Changing Sexual Orientation and Identity? The APA Report
by Andrew Goddard and Glynn Harrison
As Christians believe that all truth is God’s truth, in seeking to understand any phenomenon and to develop a practical Christian response to it, it is vital we understand and take into account the best evidence available from scientific research and other academic disciplines. This is particularly important in Christian ethics and Christian engagement in the public square. Being shaped by biblical teaching is vitally important but if we refer only to the Bible and pay no attention to other sources of authority and truth we will weaken the credibility of what we say and undermine Christian discipleship and mission.
Within debates about human sexuality in both church and society this aspect of relating wider knowledge about sexuality to our understanding of Scripture has often become another major area of contention and conflict. In August 2009, the American Psychological Association (APA) published the important report of its Task Force on Appropriate Therapeutic Responses to Sexual Orientation, which was reported on various Christian websites and blogs (eg Anglican Mainstream, Fulcrum Forums, Peter Ould, and Warren Throckmorton). This article attempts to explain and evaluate some of the key findings of that report but first it is important to set it in the context of past APA decisions and the wider Christian discussion.
APA, Sexual Orientation and Christian Responses: An Introduction
As the Report states in the opening paragraph of its Introduction, the last three decades have seen a major change in understanding and responses to homosexuality and bisexuality among psychologists:
In the mid-1970s, on the basis of emerging scientific evidence and encouraged by the social movement for ending sexual orientation discrimination, the American Psychological Association (APA) and other professional organizations affirmed that homosexuality per se is not a mental disorder and rejected the stigma of mental illness that the medical and mental health professions had previously placed on sexual minorities….Through the 1970s and 1980s, APA and its peer organizations not only adopted a range of position statements supporting non-discrimination on the basis of sexual orientation but also acted on the basis of those positions to advocate for legal and policy changes….Licensed mental health providers of all professions increasingly took the perspective throughout this period that homosexuality per se is a normal variant of human sexuality and that lesbian, gay and bisexual (LGB) people deserve to be affirmed and supported in their sexual orientation, relationships and social opportunities. This approach to psychotherapy is generally termed affirmative, gay affirmative, or lesbian, gay and bisexual (LGB) affirmative.
This development has been criticised by many conservative Christians, a significant number of whom look to the work and ministry of the minority of therapists who have taken a different approach. In particular a group of therapists called NARTH (National Association for Research and Therapy of Homosexuality) offers an alternative approach within the profession, often called ‘reparative therapy’, that it is argued has been marginalised and dismissed by mainstream professional organisations such as the APA. In addition to NARTH there are also a wide range of Christian counselling and healing ministries, many linked to Exodus International, who use various techniques and programmes to provide Christian ministry to help people who want to overcome same-sex attraction and who seek sexual reorientation to function heterosexually. Their work has most fully and recently been studied by Stanton Jones and Mark Yarhouse in their book Ex-Gays? and in a paper updating that study that was also presented at the APA this year.
The issue of seeking to change sexual orientation has, as a result of these differences, become a major bone of contention within both the “culture wars” (especially in the
In this context, the APA decision to commission and publish a major report on what it terms ‘sexual orientation change efforts’ (SOCE) is a welcome development.
The APA Report: Explaining a “Systematic Review”
The Report sets out to answer 3 specific questions:
- Are sexual orientation change efforts (SOCE) effective at changing orientation?
- Are SOCE harmful?
- Are there any additional benefits that can be reasonably attributed to SOCE?
The Report itself is a large document (comprising 92 pages of often technical text and over 30 pages of appendices and references). In addition there is a three sentence abstract and 7-page Executive Summary. Inevitably much of the reporting and subsequent debate has focussed on these shorter and more accessible accounts and this, as we shall see, creates problems.
The bulk of the report seeks to answer the three questions by carrying out what is called a ‘Systematic Review’. Before reading and evaluating the APA Report it is important to understand what such a review does. A Systematic Review is a method which has been developed out of a recognition that all of us – even ‘objective scientists’ – are ‘coming from somewhere’. Our emotional investments and personal convictions impact profoundly on the observations we make, and the conclusions we reach, about the nature of the world and the world of science is not immune from these effects. Hidden biases can be detected when scientists use even the most modern clinical trial approaches. For example, the odds of recommending an experimental drug as treatment of choice increases fivefold when a profit-making organisation provides the funding, compared with research paid for out of neutral sources (see Montori,V et al (2004) Users Guide to detecting misleading claims in clinical research reports. British Medical Journal; 329:1093-1096). It is in order to try to minimise these biases that scientists have developed the method of the Systematic Review.
The Systematic Review is a rules-based approach which tries to set clear and simple questions, identify and sift all of the evidence against a set of quality criteria, and then merge the findings in as transparent and objective a manner as possible. Reviewers attempt to identify all the ‘high quality’ studies and clinical trials relevant to the question being asked and then apply various statistical methods that allow the results to be summarised. Where there are insufficient ‘high quality’ studies to make this final stage possible, the authors may carry out a ‘narrative’ review of what evidence there is. Then they make what recommendations they can with the appropriate cautions about their own bias having crept into the review and health warnings about the poor quality evidence and need for more research.
It is important to understand what a ‘high quality’ study is. The ‘gold standard’ study for determining whether or not a particular treatment or intervention ‘works’ is something called the Randomised Controlled Trial (RCT). Here subjects are randomly assigned to either a treatment group or to a ‘control’ group in the hope that this will spread factors which could alter the result (such as the placebo effect of subjects’ belief in the treatment) between the two groups. If that can be achieved (and the study has been properly designed in terms of the numbers required to make it statistically credible) then it should be possible to ‘isolate’ any additional effect in the group that received the active treatment. In a nutshell, the RCT tries to control for the placebo effect by giving both groups something that looks the same, but where only one has the active ingredient. The problem is that many areas of medicine and psychology don’t lend themselves to this type of RCT evaluation. As a result it is important to ask what conclusions can be drawn and what happens where there are insufficient ‘gold standard’ studies.
When ‘gold standard’ RCT studies are lacking, the authors of a systematic review have to decide what to say about this absence of evidence. Great care is needed to understand what has and what has not been found. Over a decade ago, Altman and Bland (Altman D and Bland J.M (1995) ‘Absence of evidence is not evidence of absence’. British Medical Journal. 311:485) set out the dangers of concluding that ‘absence of evidence is evidence of absence’. Just because we lack evidence that X works does not mean we have evidence that X does not work. There is simply insufficient reliable evidence to answer the question posed for the Review. In such situations therefore it is misleading to say ‘there is no evidence that this intervention works’. There is simply no evidence, period. The authors should therefore conclude that it is not possible to say whether a given treatment works, or whether it does not work.
As we have noted, when this happens, many systematic reviewers critique what evidence is available using lower standards and draw their conclusions with the appropriate cautionary notes and caveats. Clinicians reading the review must then make up their own minds about whether a particular treatment associated with limited potential harm (but possible benefit) should be withheld due to lack of gold-standard evidence or whether, on a case by case basis, they will offer the patient what is available. This is a common problem in medicine and the psychiatric/psychological sciences and their various professional organisations may set up an ‘expert’ panel to make recommendations about a particular treatment based on their best (but not necessarily unbiased) judgment.
Systematic reviews (such as that conducted by the APA) are therefore not panaceas for the elimination of bias and prejudice. They have their own limitations and the quality of systematic reviews varies from study to study, depending upon the experience, qualifications and skills of the authors and the quality of the raw material available for review. What then can be said about the APA’s Systematic Review in relation to SOCE?
The short answer is that the APA systematic review is a very mixed bag. There are strengths but there are also some serious weaknesses.
The Strengths of the APA Report
The APA Reviewers follow the usual rules for a systematic review: they set out their questions clearly, standardise their search methodology for all the relevant research studies since 1960, and then critically appraise these studies against a set of explicit quality criteria. Section 3 of the review, which covers methodological problems in the research literature on SOCE, is thorough, competent and fair.
In their overview (p28) of their systematic review (which we shall contrast shortly with their abstract and the executive summary) the authors identify a lack of credible ‘gold standard’ Randomised Controlled Trials (RCTs) that would allow them to answer their primary research question about whether SOCE are ‘effective at changing sexual orientation’. Having recognised this lack they go on to identify a few studies, all carried out several decades ago on only one kind of intervention (behavioural), which used experimental procedures with some elements of experimental rigour. However, none of these (bar one with only 16 subjects) had a traditional control group, none were conventional RCT’s and none could be classified as gold standard in the sense specified in the report (p2 and then pages 28-34).
In the light of this, using the methods described above, the authors correctly conclude that there is a complete lack of high level evidence that would assist in answering their questions. They thus conclude (p28) that the answer to the first of their three specific questions (Are sexual orientation change efforts (SOCE) effective at changing orientation?) is that
there is little in the way of credible evidence that could clarify whether SOCE does or does not [emphasis added] work in changing same-sex sexual attractions.
The question this raises is why (as discussed below) they then present a different conclusion in the abstract and executive summary? However, this section of the paper, in so far as it goes, is reliable and fair in terms of the objectives of a systematic review.
Given the lack of robust evidence either way, the authors finally seek to find some way of ‘affirming’ those who seek change in their sexual attractions whilst recognising that, on the basis of the absence of high level evidence, they are unable to recommend interventions that make efforts to change orientation. Here they introduce an important distinction which has been welcomed by a number of evangelicals. They recommend an affirmative approach that recognises that sexual orientation identity (as opposed to sexual orientation) is a malleable construct. They therefore propose that, regardless of their sexual desires, clients may make choices about their ‘identity’. Further, whilst highlighting the potentially harmful effects of the stigmatising attitudes emanating from religious conservatives, they nevertheless explicitly recognise the need for psychologists to recognise the diversity of values, including religious values, of those seeking their help. Thus, although the report finds no high level evidence to recommend interventions to change sexual orientation, it does seek to accommodate the religious convictions of clients and affirm their right to decide how they manage their feelings and behaviours within the constraints of their personal sense of identity and chosen lifestyle. It is this apparent shift in APA attitudes that has been so welcomed by many evangelicals.
The Weaknesses of the APA Report
As already noted, there is a marked inconsistency in the conclusions of the report. On p.28 the authors conclude there is ‘little in the way of credible evidence that could clarify whether SOCE does or does not work’ but in the abstract and the executive summary, on which much reporting inevitably focussed, they appear to reach a different conclusion. In the Abstract (page v) the authors state that they had ‘conducted a systematic review…and concluded that efforts to change sexual orientation are unlikely to be successful…’. Similarly, in the executive summary, they state (p.3) ‘Thus the results of scientifically valid research indicate that it is unlikely that individuals will be able to reduce same-sex attractions or increase other-sex attractions through SOCE’.
From these statements it appears that, despite their more accurate and much more tentative conclusion on p.28, the authors have not only discovered an absence of evidence [of any effectiveness of these interventions], but they have discovered evidence of absence [of effect]. In other words, they appear to believe that there is, after all, sufficient credible evidence to assert that all interventions (remember their primary research question included all kinds of interventions ranging from behavioural methods, more analytic ‘counselling’ approaches through spiritual exercises) are unlikely to work.
How do the authors reach this conclusion when on p.28 they state that ‘there is little in the way of credible evidence’? It appears that this stronger conclusion that SOCE ‘are unlikely to be successful’ is based on the evidence which they have already cited as not credible on p.28, namely a few studies of behavioural interventions only, carried out in the 1960s and 1970s. These older studies are certainly more rigorous than many other (more recent) studies the authors survey in that they at least try to isolate the intervention and have a comparison group of some sort (although as we have noted only one with just 16 patients had a ‘no treatment’ conventional control group). But they fall well short of the requirements for gold standard research evidence of the RCT.
It is also difficult to see how the authors can generalise from a group of studies of one type of intervention (using behaviour therapy methods), some with eye wateringly small sample sizes, to draw conclusions about the efficacy or otherwise of all SOCE. The authors recognise these issues of ‘generalisability’ in section 3 but then fall into the same trap themselves when they appear to conclude that their findings apply to all kinds of SOCE.
The authors’ treatment of these few behavioural studies is what we earlier described as a ‘narrative review’. This is the weakest approach to sifting evidence that can be used in a systematic review and is reserved for those studies that are too variable and too poorly reported to have their findings merged in a statistical analysis. This kind of narrative is especially susceptible to bias and in this case it is difficult to see how the studies reviewed can be used to support the bold conclusion of ‘no effect’ set out in the abstract and executive summary.
In fact, in the narrative summary, despite the overall conclusion of ‘no effect’, several of the better quality studies cited appear to show some ‘positive’ effects. For example, the authors write that one of the studies - McConaghy and Barr (1973) – ‘found that about half of men reported that their same sex sexual attractions were reduced’ (p. 36). The reviewers rightly point out that, because there were no controls, the experiments ‘cannot address whether men would have changed their sexual arousal pattern in the absence of treatment’ (p. 36). But we have already accepted that these are not gold standard studies and therefore unable to establish causation and the likelihood of effect. The point is that the authors do accept that a proportion of men did change. We can’t have it both ways: either we review these few studies as lower level studies that may show us something, or we stick to our original gold standard criteria and the conclusion that that there is insufficient basis for concluding whether SOCE has any effect on sexual orientation. In summary, this is a confused and confusing set of conclusions and recommendations.
The same problem applies to the APA’s treatment of the question of whether any reported change is maintained over the longer term. In the Executive Summary (p2) it is concluded that ‘enduring change to an individual’s sexual orientation is uncommon’. To allow this conclusion we would need to have identified sufficient ‘high quality’ studies that were designed to measure long term outcomes. Yet none of the behavioural studies cited so often are of sufficient quality to reach a firm conclusion in respect of long term follow up. Indeed the study by McConaghy (1976) reported that 75% of those who received the intervention had ‘reduced’ or ‘no’ same sex behaviour at one year and a 1973 paper by the same author reported 25% reduction sustained over a year (p 38). But as we have observed, these studies simply do not sufficiently overcome problems of sampling bias, attrition in follow up and control of observer bias to allow us to draw confident conclusions from such findings. Nevertheless, these constraints do not prevent the APA authors from stating that enduring change ‘is uncommon’.
In the section on potential ‘harm’ caused by SOCE, the quality of evidence that may be considered and the readiness to attribute causation undergoes something of a shift. The authors again focus on their few ‘more rigorous’ studies and concede that they were not designed to measure harm. Nevertheless, on the basis of 2 out of 46 subjects in the McConaghy and Barr (1973) study who reported ‘severe depression’ and 4 others who reported milder depression, the APA authors feel able to introduce this section with the bold statement ‘nevertheless, these studies provide some suggestion that harm can occur from [emphasis added] these aversive efforts’ (p 41).
Of course harm can occur (implying prevalence) from (implying causation) these efforts. The crucial question though is “what does the evidence actually show?”. Given the evidence presented, it is unclear how a paper purporting to be a systematic review can conclude in the executive summary ‘we found some evidence to indicate that individuals experienced harm from [emphasis added] SOCE’ (p 3). This is a serious claim about causal attribution and within the conventional rules for a systematic review it is impossible to discern how the authors reached this conclusion.
Scientific Objectivity?
Throughout the report the authors pit the objectivity, rigour and precision of ‘science’ and psychology against the subjectivities of religion and ‘values’. In so doing the report ignores the social, philosophical and value systems that the psychological sciences themselves inhabit, especially when the choice of language in the executive summary settles on words such as ‘normal’ and ‘positive’ to describe same-sex sexual attractions, behaviour and orientations (p 2). The authors seem to believe that the ‘scientific’ evidence over which they preside allows them to police the boundaries of ‘normality’ and their apparent ability to attach values (‘positive’) to psychological observations has a degree of confidence that is breath-taking.
Alongside this they also seek to set strict boundaries to the role and ability of theologians and other people working from religious faith-commitments. Thus the resolution proposed and adopted by the APA asserted that ‘those operating from religious/spiritual traditions are encouraged to recognize that it is outside their role and expertise to adjudicate empirical scientific issues in psychology, while also recognizing they can appropriately speak to theological implications of psychological science’ (pp 120-1). This appears quite unreasonably to prevent those of us operating within ‘religious/spiritual traditions’ from subjecting claims about ‘empirical scientific issues’ such as found in this report to any form of critique and demands instead that we simply accept them as some sort of inerrant revelation from which we are left to draw ‘theological implications’.
Two models of therapy?
In addition to their distinction between orientation and orientation identity, a further interesting development within the APA report is the recognition that ‘conflict between psychology and traditional faiths may have its roots in different philosophical viewpoints’ (p 18). The report speaks of some religions giving priority to ‘living consistently within one’s valuative goals’ which they label telic congruence. They note that ‘some authors propose that for adherents of these religions, religious perspectives and values should be integrated into the goals of psychotherapy’. In contrast, they acknowledge that ‘affirmative and multicultural models of LGB psychology give priority to organismic congruence (i.e. living with a sense of wholeness in one’s experiential self)’. Faced with individuals who have strong religious beliefs and who experience ‘tensions and conflicts between their ideal self and beliefs and their sexual and affectional needs and desires’ these two approaches respond in different ways: ‘the telic strategy would prioritize values, whereas the organismic approach would give priority to the development of self-awareness and identity’.
Implicit in the Report’s discussion of these as ‘philosophical differences’ and ‘philosophical assumptions’ is an acceptance, given their understanding of the role of psychologists, that a body like the APA is not ultimately competent to judge between these two contrasting approaches. They express a hope that ‘understanding this philosophical difference may improve the dialogue between these two perspectives’. It may also be that a similar difference – worked out in terms of different understandings of sin and redemption, contrasting emphases on seeking God within through a sense of wholeness and being addressed by God beyond who may unsettle and disturb us, and differing perceptions of the ‘now’ and ‘not yet’ of Christian life this side of the eschaton – can shed light on some of the debates among Christians about how best to understand and respond to same-sex attraction.
Conclusion
Some evangelicals have welcomed the APA report and this is principally for two main reasons:
- The review assumes that client’s religious values must be respected in helping them choose how they want to manage their sexual desires within their chosen identity framework
- The review asserts that the evidence allows for choice in orientation ‘identity’ (although it still does not accept the possibility of change in actual same sex desires).
While these two developments are welcome changes within the APA approach, the serious weaknesses in some of the arguments and claims, especially in the abstract and executive summary, must not be ignored. As we have seen, there are no ‘gold standard’ Randomised Control Trial data to support the reviewers’ robust conclusions that SOCE are unlikely to be effective. Furthermore, the narrative review sections covering the ‘lower quality’ evidence that does exist are problematic and leave the reader with questions about the reviewers’ own bias.
Finally, although the authors’ review of the relative stability of patterns of sexual desire across the lifecourse is a brief narrative summary, this does not prevent them making some fairly confident assertions about sexual orientation as a fixed and enduring characteristic. Strictly, this is not part of the systematic review at all but needs to be considered as a summary of the authors’ personal opinions about what the research is saying. It is in fact quite possible to cite a range of studies that find evidence of plasticity in the patterning of sexual desire over the life course, as opposed to the assertions of fixity made in this report. (For further information see Glyn Harrison’s contribution on The Witness of Science in The Anglican Communion and Homosexuality edited by Phil Groves, SPCK, 2008, introduced here). The reader is left once again with questions about the rigour and even-handedness of this section.
Given the deficits we have considered in their approach to the evidence, the authors’ reluctance (and indeed refusal) to recommend further cautious research into potential benefits and harm of SOCE is a matter of concern.
As we await such further research to be undertaken and subjected to rigorous review, what should Christians say and do in this contentious area? Although written nearly fifteen years ago, the analysis and advice of the St Andrew’s Day Statement from 1995 remains highly relevant here as in other areas. In relation to the various claims about homosexuality from psychology and other scientific disciplines that present themselves as definitive the authors wisely reminded readers:
The interpretation of homosexual emotion and behaviour is a Christian "task", still inadequately addressed. "Guided by God's Spirit", the church must be open to empirical observation and governed by the authority of the apostolic testimony. According to this testimony the rebellion of humankind against God darkens our mind and subverts our understanding of God and creation (Acts 26.18; Rom 1.19-32; Eph. 4.17-19). For the biblical writers the phenomena of homosexual behaviour are not addressed solely as wilfully perverse acts but in
generalised terms, and are located within the broader context of human idolatry (
The Revd Dr Andrew Goddard is Tutor in Christian Ethics at Trinity College, Bristol, editor of the theological journal Anvil and on the Leadership Team of Fulcrum. Among his contributions to Christian debates about sexuality is True Union in the Body? and the introduction (with Phil Groves) to The Anglican Communion and Homosexuality
Professor Glynn Harrison is Norah Cooke Hurle Professor of Mental Health in the Department of Psychiatry, University of Bristol. He is a consultant psychiatrist interested in early interventions for young adults and evidence-based practice. Currently President of the International Federation of Psychiatric Epidemiology, he is also a Diocesan Lay Minister and a member of the General Synod of the Church of England.
