from NARTH - APA Highlights

Does the APA Task Force Report
Apply its Research-Methodology Standards Consistently?
A Preliminary Examination

By Christopher Rosik, Ph.D.

The APA's 2009 Task Force report, entitled, "Appropriate Therapeutic Responses to Sexual Orientation" contains a prominent section dedicated to identifying the methodological problems in the research on sexual-orientation change efforts (SOCE).

This section is meticulous in its efforts to identify any and all limitations to SOCE research, with a clear aim of discrediting this body of literature. While no body of research is free from limitations, one measure of the degree of objectivity behind critiques of this nature is the extent to which they are uniformly applied to research affirmed by the reviewers.

In the case of the APA's report, I was able to locate two articles cited by the Task Force that were available in full text in the EBSCO database.

Research by McCord, McCord, and Thurber (1962) is cited by the Task Force to repudiate theories that a homosexual orientation is associated with particular family dynamics, gender identity, and/or trauma.

A more recent study by Kurdek (2004) is also reported by the Task Force in support of the Task Force's claim that gay, lesbian, and heterosexual relationships are essentially similar.

A review of the Task Force's methodological critique of SOCE identified 16 separate concerns that, in their eyes, are each significant enough in themselves to call the SOCE research findings into question. (I have listed these concerns in Table 1.) My preliminary methodological examination of these articles suggested that, by the APA Task Force standards, the McCord et al (1960) research exhibits 10 of the 16 (63%) methodological problems, while 2 (13%) additional problems could not be evaluated.

The Kurdek research faired slightly better, with only 8 (50%) methodological problems identifiable and another 3 (19%) either not applicable or not able to be evaluated. I will review some of the problems in these studies below.

The reader should keep in mind the key point that that these studies were cited uncritically in the APA Task Force report.

McCord et al study. The McCord et al (1962) study examined data among a sample of boys between the ages of 10 and 15 which was culled from observational records charted 12 to 18 years prior to their investigation. The researchers examined a number of variables ascertained from the records and generally sought to determine whether these variables differed among boys from homes with or without a father present. They reported that homosexuality did not differentiate between boys with fathers present and those with absent fathers. However, the same methodological problems highlighted by the APA Task Force as evidence of weak research design in sexual-orientation change studies were evident in this same study from the start.

Attrition pared the original convenience sample of 325 down to 255. The final sample included 150 boys from intact families and only 55 who had father-absent families, with no randomization process used in selecting these groups. Consequently, some of the cell sizes were very small. Nearly all dependent measures (e.g., "affectional interaction," "homosexual tendencies") were not clearly defined and where defined, the terms used in these definitions were similarly vague. Moreover, no validity or reliability information was presented relative to these set of ratings that comprised variables such as homosexual tendencies. The relationship between homosexual tendencies and sexual orientation (a term never used in McCord, et al) is far from clear, even though the APA Task Force appears to assume (in this study, which is supportive of APA's position) that they are commensurate. The sample was restricted to low SES boys. The article further does not make clear to what extent the researchers reviewing the records were aware of the study's purposes.

Kurdek study. Kurdek (2004) compared longitudinal data obtained from gay and lesbian cohabiting couples and partners, with heterosexual married couples (with and without children). They examined five domains of relationship health to determine if similar variables predicted relationship stability for these couples. Kurdek found that where differences between same-sex and heterosexual couples did exist, over two-thirds of these differences indicated that gay and lesbian partners functioned better than heterosexual partners. The author concluded that the processes that regulate gay and lesbian relationships are the same as those that regulate heterosexual partners.

Methodological problems that--if consistently applied by APA--would lead the APA Task Force to raise questions regarding Kurdek's (2004) conclusions, begin with his sampling procedures. Different methods were used to obtain the convenience samples of heterosexual and same-sex couples. Heterosexual couples were recruited through marriage announcements published in a daily newspaper. Same-sex couples were recruited through gay and lesbian periodicals, and these participants in turn were encouraged to recruit additional same-sex couples. Thus, selection and response bias may well have been a factor, especially in the recruitment and responses of same-sex participants. However, no measure of test-taking attitude was included that could have addressed this concern.

The longitudinal waves consisted of subsamples of participants, as attrition appeared to take a significant toll over the eight assessment periods. At first assessment, there were 80 heterosexual couples with children, but by the eighth assessment, only 50 remained. The N for heterosexual couples without children declined from 146 to 29, gay couples decreased from 80 to 33, and lesbian couples diminished from 53 to 52. The sample was also restricted primarily to White and college-educated individuals. The article did not present descriptive information for the correlational analyses that would permit evaluation of the extent to which univariate and multivariate assumptions had been met. Nor was the global evaluation outcome variable defined in a clear manner. In all instances, the variables studied were derived from self-report measures.

Other methodological concerns were evident in this study, beyond those identified as problematic by APA in the sexual-orientation research that the Task Force rejected. While these will not be detailed for this analysis, one does bear mentioning in the present context. Specifically, Kurdek (2004) noted that same-sex couples were added to the sample at two points over the entire assessment period, meaning gay and lesbian couples did not have the same number of possible assessments. This is reminiscent of the Jones and Yarhouse (2007) study of ex-gays, where the authors added to their sample of participants in religiously based SOCE. This sample addition was touted by critics as a serious methodological flaw in Jones and Yarhouse's research that had introduced bias into their study of successful sexual-orientation change.

While the two studies have different aims and foci, an equally applied methodological critique would certain raise concerns about the bias that Kurdek might have introduced into his own sample of same-sex couples by adding additional subjects after the initial assessment period.

Conclusion. Serious concerns about the APA Task Forces' objectivity have to be raised if this preliminary investigation is at all indicative of the methodological problems which exist in the literature that was cited uncritically by APA, in order for APA to make its case that homosexual and heterosexual relationships are equivalent, and to dismiss traditional theories concerning sexual-orientation etiology.

Certainly, in the present analysis of the McCord et al (1962) and Kurdek (2004) studies, had the Task Force applied their SOCE methodological critique with similar rigor, they would have been unable to cite these studies in any sort of generalized or conclusive manner. Yet such certainty is precisely what the APA Task Force seems to imply in their report.

This disparate treatment of the SOCE literature in comparison to other sexual-orientation research reflects both the lack of ideological diversity on the Task Force, and the essential sociopolitical nature of the report. This, in turn, casts significant doubt upon the impartiality and accuracy of the APA's conclusions regarding the efficacy of sexual-orientation change efforts.

References

Jones, S. L., & Yarhouse, M. A. (2007). Ex-gays?: A longitudinal study or religiously mediated change in sexual orientation. Downers Grove, IL: InterVarsity Press Academic.

McCord, J., McCord, W., & Thurber, E. (1962). Some effects of paternal absence on male children. Journal of Abnormal and Social Psychology, 64, 361-369.

Kurdek, L. A. (November, 2004). Are gay and lesbian cohabiting couples really different from heterosexual married couples? Journal of Marriage and Family, 66, 880-900.

Table 1

Updated: 25 March 2010

The 2010 NARTH Convention
November 5, 6 and 7
Renaissance Philadelphia Hotel
CONFERENCE REGISTRATION & INFO