Analysis in Practice, 8:7–15, 2015) outlined methods to evaluate research rankings of behavior analytic graduate training programs and their graduate faculty. Publications across five behavior analytic journals were evaluated and cumulative totals were obtained for each program and its faculty members. Critchfield (Behavior Analysis in Practice, 8:3–6, 2015) noted that these measures indicated that many graduate students in behavior analysis are receiving training by faculty members with limited research productivity. This commentary offers some suggestions and refinements of the procedures used by Dixon et al. that might provide a more sensitive measure of scholarly productivity in our field. The implications of these suggestions are discussed.

In their recent paper, Dixon et al. (2015) outlined methods to evaluate research rankings of behavior analytic graduate training programs, specifically those that provide Board Certified Behavior Analyst (BCBA) approved course sequences and field experiences, along with the faculty in those programs. As the authors note, the substantial growth of the field of behavior analysis in recent years, particularly with regard to the increase in graduate training programs, requires indices of quality so that consumers (potential graduate students and those who hire them as professionals) can differentiate among the different graduate programs. Based on their results, Dixon et al. reported that in only approximately 50 % of BCBA programs have the faculty members published ten or more articles in what the authors refer to as “our field’s primary empirical journals” (i.e., Journal of Applied Behavior Analysis, Journal of the Experimental Analysis of Behavior, The Behavior Analyst, Behavior Analysis in Practice, The Psychological Record, and The Analysis of Verbal Behavior). In addition, many of the faculty in the programs were found to have no publications at all in these journals.

At face value, the research indices reported by Dixon et al. are alarming. One possibility of these data is that the value of research activity in our field is decreasing relative to practitioner activities. Moore and Cooper (2003) conceptualized behavior analysis as a continuum with basic research at one end, applied research in the middle, and service delivery at the other end. It is possible that demand for service delivery size has shifted the emphasis of graduate training programs. Alternatively, it is possible that research competencies in some of the programs are lacking and that students are subsequently not receiving adequate mentorship in scholarly pursuits.

Although I applaud the establishment of quality indices for our field and appreciate the efforts of Dixon et al., certain refinements to their methodology may lead to a more accurate “big picture” with regards to scholarly productivity, thereby increasing the utility of such indices. First, a more comprehensive measure of research productivity of each program and faculty member might have been achieved had Dixon et al. included additional journals that publish behavior analytic research in their analysis. Carr and Briggs (2010) provide a reasonable list of over 20 such journals. Although it is certainly the case that particular journals in our field are ranked more favorably than others with regards to scientific rigor, a perusal of the table of contents in the journals listed by Carr and Briggs shows numerous publications by highly respected members of our field. An argument can also be made that we should include publications in journals outside of behavior analysis when we consider research productivity indices. For example, Friman (2010) has challenged behavior analysts to increase our “mainstream relevance” by disseminating behavior analytic research outside of our own journals. In addition, Vyse (2013) noted that Herrnstein and his colleagues published papers on delay discounting in prominent mainstream psychology journals and equally prominent economics journals, thereby reaching a broader audience for their work.

Second, the use of cumulative publications as a measure of research productivity is not a sensitive enough measure, and, as Critchfield (2015) noted, it biases research productivity measures toward programs that have existed for longer periods of time. Rather, it may be better to use an annual rate of publications by programs and by faculty members because such a measure is independent of a program’s duration of existence and a faculty member’s years in the field. If rate is used, however, I would suggest that new programs (e.g., those in existence for 5 years or fewer) be excluded from such analyses until these programs have addressed some of the “growing pains” often faced with new program development (e.g., marketing, recruitment, alignment of courses with ABAI accreditation and BACB standards). Having developed our own behavior analysis programs at Caldwell University, which includes an ABAI-accredited M.A. program, I can certainly speak to the frustrations of needing to allocate time to endeavors other than research for the sake of the vitality of a program.

The sensitivity of annual publication rate can also be further refined by examining changes over time. In this way, programs that have either increased or decreased their productivity can be compared more meaningfully. One would expect that a relatively new program should increase its publication rate over time once the program has become more established.

In his commentary regarding perspectives on the future of behavior analysis, Schlinger (2010) stated that our field’s survival is dependent on training more behavior analysts. Clearly, the growth of graduate training programs is a step in a positive direction to ameliorate such concerns. Growth at the expense of quality training, however, might do more harm than good for the reputation for our field and our likelihood of obtaining “mainstream relevance” (Friman 2010). Indices of training quality, such as those proposed by Dixon et al. (2015), should be analyzed to identify variables that allow behavior analysts to best identify variables leading to the development of successful scientist/practitioners in our field. Hopefully, discussions afforded by commentaries such as this one and those by my colleagues (this issue) will move us further toward that goal.