The gender of participants in articles published from 1991 through 2000 in eight journals relevant to mental retardation was determined. Overall, participants were used in 65% of the articles; gender was not reported in 26% of them. When gender was reported, 6%, 8%, and 60% of the investigators used females only, males only, and both sexes, respectively. Unless gender is reported, one cannot ascertain to whom results should generalize or whether gender affects the variable under investigation. For these and other reasons, we recommend that researchers routinely specify how many of their participants are males and how many are females. This information requires little space and in most cases is easy to obtain.
The Publication Manual of the American Psychological Association [APA] (American Psychological Association, 2001) dictates the style of manuscripts published in all APA journals and many other professional journals. For example, eight of the primary outlets for research dealing with mental retardation and related topics, American Journal on Mental Retardation (AJMR), Exceptional Children (EC), Education and Training in Mental Retardation (ETMR), Journal of the Association for Persons with Severe Handicaps (JASH), Journal of Developmental and Physical Disabilities (JDPD), Journal of Special Education (JSE), Mental Retardation (MR), and Research in Developmental Disabilities (RDD), specify that manuscripts be prepared according to the APA manual. It is specifically stated in the manual that when human beings are the participants in a study, “report major demographic characteristics such as sex” (APA, 2001, p. 18).
Despite this recommendation, authors of articles published in psychology journals often fail to specify the sex (gender) of their research participants (e.g., Carlson & Carlson, 1960; Holverstodd et al., in press; Jarema, Snycerski, Bagge, Austin, & Poling, 1999; Pepinsky, Hill-Frederick, & Epperson, 1978; Wann & Hamlet, 1995). For instance, Gannon, Luchetta, Rhodes, Pardie, and Segrist (1992) reported that of the articles in eight psychology journals that included participants, 25% and 34% did not contain information on the gender of participants in 1970 and 1990, respectively. Similarly, Ader and Johnson (1994) found that 30% of the authors of articles published in eight APA journals in 1990 did not report the gender of participants.
None of the investigators concerned with the gender of participants in articles focused on journals devoted primarily to mental retardation and related topics. Our purpose in the present investigation was to obtain information about the gender of participants in articles published in such journals. To do so, we evaluated each article published in AJMR, EC, ETMR, JASH, JDPD, JSE, MR, and RDD from 1991 through 2000. Although articles relevant to mental retardation appear in other journals, the eight journals selected for analysis cover a broad spectrum of prestigious outlets for research, theory, and opinion relevant to mental retardation.
Several authors have supported the notion that researchers in psychology and related fields should routinely report the gender of their participants, and they have provided good reasons for doing so (Ader & Johnson, 1994; Carlson & Carlson, 1960; Denmark, Russo, Frieze, & Sechzer, 1988; Gannon et al., 1992; Holverstott et al., in press; Jarema et al., 1999; Pepinsky et al., 1978; Wann & Hamlet, 1995). They noted that if the gender of participants is not specified in an article, readers cannot ascertain whether the reported results should generalize to males only, to females only, or to both genders. Similarly, they cannot determine whether participants' gender influences the behavioral characteristic or disorder of interest in the investigation. Finally, unless the gender of participants is reported, readers cannot ascertain whether males or females were excluded from participation in a study. Males appear to be overrepresented as participants in several research areas, which has caused concern (e.g., Carlson & Carlson, 1960; Etaugh & Bohn Spandikow, 1979; National Institutes of Health, 1994; Pepinsky et al., 1978; Wann & Hamlet, 1995).
Every article published in AJMR, EC, ETMR, JASH, JDPD, JSE, MR, and RDD from 1991 through 2000 was evaluated by one of two raters, each the holder of an advanced degree in psychology, to determine whether any participants were described in the article. If so, the raters recorded on standardized data sheets the number of participants and their gender (i.e., gender was coded as male, female, or unknown). To allow interrater agreement to be calculated, the two raters independently scored the same 306 articles (10% of the 2,972 total articles rated). They agreed perfectly with respect to which articles used participants (N = 188) and whether males only, females only, or both males and females (N = 131) served as participants. They agreed almost perfectly on whether or not the gender of participants was reported (N = 182, 97% agreement). In the 6 cases where the two raters disagreed, a third independent rater evaluated the article with the two primary observers, and a consensus was reached. The consensus value was used in calculating the data reported in this article.
Of the 2,976 articles evaluated, 1,929 (65% of the total) included participants. Of those articles, 429, 253, 263, 136, 183, 138, 268, and 259 appeared in AJMR, EC, ETMR, JASH, JDPD, JSE, MR, and RDD, respectively. Across all years and all journals, of the 1,929 articles with participants, 26% included no description of the gender of the participants, 6% reported using only female participants, 8% reported using only male participants, and 60% reported using both male and female participants. The level of nonreporting of participants' gender was 24, 32, 22, 30, 19, 29, 30, and 20% in AJMR, EC, ETMR, JASH, JDPD, JSE, MR, and RDD, respectively.
The potential problems associated with a failure to report the gender of participants, summarized previously, are real and potentially serious. Therefore, the fact that 26% of the articles that we evaluated failed to provide this information may be a cause for concern; at least, similar findings for other journals have concerned other authors (e.g., Ader & Johnson, 1994; Denmark et al., 1988; Gannon et al., 1992; Pepinsky et al., 1978). Whether such concern is justified ultimately requires careful analysis of why the gender of participants was not reported in individual articles, and the possible implications of the omission. For example, researchers may fail to mention participants' gender because they (a) cannot obtain this information; (b) are unaware of the potential importance of reporting the gender of participants; (c) feel it is meaningless to report gender unless data are analyzed as a function of it, which is misleading or impossible due to the size or other characteristics of the data set; or (d) intentionally excluded members of one gender from participating in their study and wanted to hide this fact.
The present data are of value if they encourage other researchers to examine the kinds of studies in which investigators characteristically fail to report gender or if they encourage journal editors and referees to ask routinely whether participants' gender could and should be reported in each article that they evaluate. They also are of value if they encourage future researchers to examine how often authors who report the gender of participants also analyze their data as a function of this variable and, if not, why not. There are several reasons why researchers do not analyze data as a function of the gender of participants; most are similar to the reasons they do not report the gender of participants. In both cases, some of the reasons are good, but others are questionable from a scientific or ethical perspective. For example, if there is no good theoretical or practical reason to analyze data as a function of gender, or if the data set is too small to support meaningful conclusions about the effects of gender, it is appropriate not to analyze data separately for male and female participants. It is, however, inappropriate to ignore gender if there is reason to believe that results might vary for males and females, or concern that this might occur. It also is inappropriate to fail, for theoretical or political reasons, to not report gender differences if they are observed. Although it would be difficult to evaluate on a post hoc basis why researchers who reported participants' gender failed to report their findings as a function of this variable, it would be easy enough to have experts judge whether in their opinion this should have been done. To our knowledge, however, this has not been done in any research area. In fact, even data on how often results are analyzed as a function of participants' gender are lacking.
A criticism that has been levied against psychological research is that males have been overrepresented as research participants and that this overrepresentation raises significant problems concerning the generality of applied research results (Carlson & Carlson, 1960; Etaugh & Bohn Spandikow, 1979; Pepinsky et al., 1978; Wann & Hamlet, 1995). In the articles examined here, we did not find that males were overrepresented as participants. Specifically, 91.7% of those articles included male participants and 89.4%, female participants. Although the reason for comparable representation of males and females as participants is not apparent, this finding is encouraging inasmuch as we suspect it reflects awareness among researchers of the importance of studying the variables of interest in males and females alike. If this is true, there may well be good reasons for not reporting participants' gender in most of the articles in which investigators failed to do so. Only further and more detailed research can determine whether this is, in fact, the case.
NOTE: The reported data were collected as part of the dissertation presented by the first author to the Department of Psychology, Western Michigan University, in partial fulfillment of the requirements for the doctoral degree.
Authors:Cari L. Porter, PhD, Research Associate, and Alan Poling, PhD ( firstname.lastname@example.org), Department of Psychology, Western Michigan University, Kalamazoo, MI 49008. LeeAnn Christian, PhD, Director of Research, Institute for Applied Behavior Analysis, 5777 W. Century Blvd., Los Angeles, CA 90045. Requests for reprints should be sent to the third author