Longitudinal Ex-Gay Study Update – Can Sexual Orientation be Changed?
August 12th, 2009
This week the American Psychological Association released a report that said that while religion and its value in a patient’s life should be considered and respected, therapists should not encourage clients to seek a change in sexual orientation and that there was no evidence to suggest that such efforts are successful.
This did not sit well with those organizations who build their existence on convincing their public that gay persons can “change” and that because such change is possible then public policy can be punitive to gay persons that do not submit themselves to such a change.
In response to the APA’s Resolution on Appropriate Affirmative Responses To Sexual Orientation Distress and Change Efforts, NARTH (the National Association for Research and Therapy of Homosexuality) declared
They selected and interpreted studies that fit within their innate and immutable view. For example, they omitted the Jones and Yarhouse study, the Karten study, and only gave cursory attention to the Spitzer study. Had the task force been more neutral in their approach, they could have arrived at only one conclusion: homosexuality is not invariable fixed in all people, and some people can and do change, not just in terms of behavior and identity but in core features of sexual orientation such as fantasy and attractions.
And Alan Chambers, president of Exodus International spoke to USA Today.
Its president, Alan Chambers, describes himself as someone who “overcame unwanted same-sex attraction.” He and other evangelicals met with APA representatives after the task force formed in 2007, and he expressed satisfaction with parts of the report that emerged.
“It’s a positive step — simply respecting someone’s faith is a huge leap in the right direction,” Chambers said. “But I’d go further. Don’t deny the possibility that someone’s feelings might change.”
So it was with great joy that those opposed to equality received news of evidence of change. The Baptist Press is crowing. Just “four days after an American Psychological Association task force released a 130-page report that said “gay-to-straight” therapies are unlikely to work”, they are trumpeting some amazing results of a study on Exodus International and their ex-gay ministries.
In findings that directly contradict mainstream academic thought, 53 percent of subjects in a new seven-year study reported successfully leaving homosexuality and living happily as heterosexual or celibate persons.
These “latest findings” are actually an update of the multi-year study of participants in Exodus ministries presented by Stanton Jones and Mark Yarhouse in their 2007 book, Ex-Gays? A Longitudinal Study of Religiously Mediated Change in Sexual Orientation.
The authors were not pleased that the study which they proclaimed throughout Christian media as an evidence of change in sexual orientation did not convince the APA.
“They selectively apply rigorous scientific standards,” he said. “So when it comes to examining the evidence that sexual orientation change can occur, they apply extraordinarily rigorous standards, and those standards allow them to disregard significant evidence that sexual orientation change can occur. That’s what happens with our study. They, I think, invalidly applied several methodological concerns to dismiss our study.
Indeed, the APA did apply concerns and dismiss the study.
Dr. Judith Glassgold, a clinical psychologist who led the APA task force, said the paper was not written in response to Dr. Jones’ work, though it did dismiss his findings.
“We don’t believe the claims were proven, to be honest,” said Dr. Glassgold in an interview. “In our looking at all the research we find that people don’t change their underlying sexual attraction. What they do is figure out a way to control their attractions. And some learn to live a heterosexual life but mostly for religious motivation.”
Presented as a counter-point to the APA’s declaration that there is inadequate evidence that therapies designed to change sexual orientation are effective, Jones and Yarhouse argue that:
the findings of this study would appear to contradict that commonly expressed view of the mental health establishment that sexual orientation is not changeable and that the attempt to change is highly likely to produce harm for those who make such an attempt.
I won’t speak to the likelihood of harm, but when it comes to change in orientation, their study convinced me of exactly the opposite.
My synopsis of the results, as published in the book, was
the Jones and Yarhouse study revealed little to no statistically measurable change in orientation in the prospective sample. The much touted “successes” were either in recollection (which again were quite small) or were those who had decided to no longer call themselves “gay”. However, they still identified their orientation as homosexual (”I’m not gay but my attractions are”).
In short, the Jones and Yarhouse study was funded and fully supported by Exodus and conducted by two researchers who were avid supporters of ex-gay ministries. They wanted to study 300 participants, but after more than a year, they could only find 57 willing to participate. They then changed the rules for acceptance in order to increase the total to 98. After following this sample for 4 years, 25 dropped out. Of the remainder, only 11 reported “satisfactory, if not uncomplicated, heterosexual adjustment.” Another 17 decided that a lifetime of celibacy was good enough.
What the 2007 Jones and Yarhouse book revealed, and what this update further confirms, is that the “change” which NARTH and Exodus loudly proclaim is not a change in sexual orientation at all and, in fact, may be nothing more that a change in identity or recollection.
Prospective v. Retrospective
In order to understand the J&Y study, you have to understand the distinction between retrospective and prospective sampling. Prospective uses currently measured data, and retrospective uses recollections about the past.
For example, if one were wanting to compare changes in the length that a student has to walk in their morning commute to school, a prospective study would select a random sample (say a selection of schools), measure the distance the students walked, and repeat the process over a long enough period of time to determine if there is change. A retrospective study would go ask Grandpa and compare today’s walking distance to “ten miles through the snow, barefoot, uphill both ways!”
Obviously, measured differences are far more accurate than recollected differences. Time has a way of providing support for what we want to believe and recollections tend to be very selective. Things were simpler then, or tougher; summers were hotter, or milder; politicians were more honest, or scoundrels.
So the best studies are prospective rather than retrospective. Oddly, the J&Y study is both.
Of the 98 participants, 57 were more-or-less prospective. These were persons who had been involved with an Exodus (ex-gay) ministry for less than a year. So while there may have been some recollection error, it was at least a discussion of reasonably recent events. These participants are called Phase 1.
Unable to get a sample size that the authors felt was sufficient, they then recruited Exodus participants that had been in the programs for one to three years. These participants are called Phase 2 and to the extent that there is a measurement from a recollected starting point, their participation is retrospective.
Fortunately, it is possible to distinguish between the results for these two groups.
The paper presented by Jones and Yarhouse to the APA is significantly less detailed than was their book, as could be expected. Specifically, the several scales of measurement were reduced to two, Kinsey and Shively & DeCecco, and while the book provided information on interim points, the paper uses only the starting point (T1) and the final point (T6).
The results in the book are based on 75 of the original 98 participants. Since J&Y presented their results in their book, an additional 14 participants have dropped out of the study, bringing the sample size down to 61. The remaining participants have now been part of the study for six to seven years.
Because the total sample is a hodge-podge of two very different subpopulations, it is informative only to the extent that it reveals information about the difference between those subpopulations.
The group that is most accurately studied, and that whose results are most revealing about the extent to which Exodus is successful is Phase 1, the prospective study. And this is what Jones and Yarhouse report about that subpopulation:
- There was, on average, virtually no change in sexual orientation on the Kinsey scale using measures of behavior, sexual attraction, emotional/romantic attraction, and fantasy.
- There was, on average, a small but not significant increase in homosexual behavior.
- There was, on average, a slight but not significant reduction in homosexual attraction.
- There was, on average, virtually no increase in heterosexual attraction.
In other words, on average, after six to seven years of participation, those who went through Exodus ministries reported over the period of their involvement no change in sexual orientation at all.
Averages v. Individuals
Averages, while meaningful to statisticians and to those who are evaluating the effectiveness of Exodus International, do not tell the full story. We must also look at individual results.
For their book, Jones and Yarhouse classified their participants into categories based on their individual reports. They came from a religious evaluation model and defined two groups as successful (conversions and chaste), two as failure (identifying as gay and considering identifying as gay), and two in the middle that were still trying but seeing little to no results. For the final report, the authors changed their procedure and allowed participants to select their own category.
As I am less interested in adherence to religious identities and more interested in sexual orientation change, I’ll group the failure and the middle two together. After time T6, J&Y report:
- Success: Conversion – 14 (23%)
Success: Chastity – 18 (30%)
Non-Success – 29 (48%)
This does not, however, present an accurate story of the study participants. It does not account for those who dropped out of participation and thus overstates the success rates. One could extrapolate from this reporting method that eventually only those who are successes of some sort will remain and the authors could declare with great fanfare that 100% of all Exodus participants eventually succeed.
But that would not be truthful.
When I made the observation that drop-outs should be considered a likely failure, those who defended the skewing upward of success rates argued that because we don’t know the reasons for discontinued participation, it was just as easy to believe that these individuals were now happily heterosexually married and not wanting to be reminded of their old life as that they had embraced a gay identity. But additional information in this report reveals otherwise.
Of the 14 persons who left the study between T3 and T6, two were Conversion, one was Chastity, and the remaining were Non-success. We know from the book that one of the Conversion drop-outs reported that he had never been heterosexual and was simply reporting what he thought the authors wanted to hear. So it is rather unlikely that these drop-outs went off to live heterosexual lives. Nor is it (or ever was it) likely that any sizable chunk of those who dropped out before T3 left because they are now happily hetero.
Considering drop-outs as their own category, a more accurate reporting of the self-identified placement into categories would look like this:
- Success: Conversion – 14 (14%)
Success: Chastity – 18 (18%)
Non-Success – 29 (30%)
Drop-Outs – 37 (38%)
And considering just the Phase 1 participants, the results are
- Success: Conversion – 5 (9%)
Success: Chastity – 6 (11%)
Non-Success – 18 (32%)
Drop-Outs – 28 (49%)
When looking at these numbers, we should consider two things about the “conversion” category shown above.
First, much of Exodus’ efforts go into changing identity. They view a “gay identity” as sinful and contrary to a “Christian identity”.
So this change in identity may not be related to an actual change in orientation. As I noted above, the first measurement of “change” reported in the book – the one trumpeted in anti-gay press upon the book’s press release – was a change in self identification. Yet is was accompanied by a measurement that spoke of one’s orientation as separate from one’s identity and found that those who claimed that they were not homosexual were willing to admit that their orientation is homosexual. It was literally a declaration that, “I’m not gay but my orientation is.”
We should be careful to recognize that those claiming conversion at T6 may be doing so for themselves and not for their orientations. The authors do acknowledge that such success may be seen as relating more to identity than to orientation:
Some may see these results as reflecting not a change in sexual orientation for most participants who reported such change, but rather a change in sexual identity. Such a change might result from how one thinks of oneself and labels one’s sexual preferences (that is, attributions and meaning-making).
But with Exodus placing heavy emphasis on identity, by allowing unanalyzed self-assignment the authors may have created a scenario in which there is an inflated increase in the “success” categories.
And second, this report differs from the book in that the qualifiers are removed. The book provided discussion of the non-traditional definitions of “heterosexual” used in the study and how those who were so identified also experienced wandering eyes, erotic dreams, and other situations that are most often associated with a homosexual orientation. In the paper, it is limited to
[W]hile we found that part of our research population experienced success to the degree that it might be called (as we have here) “conversion,” our evidence does not indicate that these changes are categorical, resulting in uncomplicated, dichotomous and unequivocal reversal of sexual orientation from utterly homosexual to utterly heterosexual. Most of the individuals who reported that they were heterosexual at T6 did not report themselves to be without experience of homosexual arousal, and they did not report their heterosexual orientation to be unequivocal and uncomplicated.
I don’t think that I’m alone in noting that few of the heterosexuals that I know experience much homosexual arousal. Perhaps Stanton Jones himself said it best in an interview.
“A typical hetero male finds himself attracted to a wide range of females. But among the successful people who reported conversion the typical response was I’m very happy with my sexual responses to my wife, but I don’t experience much hetero attraction to other women. Also, when asked and pressed about whether they still find attraction to men, they will say: ‘Yes, if I let my mind go in that direction.’ “
And finally, when comparing the individual with the average, it must be noted that without an average change, any individual change experienced is offset by an equal and opposite experience.
In other words, for every person who came to Exodus and found that they became one Kinsey point more heterosexual, there was a person who found that Exodus made them one Kinsey point gayer. If Exodus sees their mission as rescuing those sinking in a sea of sin, for each person they pull into the lifeboat, they hit another over the head with an oar.
Failure to report Phase 2 Results
Jones and Yarhouse report the “whole population”, a commingled combination of Phase 1 and Phase 2 as though it is informative. While they do break out Phase 1 results, they do not disclose Phase 2 results.
I believe that were Phase 1 to be visually compared to Phase 2, the variances between the two would be startling. The question jumping out from the report might shift from whether there is a significant effect size in responses to why these two subpopulations are reporting opposite conclusions.
And, indeed, the results from Phase 2 can be deduced to be significantly different from Phase 1. If we know the average response of the 29 remaining participants in Phase 1 and the total average responses of all 61 remaining participants, we can back into the Phase 2 reported change.
A comparison of the two would show:
- Kinsey – behavior only
Shively & DeCecco homo
Shively & DeCecco hetero
As we can see, there are sharp differences in the results of these two subpopulations. And although no information on Phase 2 is directly reported, the authors somewhat acknowledge that the two subpopulations vary in results
We expected that the results of change would be somewhat less positive in [Phase 1], as individuals experiencing difficulty with change would be likely to get frustrated or discouraged early on and drop out.
That is a round-about way of admitting that the Phase 2 subpopulation does not include those who got frustrated early and dropped out in the first one to three years. It avoids pointing out that results for the Phase 2 subpopulation are already skewed towards those who either believe they are experiencing “change” or have a stronger more deeply dedicated commitment to Exodus ministries.
But even so, with such astounding results in this subpopulation, why wouldn’t the authors include this separate information. It may be that isolation of Phase 2 raises questions about the validity of including them at all and, more importantly, what it says about the claims made by Exodus members both included and not included in the study.
The real difference between Phase 1 and Phase 2 is more than just that P2 has been in the program for a few more years. It is more than that they have fewer drop outs. The real difference is that P2 is based on recollection to a much greater extent than P1.
And Phase 2 participants recalled being more gay than Phase 1 reported. Significantly, especially in the area of behavior. The Kinsey 1 report was 4.52 for Phase 1 and 5.49 for Phase 2.
There is no reason to believe that those in Phase 2, having eliminated the drop outs, actually were any more homosexually oriented than those in Phase 1. Rather, it seems likely that they simply recalled being more homosexually oriented when they established their base point some one to three years later.
So all reported change in Phase 2 – and indeed all reported change – may be attributable to this variance in starting point due to reliance on recollection. Ultimately, all of Jones’ and Yarhouse’s announced success may be nothing more than, “I remember being much more gay three years ago than I am today.”
The one subpopulation that Jones and Yarhouse are excited about is what they call the “truly gay subpopulation.” These are defined as those who “scored above the scale midpoint at T1 for homosexual attraction, and for homosexual behavior in the past, and for having previously embraced full homosexual or gay identity.” This subgroup reported the most change.
It is difficult to know whether these persons are mostly Phase 1 or Phase 2, but it would appear that they are a combination of both. We know from the break out of results in the book that a number of the non-successes in the truly gay subpopulation were also Phase 1. This lends itself to assumptions that those in the truly gay subpopulation that reported progress were likely in Phase 2 and that much, if not all, of their progress consisted solely of exaggerated recollection.
This is further supported by noting that most of the change reported over the seven year life of the study was between the first measurement (often as recollected) and the second. In discussing the possibility that reported change is largely identity, the authors noted:
This might also explain to some why the Truly Gay subpopulation showed more dramatic change, as their shift was away from a more pronounced gay identity. Such a departure may have been measured as a greater movement away from something that had previously been more salient to them.
Or, more likely, a greater movement away from the recollection of being very gay three years ago.
Based on the Jones and Yarhouse book, Ex-Gays? A Longitudinal Study of Religiously Mediated Change in Sexual Orientation, and on their follow up report, Ex-Gays? An Extended Longitudinal Study of Religiously Mediated Change in Sexual Orientation, we can observe the following:
- The prospective sample reported, on average, virtually no change in attractions and a small increase in homosexual behavior.
- A retrospective look at ones perceptions of prior orientation from the perspective of one to three years yields a sharply different result from that seen by a prospective sample. This change in perspective may account for all reported change in Exodus ministries.
- Most change reported away from homosexuality and towards heterosexuality was in the interval between the starting point (T1) and the second measurement point (T2). This change occurred most strongly in the retrospective sample and may be due to variances in recollection.
- A combined prospective and retrospective sample experienced, on average, no significant increase in opposite sex attraction.
- A small percentage (perhaps 9%) of those who start Exodus programs may eventually self-categorize themselves as “experiencing substantial reductions in homosexual attraction and substantial conversion to heterosexual attraction and functioning. These persons will be unlike other heterosexuals in that they will continue to experience homosexual arousal and not experience much attraction to the opposite sex.
- Another small percentage (perhaps 11%) of those who start Exodus programs may eventually achieve a life of manageable homosexual attraction and chastity.
- Others may continue perpetually in Exodus programs without ever achieving any significantly reduced homosexual attractions.
- Eventually, most of those who start Exodus programs will drop out.
- On average, for each person who enters and Exodus program and finds any movement away from homosexual attraction, another will find movement towards homosexual attraction.
But these observations are not readily obvious from the media reports of either the 2007 Jones and Yarhouse book nor this follow-up report. And those seeking “proof” that homosexuals can “change” have used both to advance a false image of the results of this study. Exodus, NARTH, and many others will spin this study to come to conclusions that are far from of those I’ve stated above.
The authors have a moral responsibility to discourage those who will make false statements or who will falsely claim that this study justifies their ex-gay or anti-gay endeavors. And they have a moral obligation not to allow their wishes about the mutability of sexual orientation cloud the results of their study and give false hope to those who believe Exodus’ slogan that “change is possible”.