Establishing the Validity of the Affirmative Training Inventory: Assessing the Relationship between Lesbian, Gay, and Bisexual Affirmative Training and Students’ Clinical Competence

Authors

  • Thomas Stone Carlson,

  • Christi R. McGeorge,

  • Russell B. Toomey


  • Thomas Stone Carlson, PhD, and Christi R. McGeorge, PhD, are faculty members in the Human Development and Family Science Department at North Dakota State University; Russell B. Toomey is a Postdoctoral Researcher in the Prevention Research Center at Arizona State University.

Address correspondence to Thomas Stone Carlson, Human Development and Family Science Department at North Dakota State University, Dept. 2615, PO Box 6050, Fargo, North Dakota 58108-6050; E-mail: tom.carlson@ndsu.edu

Abstract

This study established the validity and factor structure of the Affirmative Training Inventory (ATI; T. S. Carlson, C. R. McGeorge & M. Rock, unpublished) as a measure of lesbian, gay, and bisexual (LGB) affirmative clinical training. Additionally, this study examined the latent associations among the subscales of the ATI and the Sexual Orientation Counselor Competency Scale (Bidell, 2005) utilizing a sample of 248 master’s and doctoral couple and family therapy students. The findings from this study provide empirical support for the relationship between specific classroom-related content associated with LGB affirmative therapy and students’ perceptions of their own ability to work competently with LGB clients. This study also found a positive association between the degree to which couple and family therapy programs adopt a LGB-affirmative stance and students’ beliefs, knowledge, and skills associated with competent therapy with LGB clients.

The available data suggest that lesbian, gay, and bisexual (LGB) identified individuals seek therapy services at a higher rate than heterosexual individuals (Bieschke, McClanahan, Tozer, Grzegorek, & Park, 2000; Bradford, Ryan, & Rothblum, 1994; Cochran, Sullivan, & Mays, 2003; Liddle, 1997). Further, research has found that a significant proportion of couple and family therapists (CFT) report that they work with LGB clients (Green & Bobele, 1994; Henke, Carlson, & McGeorge, 2009). However, it is interesting to note that just over half of the CFTs surveyed by Doherty and Simmons (1996) report that they did not feel competent to work with lesbian and gay clients in therapy. While the Doherty and Simmons’ (1996) study is 15 years old, a more recent study of CFT students supports their findings (Rock, Carlson, & McGeorge, 2010). In particular, Rock et al. (2010) found that student therapists reported relatively low levels of clinical skills related to working with LGB clients. Thus, while LGB clients appear to be seeking the services of CFTs, it appears that CFTs may not be prepared to provide competent services to this group of clients.

Scholars and educators have argued that CFT training programs have failed to prepare students to competently work with LGB clients (Bepko & Johnson, 2000; Bernstein, 2000; Godfrey, Haddock, Fisher, & Lund, 2006; Long, 1996; Long & Serovich, 2003). In particular, CFT programs have failed to incorporate LGB affirmative training practices, which have been shown to predict students’ clinical competence with LGB clients (Rock et al., 2010). These assertions are supported by a recent study which found 60.5% of students surveyed reported that they received no training whatsoever on LGB affirmative therapy practices (Rock et al., 2010). Furthermore, scholars have argued that in addition to a lack of training to work with LGB clients, heterosexual biases are pervasive in mainstream CFT theories, practices, and approaches to training (Godfrey et al., 2006; Long & Serovich, 2003). This lack of affirmative training and the heterosexual biases that students may be regularly exposed to leaves students ill-equipped to work with LGB clients, let alone provide competent and affirmative services to this population. This overall lack of training and preparation led McCann (2001) to observe that LGB clients “are essentially at the mercy of the therapist’s own struggles, prejudices, and intolerance” (p. 80).

While there is a consistent call in the literature for CFT training programs to better prepare students to work with LGB clients, the type of training that is recommended involves moving beyond simply teaching about LGB topics and instead encourages students to reflect on their heterosexual biases and develop a positive belief system about LGB individuals and relationships (Godfrey et al., 2006; Long & Serovich, 2003; Matthews, 2007; Phillips & Fischer, 1998; Rock et al., 2010). Assisting students in exploring heterosexual biases entails both helping students become aware of the unconscious biases and heteronormative assumptions that they hold about sexual orientation as well as teaching students about the concept of heterosexual privilege and its influence on the therapy process (Matthews, 2007; McGeorge & Carlson, 2011). Additionally, preparing students to work with LGB clients involves helping them gain an awareness of the negative impact that living in a heterosexist society has on the lives and relationships of LGB individuals, couples, and families (Matthews, 2007; McGeorge & Carlson, 2011; Meyer, 1995; Phillips, 2000). Thus, adopting LGB affirmative training practices is essential in preparing students to competently work with LGB clients.

The importance of integrating LGB affirmative training practices into CFT programs has been clearly articulated in the literature (Godfrey et al., 2006; Long & Serovich, 2003; Matthews, 2007; Rock et al., 2010); however, there is limited research about the degree to which CFT programs integrate affirmative training practices into their curriculum. More specifically, a formal mechanism for measuring the degree to which clinical training programs are incorporating LGB affirmative training practices has not previously existed. Therefore, the purpose of this study was to explore the utility of the Affirmative Training Inventory (ATI; T. S. Carlson, C. R. McGeorge & M. Rock, unpublished). In particular, given the relationship established previously between LGB affirmative training and clinical competence (Rock et al., 2010), this study aims to examine the validity and factorial structure of a new measure of LGB affirmative training and a measure of sexual orientation counselor competency (i.e., Sexual Orientation Counselor Competency Scale, SOCCS; Bidell, 2005). Additionally, we sought to examine latent associations among the subscales of a LGB affirmative training scale and the SOCCS. Given the gender differences that have been documented in the literature related to homophobia and attitudes toward LGB individuals (Davies, 2004; Gormley & Lopez, 2010; Herek, 1988; Nagoshi et al., 2008), the third aim of this study was to examine whether the measurement of these two scales is equivalent across gender (i.e., do men and women have the same concept of LGB affirmative training and sexual orientation counselor competency?). Apart from the basic need to demonstrate the construct validity of the ATI (T. S. Carlson, C. R. McGeorge & M. Rock, unpublished) and the SOCCS (Bidell, 2005), demonstrating measurement equivalence is critical for accurate interpretations of results for full samples using these measures as well as any resulting gender differences. Use of structural equation modeling (SEM), specifically confirmatory factor analysis (CFA), to complete these aims is superior to more traditional methods because this method takes into account measurement unreliability leading to more accurate estimates of factorial structure and associations among constructs (Kline, 2005).

Method

Participants

Master’s and doctoral-level CFT students were recruited to participate in this study through e-mails sent to all of the directors of CFT programs accredited by the Commission on the Accreditation of Marriage and Family Therapy Education (COAMFTE). The e-mails asked the program directors to forward information about the study to all of their students. The e-mails received by the students included a link to an electronic version of the survey used for this study. Paper copies of the survey were also distributed to students at an annual conference of the American Association for Marriage and Family Therapy. Stamped envelopes addressed to the principal investigator for this study were attached to each paper copy of the survey.

The sample is comprised of 248 participants, with 71.1% identifying as master’s students and 28.3% as doctoral students. The mean age of the participants was 29.5 years (SD = 7.8), with a range of 21–61 years. The majority of the sample identified as women (76.6%), White (80.2%), and heterosexual (87.5%). In terms of the type of educational institution participants reported attending, 64.4% reported attending a public university, 10.9% a private non-religious university, and 24.7% a private religious university. Finally, our sample is consistent with existing data describing the student population for COAMFTE accredited programs. In particular, COAMFTE reports that 79% of student trainees are White, which is similar to our sample (Northey, 2004). In regard to gender, previous studies of AAMFT student members have reported percentages of female students ranging from 66% to 77%, which is similar to the gender make-up of the sample for this study (Inman, Meza, Brown, & Hargrove, 2004; Polson & Nida, 1998; Schultz & Leslie, 2004).

Measures

Affirmative training inventory (ATI).  The ATI (T. S. Carlson, C. R. McGeorge & M. Rock, unpublished) measures the extent to which CFT programs integrate LGB affirmative training practices into their educational processes. In particular, the ATI assessed the following: (a) course material on LGB topics, heterosexism, heterosexual bias, heterosexual privilege, and affirmative therapy practices, (b) opportunities for personal reflection of the students’ own heterosexual biases and privileges, and (c) opportunities to work with LGB clients in therapy. The ATI is comprised of nine items and used a 6-point Likert scale ranging from 1 (strongly disagree) to 6 (strongly agree). The alpha coefficient for the ATI was .86.

The ATI is comprised of two subscales: Classroom and Program (See Table 1 for a list of the items comprising each subscale). The Classroom subscale included six items, which assessed students’ perceptions of the type and level of LGB affirmative content that they were exposed to in their coursework. The Cronbach’s alpha for the Classroom subscale was .87. To create a structurally stable latent construct, three parcels were created with the item-to-construct balance approach (Little, Cunningham, Shahar, & Widaman, 2002).

Table 1. Items Comprising the Affirmative Training Inventory
Coursework subscale
 Content related to the experiences of LGB individuals is specifically addressed in each of my family therapy courses.
 I learned about the presence of heterosexual bias (i.e., the act of conceptualizing human experiences in heterosexual terms, thereby discounting LGB experiences and relationships) in my family therapy training program.
 I learned about the concept of heterosexism (i.e., a belief system supported by laws and societal customs that legitimizes heterosexuality as the only acceptable way of being which leads to the unequal treatment of LGB individuals) in my family therapy training program.
 I learned about the concept of heterosexual privilege (i.e., the unearned advantages given to heterosexual individuals based solely on their sexual orientation) in my family therapy training program.
 The faculty in my family therapy program encourage students to explore their own heterosexual biases (i.e., the act of conceptualizing human experience in heterosexual terms, thereby discounting lesbian, gay, and bisexual lifestyles and relationships).
 My program provides students with information on LGB affirmative therapy (i.e., an approach to therapy that embraces a positive view of LGB identity and relationships and addresses the negative influences that homophobia and heterosexism have on the lives of LGB clients) through readings, lectures, supervision, etc.
Program subscale
 The faculty in my family therapy program would be supportive of students pursuing research on topics related to LGB individuals, couples, and/or families.
 My program provides students with the opportunity to work with LGB clients.
 My program takes an affirmative (i.e., a positive view of LGB identity and relationships) stance toward LGB individuals and relationships.

The Program subscale included three items, which were designed to measure the extent to which students perceive that their training program adopts a LGB affirmative stance and provides experiences engaging with the LGB community. The Cronbach’s alpha for this subscale was .67. These items were used as the three indicators for the latent construct.

Sexual orientation counselor competency scale (SOCCS).  A revised version of the SOCCS was used to assess CFT students’ self-reported competency in working with LGB clients (Bidell, 2005). In its original form, the SOCCS was comprised of 29 items and three subscales measuring attitudes, skills, and knowledge related to therapy with LGB clients. The SOCCS appears to demonstrate good reliability and validity. Specifically, the overall scale had an alpha coefficient of .90, and the alpha coefficients for the three subscales were as follows: Awareness (α = .88), Skills (α = .91), and Knowledge (α = .76; Bidell, 2005). Additionally, the 1-week test–retest reliability of the SOCCS was .84 (Bidell, 2005).

For this study, five items from the SOCCS were modified to be more relevant to a student population. Below the original item is listed first followed by the text for the revised item:

Original Item: I check up on my LGB counseling skills by monitoring my functioning/competency—via consultation, supervision, and continuing education.
Revised Item: I know where to find resources to enhance my therapy skills when working with LGB clients by monitoring my functioning/competency—via consultation, supervision, and continuing education.

Original Item: I have experience counseling gay male clients.
Revised Item: I have had the opportunity to work with gay male clients in therapy.

Original Item: I have experience counseling lesbian and gay couples.
Revised Item: I have had the opportunity to work with lesbian and gay couples in therapy.

Original Item: I have experience counseling lesbian clients.
Revised Item: I have had the opportunity to work with lesbian clients in therapy.

Original Item: I have experience counseling bisexual (male or female) clients.
Revised Item: I have had the opportunity to work with bisexual (male or female) clients in therapy.

The five modified items were all part of the Skills subscale, and the alpha coefficient for the Skills subscale in this study is comparable to the alpha reported in Bidell (2005). Additionally, to better measure students’ beliefs about all members of the LGB community, two items (i.e., 28, 31) were added to the SOCCS to assess participants’ beliefs about bisexual clients. For example, item 28 asks participants the degree to which they agree with the following statement “Personally, I think bisexuality (both female and male bisexuality) is a mental disorder and/or a sin and can be treated through therapy or spiritual help.”

Additionally, the 7-point Likert scale used in the original version of the SOCCS was converted to a 6-point scale ranging from 1 (strongly disagree) to 6 (strongly agree) in order to remove the neutral or middle point in the scale. The alpha coefficients for the revised subscales were as follows: Awareness (α = .96), Skills (α = .89), and Knowledge (α = .67). The alpha coefficient for the overall revised SOCCS was .90. Finally, similar to the approach used with the ATI, we used the item-to-construct balance approach to create three parcels for each of the three latent constructs of the SOCCS (Little et al., 2002).

Plan of Analysis

We first examined bivariate associations among the subscales of the ATI and the SOCCS. After establishing initial associations among the constructs of interest, we explored the measurement and factorial structure of each of the constructs using CFA in LISREL (Jöreskog & Sörbom, 2006). First, we utilized CFA to examine the mean and covariance structure of the ATI: Examining the Classroom and Program subscales. Second, we examined the measurement properties of the SOCCS subscales using CFA. Third, we included the subscales of both the ATI and SOCCS in our final model to examine the covariation among subscales. Finally, we examined a multiple-group CFA to explore potential gender differences in the means and covariance structures (MACS; Little, 1997).

We evaluated each model using the following established fit indices: Chi-square test, the root-mean-squared error of approximation (RMSEA), the comparative fit index (CFI), and the nonnormed fit index (NNFI; Kline, 2005). RMSEA values of ≤.05 are considered to be an excellent fit, and values of ≤.08 are considered to be an acceptable fit. CFI and NNFI values of ≥.95 are considered to be an excellent fit, and values of ≥.90 are considered to be an acceptable fit. To establish invariance between groups in the multi-group CFA, we established configural, weak, and strong levels of invariance (Little, 1997). Tests of invariance utilized nested model comparisons where a non-significant change in chi-square suggests that the restricted model is just as good as and more parsimonious than the original, more variant model. A significant change in chi-square suggests that one of the constraints in the model is not supported and therefore cannot be equivalent across groups.

Prior to data analysis in LISREL, all missing data points (<7%) were imputed using the expectation maximization algorithm in PRELIS, a component of LISREL (Jöreskog & Sörbom, 2006). This process minimizes the exclusion of study participants because of missing data and maximizes statistical power (Graham, Cumsille, & Elek-Fisk, 2003).

Results

Descriptive Findings

Correlations between the subscales of the ATI and SOCCS can be found in Table 2. There appears to be a significant positive association between the Classroom subscale of the ATI and the Skills subscale of the SOCCS for both women and men. While there appears to be a significant positive association between the Classroom subscale of the ATI and the Knowledge subscale of the SOCCS for women, there appears to be a negative association between the same subscales for men. However, these associations will be explored further through use of multi-group CFA.

Table 2. Correlations Among Key Study Constructs
 12345
  1. Note. N = 248. ***< .001. **< .01. *< .05. Correlations of variables among female participants are shown under the diagonal and correlations for male participants are shown above the diagonal. ATI, Affirmative Training Inventory; SOCCS, Sexual Orientation Counselor Competency Scale.

ATI: Classroom.32*−.22−.35*.33*
ATI: Program.45***−.05−.17.14
SOCCS: Awareness.05.11.16.22
SOCCS: Knowledge.25**−.03.20*−.08
SOCCS: Skills.44***.13.16*.24**

Confirmatory Factor Analysis: ATI

The first model had excellent model fit: χ2(df = 8) = 10.56, > .05; RMSEA = .04 (90% CI: 0.00–0.09); NNFI = 0.99; CFI = 0.99. These fit statistics suggest that no modifications are necessary for measurement of the two subscales of the ATI. Factor loadings and the covariance estimation between the subscales are presented in Figure 1. As noted in Figure 1, reports of receiving LGB affirmative training in the classroom were positively associated with reports of experiencing an overall LGB affirmative stance at the program level.

Figure 1.

 Confirmatory Factor Analysis of the Affirmative Training Inventory. Standardized estimates are shown. All loadings are significant (p < .001).

Confirmatory Factor Analysis: SOCCS

The first attempt to examine the factor structure of the SOCCS resulted in poor model fit: χ2(df = 24) = 139.86, < .001; RMSEA = .15 (90% CI: 0.13–0.18); NNFI = 0.72; CFI = 0.81. Given the poor fit of the model, we examined an alternative subscale structure to the SOCCS that only includes two subscales. These two new subscales include: (a) the original Awareness subscale (Bidell, 2005) and (b) a new subscale that combines the items from the original Knowledge and Skills subscales. Combining the Knowledge and Skills subscales makes sense at a conceptual level, as some of the items comprising the original Skills subscale appear to be measuring participants’ knowledge about LGB-related topics (e.g., LGB identity development), and the items comprising the initial knowledge subscale have been associated in the literature with the ability (i.e., skill) to provide LGB affirmative therapy (Godfrey et al., 2006; Long & Serovich, 2003; Matthews, 2007; Phillips & Fischer, 1998; Rock et al., 2010).

The Awareness subscale included 12 items (Bidell, 2005). The item-to-construct balance approach was used to create three parcels for the Awareness subscale (Little et al., 2002). The Knowledge/Skills subscale included 16 items (see Table 3). Preliminary analyses revealed that one of the original 19 items negatively loaded onto the Knowledge/Skills subscale and two of the original items did not significantly load onto the construct (these results are available upon request). Therefore, we decided to remove these items from further analyses. The items removed were as follows:

Table 3. Items Comprising the Combined Knowledge/Skills Subscale of the SOCCS
I have received adequate clinical training and supervision to provide therapy to lesbian, gay, and bisexual (LGB) clients.
I know where to find resources to enhance my therapy skills when working with LGB clients by monitoring my functioning/competency—via consultation, supervision, and continuing education.
I have had the opportunity to work with gay male clients in therapy.
At this point in my professional development, I feel competent, skilled, and qualified to provide therapy to LGB clients.
I have had the opportunity to work with lesbian or gay couples in therapy.
I have had the opportunity to work with lesbian clients in therapy.
I am aware some research indicates that LGB clients are more likely to be diagnosed with mental illnesses than are heterosexual clients.
I have received course work that focused on LGB issues in family therapy.
Heterosexist and prejudicial concepts have permeated the mental health professions.
I feel competent to assess the mental health needs of a person who is LGB in a therapeutic setting.
I am knowledgeable about LGB identity development models.
I have had the opportunity to work with bisexual (men or women) clients in therapy.
I am aware of institutional barriers that may inhibit LGB people from using mental health services.
I am aware that therapists frequently impose their values concerning sexuality upon LGB clients.
Currently, I do not have the skills or training to do a case presentation if my client was LGB.
I have done a therapeutic role-play as either the client or therapist involving a LGB issue.

LGB clients receive less competent treatment than heterosexual clients.

Being born a heterosexual person in this society carries with it certain advantages.

I feel that sexual orientation differences between therapist and client may serve as an initial barrier to effectively working with LGB individuals.

The remaining 16 items were also parceled using the item-to-construct balance approach (Little et al., 2002).

The second attempt to examine the SOCCS in CFA resulted in acceptable model fit and was a drastic improvement over the initial model: χ2(df = 8) = 20.65, < .01; RMSEA = .08 (90% CI: 0.04–0.12); NNFI = 0.98; CFI = 0.99. Factor loadings and the correlation between awareness and knowledge/skills can be found in Figure 2. The association between awareness and knowledge/skills suggests that these two constructs are positively associated, but this correlation is small (= .21).

Figure 2.

 Confirmatory Factor Analysis of the Revised Sexual Orientation Counselor Competency Scale. Standardized solution shown. All factor loadings are significant (p < .001); the correlation between Aware and Knowledge/Skills was significant (p < .01).

Confirmatory Factor Analysis: Final Model

To examine the covariation among the four subscales present between the ATI and SOCCS, we explored these two scales in a final CFA. The model fit statistics suggest that this combined model is acceptable: χ2(df = 48) = 126.21, < .001; RMSEA = .08 (90% CI: 0.06–0.10); NNFI = 0.95; CFI = 0.96. Factor loadings and correlations for the final model can be found in Figure 3. In this model, classroom was not significantly associated with awareness (= .02, > .05); however, classroom was significantly associated with knowledge/skills (r = .50, < .001). This suggests that greater levels of affirmative training at the classroom level are associated with greater levels of knowledge/skills; however, the direction of this association cannot be determined because of the cross-sectional nature of this data. Affirmative training at the program level was significantly associated with awareness (= .19, < .01) and with knowledge/skill (= .16, > .05). Thus, greater levels of inclusion of LGB affirmative practices at the program level are associated with students’ reports of greater levels of awareness and knowledge/skills.

Figure 3.

 Associations among the Affirmative Training Inventory and the revised Sexual Orientation Counselor Competency Scale subscales. Standardized solution shown. ***< .001. **< .01. *< .05.

Measurement Invariance and Moderation: The Issue of Gender

We fit a multi-group CFA to test for invariance of measurement across gender. The first step examined configural invariance across groups, the second step examined weak invariance, and the third step examined strong invariance. Each of these steps constrained model parameters to make female and male groups invariant on factor loadings and means. The results of these steps suggest that measurement equivalence exists for both the ATI and SOCCS across gender (see Table 4). After measurement invariance was established, we examined the equality of latent variances, covariances, and means. The tests revealed that women and men did not significantly differ on any of the parameter(s) included in the model (see Table 4). Thus, the ATI and revised SOCCS operate equivalently for women and men.

Table 4. Multiple Group Factorial Invariance Comparisons and Tests of Equivalence of Latent Variances, Correlations, and Means
Modelχ2 df p RMSEANNFICFIConstraint tenable
  1. CFI, comparative fit index; NNFI, nonnormed fit index; RMSEA, root-mean-squared error of approximation.

Configural invariance164.5296<.05.08 ± .020.930.97
Weak invariance169.29104<.05.07 ± .020.960.97Yes
Strong invariance176.91112<.05.07 ± .020.960.97Yes
Modelχ2 df p Δχ2Δdf p Constraint tenable
Homogeneity of variance and covariance182.43122<.055.5310>.05Yes
Equality of variances178.15116<.051.244>.05Yes
Equality of correlations180.30118<.053.406>.05Yes
Equality of means180.30118<.053.406>.05Yes

Discussion

The results of this study support the validity and factorial structure of the ATI as a measurement of LGB affirmative training. While family therapy scholars have long argued the importance of teaching students the content and skills associated with LGB affirmative therapy (Godfrey et al., 2006; Long & Serovich, 2003; Matthews, 2007; Phillips & Fischer, 1998; Rock et al., 2010), there has not been a mechanism for assessing the degree to which such training is being provided. Therefore, the establishment of the ATI as a valid measure provides an important tool for the CFT field to begin to assess the degree to which progress is being made in preparing students to competently work with LGB clients. Moreover, the results of this study suggest that the ATI is an equally valid measure for both women and men and therefore can be used with all students.

This study also sought to explore the latent associations between the subscales of the ATI and the revised SOCCS. The results of this study revealed a significant positive association between the Classroom subscale of the ATI and the Knowledge/Skill subscale of the revised SOCCS. Thus, when students receive more content on LGB affirmative therapy concepts and were provided with more opportunities for self-reflection on their own heterosexual biases, greater levels of self-reported knowledge and skills related to working with LGB clients were reported. This finding is noteworthy as this association has only been argued from a theoretical standpoint in the existing literature. In particular, the findings from this study provide empirical support for the relationship between specific LGB affirmative classroom-related content and practices and students’ perceptions of their own ability to work competently with LGB clients. While the directionality of the relationship between these two factors cannot be determined because of the cross-sectional nature of the data, we hope that this finding provides further encouragement to CFT training programs to be more intentional about integrating specific content and opportunities for reflection on heterosexism, heterosexual privilege, and heterosexual bias into their training of CFT students.

While a positive association existed between the Classroom subscale of the ATI and the Knowledge/Skill subscale of the revised SOCCS, there was not an association between the Classroom subscale of the ATI and the Awareness subscale of the revised SOCCS. There are a number of ways of making sense of the lack of relationship between these two subscales. Given that the Awareness subscale assesses the degree to which participants hold negative or prejudicial beliefs about LGB individuals and relationships, it could be that the amount of training on LGB affirmative therapy does not impact students’ personal beliefs about sexual orientation. However, it is also possible that the sample for this study held such positive views of LGB individuals prior to entering their training programs that the amount of training they received did not affect their beliefs. While on the surface this finding may seem surprising and difficult to explain, the existing literature suggests that most effective way of altering negatively held beliefs about LGB individuals and relationships is through increased personal contact with individuals from the LGB community (Green, 1996; Green & Bobele, 1994), which is not measured by the Classroom subscale.

The analyses for this study also found a positive association between the Program subscale of the ATI and both the Knowledge/Skill and Awareness subscales of the SOCCS. This suggests that the greater degree to which a training program takes a LGB affirmative stance is associated with students’ reports of increased knowledge and skill to competently work with LGB clients. These findings provide empirical support for the importance that taking a LGB affirmative stance as a training program can have on students’ self-reported clinical competency with LGB clients. The positive association between training programs taking a LGB affirmative stance and students’ beliefs about LGB individuals and relationships can be interpreted in a number of ways. It is possible that students who hold positive beliefs about LGB individuals and relationships are more likely to seek out training programs that identify themselves as LGB affirmative. It is also possible that being in a LGB affirmative training environment has a positive influence on students’ beliefs about sexual orientation. This possibility is supported by the literature as researchers have found that the LGB affirmative nature of a clinical agency predicts the level of positive beliefs that therapists hold about LGB clients (Bieschke & Matthews, 1996; Matthews, Selvidge, & Fisher, 2005). It is further possible that as the Program subscale assesses the degree to which training programs provide opportunities for students to work with LGB clients, this positive association between the Program and Awareness subscales could be explained by the level of contact that these students have with members of the LGB community.

In regard to the positive association between the Program and Knowledge/Skill subscales, this finding is supported by the existing literature, as researchers have found that the degree to which a clinical agency is LGB affirmative predicts therapists’ ability to provide LGB affirmative therapy (Bieschke & Matthews, 1996; Matthews et al., 2005). This finding is important because it suggests that preparing students to provide competent services to LGB clients involves more than just the training that takes place in the classroom and in supervision, but involves the development of an overall LGB affirmative environment in the training program. Specifically, this finding highlights the importance of training programs supporting students in researching LGB topics, providing opportunities for students to work with LGB clients, and communicating an affirmative stance toward LGB individuals and relationships.

Limitations

Although the models that we presented here ranged in fit from excellent to acceptable, our sample size was small, which resulted in low statistical power (MacCallum, Browne, & Sugawara, 1996). Thus, future studies should try to replicate our findings with larger and more diverse samples. Another possible limitation of this study involves the possible bias of self-selection in that students with more affirmative LGB beliefs could have been more likely to complete our survey. Additionally, because of the fact we used self-report measures, we were only able to assess students’ perceptions of both their training and clinical competence. Therefore, future studies could seek to assess the relationship between the level of affirmative training as measured by the ATI and students’ actual clinical competence working with LGB clients.

Implications for Future Research

Based on the results of this study, we have several recommendations for future research. As previously stated, one suggestion for future research would be for researchers to replicate this study with a more diverse sample, particularly related to sexual orientation and the type of educational institution in which students are enrolled, as our sample lacked adequate diversity in these areas. For example, future researchers could explore whether the ATI and revised SOCCS are equally valid across all sexual orientations. Researchers could also explore whether accredited CFT programs located in different types of educational institutions (e.g., public university, private non-religious university, and private religious university) differ in the level of LGB affirmative training that they provide to their students. The obvious challenge with this suggestion is recruiting a larger more diverse sample. Future researchers could attempt to recruit this sample by placing an ad in the AAMFT Family Therapy Magazine, accessing listserv that target diverse students (e.g., AAMFT LGBTQIA open forum group), and mailing paper copies of the survey to all student members of the AAMFT.

Given that our study found that the overall affirmative nature of training programs was associated with students reporting higher level of competence working with LGB clients, another possible area for future research would be to survey accredited CFT programs to explore the type of LGB affirmative policies that programs have and how those policies have impacted their curriculum and training processes. Finally, another area of future research would be to survey CFT faculty to determine the level of LGB affirmative training that they feel they offer students in their training programs.

Implications for Clinical Training

Given the results of the current study, there are a number of implications for CFT training programs and faculty members to better prepare students to work with LGB clients. In particular, our results highlight the importance of developing an overall LGB affirmative program environment. One important way to develop a LGB affirmative environment is through the adoption of official program statements and policies. For example, it seems important for training programs to adopt an official statement that clearly identifies the program as being LGB affirmative and committed to training students to provide competent services to the LGB community. Additionally, such a statement could also highlight that the program is affirming of LGB students and their relationships, and supports the equal rights of LGB couples and families. After creating such a statement, program faculty need to post the statement on the official program website and place that statement in their student handbook. Another important statement programs need to consider adopting in order to foster a LGB affirmative environment is a statement of opposition to the practice of conversion therapy. The rationale behind this recommendation is based on the existing research that documents the many harmful effects of conversion therapies on LGB clients (Serovich et al., 2008). Additionally, conversion therapy is based on the belief that LGB sexual orientations are inherently pathological, which is inconsistent with official policies of all major mental health organizations, including the AAMFT (2009).

In addition to developing these statements, programs should also develop specific official policies to establish a LGB affirmative environment. For example, it might be important for programs to develop a policy that requires students to attend campus-supported LGB ally trainings (e.g., Safe Zone, Safe Space, Safe Harbor, ALLIES, and Safe on Campus). Another official program policy that faculty may want to adopt is that homophobic language and behaviors will not be tolerated. Additionally, it seems important for programs to develop a policy that outlines a remediation process for students who hold homophobic beliefs that impact their ability to provide competent services to LGB clients (Long & Serovich, 2003). This remediation process could involve requiring students to do self-of-the-therapist work to intentionally reflect on their attitudes and beliefs about sexual orientation, observe LGB affirmative therapy sessions, and participate in focused supervision that teaches them how to use the specific skills associated with LGB affirmative therapy (Long & Serovich, 2003).

In addition to creating a LGB affirmative training environment, the results of this study highlight the need for CFT programs to include specific content on heterosexism, heterosexual privilege, and heterosexual bias throughout the curriculum. In particular, the results of this study indicate the need for training programs to move beyond simply teaching about LGB topics and specifically focus on helping student therapists critically reflect on the beliefs that they hold about sexual orientation and LGB individuals, couples, and families. A process of critical self-reflection for heterosexual therapists and suggestions for self-reflection activities can be found in McGeorge and Carlson (2011).

Another critical training implication from our study is the importance of providing students with the opportunity to work directly with LGB clients. This means that programs have a responsibility to develop intentional strategies for recruiting LGB clients. An important first step in this process is to develop relationships with organizations that serve their local LGB community. These relationships can be fostered by actively participating as a program and clinic in events sponsored by the LGB community (e.g., Pride, Coming out week, regional LGB conferences). In addition, CFT programs need to develop marketing materials for their clinics that contain their non-discrimination statement, which should include sexual orientation (Long & Serovich, 2003). These marketing materials also need to state that LGB affirmative services are provided and LGB individuals and relationships are encouraged to seek services. Additionally, brochures, flyers, and websites for CFT training clinics need to include images of same-sex relationships. Once these marketing materials are developed, they can be placed in local LGB newspapers, websites, and magazines.

A final implication for clinical training involves the need for faculty to support and promote student research on LGB topics. The process of promoting student research on LGB topics can begin in research method courses by including existing LGB research as examples of various research methodologies. This legitimizes LGB research as valid and may encourage students to consider doing such research for their thesis or dissertation. It may be important for research method courses to also include content on recruitment strategies for researching LGB populations and introduce students to queer theory as a legitimate research methodology (Oswald, Blume, & Marks, 2005). It is also essential that students have access to journals that regularly publish articles related to LGB topics and populations (e.g., Journal of GLBT Family Studies, Journal of Gay & Lesbian Mental Health, Journal of Feminist Family Therapy, Journal of LGBT Issues in Counseling). Finally, another way to promote student research on LGB topics is for faculty to do their own research on LGB populations or topics related to heterosexism and affirmative therapy in order to model the importance and value of researching such topics.

Implications for Clinical Practice

The results of this study can also provide insight into implications for clinical practice. A clear implication from this current study is the importance of therapists holding affirmative beliefs about LGB individuals and relationships to practice competent LGB therapy. This requires that therapists not only commit to providing competent therapy but they also need to commit to exploring their own heterosexism, heterosexual privilege, and heteronormative assumptions (McGeorge & Carlson, 2011). Our findings further highlight the need for postgraduates to seek out LGB affirmative therapy training, as the literature suggests that it is unlikely that they received such training in their graduate studies (Rock et al., 2010). Therapists can seek out this training from a number of different sources and, according to the AAMFT Code of Ethics, have an ethical responsibility to do so (AAMFT, 2001, Principle 3.1). For example, therapists can attend workshops on LGB affirmative therapy or LGB topics at the national AAMFT conference or the national conferences of other professional organizations. Additionally, therapists can attend regional and national conferences on LGB affirmative therapy and mental health (e.g., Creating Change Annual Conference, LGBT Affirmative Psychotherapy Annual Conference). Therapists can also educate themselves through readings on LGB affirmative therapy (e.g., Handbook on LGBT Affirmative Counseling and Psychotherapy, LGBT Affirmative Therapy, Queer Counseling and Therapy, Handbook of LGBT Affirmative Couple and Family Therapy). Moreover, therapists have a responsibility to become familiar with the Guidelines for Psychological Practice with Lesbian, Gay, and Bisexual Clients (American Psychological Association, 2011) and the LGBT competencies detailed by the Association for Lesbian, Gay, Bisexual, and Transgender Issues in Counseling, a division of the American Counseling Association (Association for Lesbian, Gay, Bisexual and Transgender Issues in Counseling, 1999). Another source for receiving LGB affirmative therapy training would be to seek out a supervisor who is LGB affirmative and has experience working with LGB clients.

Conclusion

This study established the validity and factor structure of the ATI (T. S. Carlson, C. R. McGeorge & M. Rock, unpublished) as a measure of LGB affirmative clinical training. Additionally, this study examined the latent associations among the subscales of the ATI and the revised SOCCS (Bidell, 2005). The findings from this study provide empirical support for the relationship between specific classroom-related content and practices associated with LGB affirmative therapy and students’ perceptions of their own ability to work competently with LGB clients. In particular, the more content students receive on LGB affirmative therapy concepts and the more opportunities they were provided with for self-reflection on their own heterosexual biases were associated with greater levels of self-reported competency working with LGB clients. This study also found a positive association between the degree to which CFT programs adopt a LGB affirmative stance and students’ beliefs, knowledge, and skills associated with competent therapy with LGB clients. Our hope is that this study provides CFT training programs with ideas about how to enhance their ability to better prepare CFT students to provide competent and affirmative services to LGB clients. Additionally, it is our hope that this study will encourage the COAMFTE to provide clearer guidelines for CFT training programs regarding the inclusion of LGB affirmative content and practices in the standard curriculum.

Ancillary