The Gay Affirmative Practice Scale (GAP): a new measure for assessing cultural competence with gay and lesbian clients.
Although these studies yielded helpful information about social workers' attitudes toward gay and lesbian individuals, they ten us little about their social work practice with this population. Several authors have discussed the practice implications of homophobia in social workers, and many claim that homophobia may reduce the effectiveness of services offered to gay and lesbian individuals. Homophobia may thus lead practitioners to provide inferior treatment; minimize or exaggerate the importance of sexual orientation in the gay of lesbian person's life; change the topic when clients talk about gay or lesbian issues; devalue clients' feelings and experiences; deny clients access to a broad range of experiences; view clients strictly in terms of their sexual behavior; assume celibate adults and adolescents cannot identify as gay men or lesbians; inform clients that they are not gay of lesbian because they fail to meet some arbitrarily defined criterion; assume that gay or lesbian relationships are phases clients will move through; or perpetuate self-hatred experienced by some gay and lesbian clients (Brown, 1996; McHenry & Johnson, 1993; Messing, Schoenberg, & Stephens, 1984; Peterson, 1996). At its extreme, homophobia in social workers and other practitioners can lead to the use of conversion or reparative therapies, treatments aimed at changing the sexual orientation of the gay, lesbian, of bisexual person, which are explicitly condemned by NASW, the American Psychological Association (ApA), the American Counseling Association, and the American Psychiatric Association (ApA) (American Academy of Pediatrics, n.d.; ApA, 1998; NASW National Committee on Lesbian, Gay, and Bisexual Issues, 2000).
Despite these assertions, few studies have assessed the relationship between social workers' attitudes and practice with gay and lesbian individuals empirically, although a relationship between the two is generally assumed (Wisniewski & Toomey, 1987). Oles and colleagues (1999) claimed that although attitudes are an important component of practice with gay men and lesbians, other factors also are required for culturally sensitive practice with these individuals. Given these limitations, additional research on social workers' behaviors in practice and beliefs about practice with gay and lesbian individuals is needed. The goal of this study was to develop a two-dimensional scale that would assess both these elements and to examine the relationship between this scale and social workers' attitudes in general toward gay and lesbian individuals.
GAY AFFIRMATIVE PRACTICE
Gay affirmative practice models provide guidelines for treating gay and lesbian individuals. Historically, this approach to practice has been the domain of psychologists with an emphasis on gay affirmative psychotherapy. More recently, social workers such as Appleby and Anastas (1998); Hunter, Shannon, Knox, and Martin (1998); and Hunter and Hickerson (2003) introduced the concept of gay affirmative practice into the social work literature and broadened the model to include the many venues in which social workers are employed.
As defined by Davies (1996), gay affirmative practice "affirms a lesbian, gay, or bisexual identity as an equally positive human experience and expression to heterosexual identity" (p. 25). Tozer and McClanahan (1999) said that affirmative practitioners
celebrate and advocate the validity of lesbian, gay, and bisexual persons and their relationships. Such a therapist goes beyond a neutral or null environment to counteract the life-long messages of homophobia and heterosexism that lesbian, gay, and bisexual individuals have experienced and often internalized. (p. 736)
Thus, an absence of homophobia is not sufficient to practice affirmatively. Rather, affirmative practice requires that practitioners celebrate and validate the identities of gay men and lesbians and actively work with these clients to confront their internalized homophobia to develop positive identities as gay and lesbian individuals.
Gay affirmative practice is well suited for the many settings in which social workers assist clients. According to Appleby and Anastas (1998), "There is no particular approach to psychotherapy or other forms of mental health treatment nor any particular modality of treatment--individual, couple, family, or group--that cannot be made useful for lesbian, gay, or bisexual people if approached affirmatively" (p. 286).
In addition to being applicable across a variety of social work settings, such as case management, substance abuse treatment, child welfare, and private practice, gay affirmative practice is consistent with approaches familiar to many social workers (see Figure 1):
* Person in environment--Gay and lesbian individuals are considered in the context of the many environments in which they interact and the many roles they play. For example, when working with gay and lesbian individuals, affirmative practitioners pay attention to gay men's and lesbians' work and family settings and the degree to which they disclose their sexual orientation to others along with the roles that gay and lesbian individuals play in these environments.
* Strengths perspective--Gay and lesbian individuals are viewed as having many strengths that can assist them in addressing their presenting issue. Where appropriate, affirmative practitioners also use other components of the strengths model, including self-determination, by supporting gay and lesbian individuals in their decisions regarding when and to whom to disclose their sexual orientation (Appleby & Anastas, 1998); a focus on health, not pathology, by viewing identities as gays or lesbians as equally healthy as heterosexual identities (Davies, 1996); and consciousness raising, by encouraging gay men and lesbians to examine the impact of homophobic forces in their lives.
* Cultural competence models--Many of these models suggest that culturally sensitive practice with diverse populations requires a unique knowledge base, set of attitudes and beliefs, and skill base for a given population (Sue et al., 1982).Van Den Bergh and Crisp (2004) asserted that gay affirmative practice is a form of cultural competence, similar to culturally sensitive practice with racial and ethnic minority groups.
[FIGURE 1 OMITTED]
Other aspects of gay affirmative practice should be noted. Appleby and Anastas (1998) discussed six principles of gay affirmative practice:
1. Do not assume that a client is heterosexual.
2. Believe that homophobia in the client and society is the problem, rather than sexual orientation.
3. Accept an identity as a gay, lesbian, or bisexual person as a positive outcome of the helping process.
4. Work with clients to decrease internalized homophobia that they may be experiencing so that clients can achieve a positive identity as a gay or lesbian person.
5. Become knowledgeable about different theories of the coming out process for gay men and lesbians.
6. Deal with one's own homophobia and heterosexual bias.
Hunter and colleagues (1998) presented guidelines for affirmative practice, including, but not limited to, the following: Understand and abide by one's professional code of ethics; value clients' sexual orientations; do not view sexual orientation as the problem; do not attempt to change clients' sexual orientation; support clients in deciding how "out" to be; and do not attempt to identify the cause of clients' sexual orientation, as doing so may be destructive to the client.
Davies (1996) claimed that holding any of the following beliefs precludes practitioners from working affirmatively with gay and lesbian individuals: Homosexuality is sinful or against God's wishes; homosexuality is sick, unnatural, or perverted; homosexuality is inferior to heterosexuality; monogamy is the only healthy way to have a relationship; gay and lesbian relationships can only be short-term, sexual, or lacking depth; gay men and lesbians are more likely to sexually abuse children; gay and lesbian parents are inferior to heterosexual parents; and bisexual individuals can decide to be gay or lesbian or heterosexual. Practitioners who hold these beliefs cannot work with gay men and lesbians in ways that convey respect for them and seek to affirm their identities as gay and lesbian individuals. Concern arises that these beliefs about gay men and lesbians lead to beliefs about practice with gay men and lesbians. For example, a belief that homosexuality is against God's wishes may lead to a belief that gay men and lesbians should be treated for their homosexuality so that they may no longer be sinners; a belief that gay men and lesbians are more likely to sexually abuse children may lead practitioners to treat gay men and lesbians for pedophilia when there is no other indication of such a disorder.
Using clinical measurement theory and the domain sampling method as a guiding framework (Crocker & Algina, 1986; Nunnally & Bernstein, 1994), I used three stages to develop and validate a self-administered scale to assess the degree to which practitioners engage in principles consistent with gay affirmative practice: (1) draft of an initial pool of items, (2) administration of the items to a pool of experts to assess the content validity of the items, and (3) administration of the scale to clinicians to assess the reliability and validity of the instrument and to further reduce the number of items in the scale.
Draft of Initial Pool of Items
One commonly used path to generating items for a new scale is through a review of the relevant literature (DeVellis, 1991). A review of the literature produced 23 articles and books on gay affirmative practice. From this review, it became apparent that the scale should consist of two domains: (1) behaviors in practice with gay and lesbian individuals and (2) beliefs about practice with gay and lesbian individuals (which are distinct from attitudes about gay and lesbian individuals). Across the two domains, 543 items were created from this review. After duplicate items and items that did not appear to assess the behavior or belief domains were eliminated, 372 items remained: 167 in the behavior domain and 205 in the belief domain.
The use of expert reviewers can be invaluable in helping a scale developer test his or her perceptions about how well items capture an intended construct (Springer, Abell, & Hudson, 2002). Accordingly, using a "snowball" sampling method, nine experts on gay affirmative practice were identified and asked to evaluate the 372 items. Consistent with methods described by Lynn (1986) and Waltz and Bausell (1981), reviewers were asked to rate each item for its relevance to the construct using a four-point Likert-type scale (1 = not relevant, 2 = somewhat relevant, 3 = quite relevant, and 4 = very relevant), a method supported by Springer and colleagues (personal communication with D. Springer, PhD, University of Texas at Austin School of Social Work, September 28, 1999). Although the initial intent was to retain items with a mean score greater than or equal to 3.0, doing so would have resulted in a total of 264 items being retained for administration to the sample. Thus, the 35 items with the highest mean scores were retained, except when two items identified for retention assessed similar constructs. When this occurred, one of the two items was eliminated and the item with the next highest mean score was retained. For example, several items in the behavior domain assessed how frequently practitioners assisted clients in dealing with homophobia; the item with the highest mean score was retained. In addition, five items from each domain that were reverse scored were chosen for inclusion in the scale because including reverse-scored items in a questionnaire can reduce acquiescence response (DeVellis, 1991) and increase validity (Torabi & Ding, 1998). Following this step, 80 items were retained across the two domains.
Administration to Clinicians
Sample. Because only those who practice directly with clients were of interest in the study, NASW and APA were asked to randomly select members who met their definition of a "direct practitioner." A total of 3,000 respondents (1,500 from each organization) were selected by these organizations for participation in the study. The decision to sample this number was based on the following:
* Tinsley and Tinsley (1987) recommend that a minimum of five to 10 respondents per item are needed to conduct the factor analysis. Given that the factor analysis was conducted on 80 items, a minimum sample size of 400 usable surveys was needed.
* Earlier research using mailed surveys to assess social workers' attitudes toward gay men and lesbians yielded response rates between 32 percent and 63 percent (Hardman, 1997; Harris, Nightengale, & Owen, 1995). Using this minimum response rate of 32 percent, a return of 960 usable surveys could be expected.
* In view of the terrorist attacks on the World Trade Center and the Pentagon on September 11, 2001, and the subsequent disruption in the mail system when envelopes containing anthrax were sent to members of Congress and other individuals were infected by anthrax, concern arose that respondents might be reluctant to open mail from a source unfamiliar to them. Thus, it seemed wise to expect a much lower response rate than previous studies and to assume that the response rate might be as low as half the lowest response rate noted earlier.
* Given that 400 responses were needed and a minimum response rate of 16 percent was assumed, a minimum sample of 2,500 was needed. To further safeguard against a low response rate, this number was rounded up to a total of 3,000 total packets mailed.
Measurement Package. Along with the 80-item Gay Affirmative Practice Scale (GAP) developed from the expert review stage, a packet consisting of the following items was mailed to each potential respondent:
* The Attitudes Toward Lesbians and Gay Men Scale (ATLG) (short form), developed by Herek (1988), was administered to assess the convergent construct validity of the GAP's behavior domain. The ATLG has been validated with college students and members of gay and lesbian organizations and demonstrated high internal consistency (alpha of .90). Herek found that the scale correlates significantly in the expected direction with attitudes toward sex roles, with traditional family ideology, and with reports of positive contact with lesbians and gay men. Although the research correlating attitudes and behaviors provides mixed results, it was expected that people who held more positive attitudes toward gay men and lesbians would be more affirming in their practice with this population than those who held more negative attitudes toward gay men and lesbians. Thus, the behavior domain of the GAP was expected to correlate significantly with the ATLG.
* The Heterosexual Attitudes Toward Homosexuals Scale (HATH), developed by Larsen, Reed, and Hoffman (1980), was administered to assess the convergent construct validity of the GAP's belief domain. The HATH has been validated with college students, resulting in a split-half reliability of .86, and when corrected with the Spearman-Brown prophecy formula, has a split-half reliability of .92 (Larsen et al.). Larsen and colleagues found that the scale correlated significantly in the expected direction with instruments that assess religious ideology, authoritarianism, and feelings of inadequacy. As with the ATLG and the GAP's behavior domain, significant correlations between the HATH and the GAP's belief domain were expected.
* The Marlowe-Crowne Social Desirability Scale (SDS), developed by Crowne and Marlowe (1960) and reduced to a shorter version by Strahn and Gerbasi (1972) using 10 of the original 33 items, was used to assess discriminant construct validity. The short version of the scale has high internal consistency (.876) and a correlation of .958 with the original scale (Fischer & Fick, 1993). A nonsignificant correlation with the SDS and the entire GAP was expected and would thus provide initial evidence of discriminant construct validity.
* Twenty demographic items were chosen based on the literature review and other items of interest. These items inquired about the respondents' personal characteristics such as gender, race, and sexual orientation; religious and political affiliation; and contact with and feelings about gay men and lesbians.
Mailing Procedure. On January 25, 2002, the University of Texas at Austin University Mailing Services (UMS) distributed the instrument packets to the potential respondents using mailing labels provided by NASW and APA. A cover letter and a self-addressed business reply envelope in which to return the survey were included in the packets. According to UMS, materials should have been received by the respondents no later than February 1,2002. Respondents were requested to return the materials to the researcher by February 15, 2002.
Assessing Reliability. Internal consistency reliability was established through an analysis of responses obtained in the administration to clinicians. Internal consistency reliability is considered a more effective method for computing reliability (Nunnally & Bernstein, 1994) and was thus the most logical method for establishing reliability in this study.
The standard error of measurement (SEM) was also computed for each domain. The SEM is an estimate of the standard deviations of error of measurement and is less influenced by differences in variance and standard deviation in different samples or populations than coefficient alpha (Springer, Abell, & Nugent, 2002).The SEM should be computed to compensate for differences in sample standard deviations. A small SEM provides evidence that the scale is reliable.
Assessing Validity. Content validity was examined and can be thought of as the degree to which items in an instrument represent the construct of interest. It is often established by having independent experts assess whether the items adequately assess the construct (Springer, 1997). Convergent construct and discriminant construct validity were assessed to examine the degree to which the GAP correlated with measures that were theoretically related to it. Construct validity is closely tied to theory and is concerned with theoretical relationships between variables (DeVellis, 1991; Springer). As multidimensional scales are collections of unidimensional scales (Hudson, 1985), convergent construct validity was established for each domain rather than for the scale as a whole, and discriminant construct validity was examined for the entire scale. Confirmatory factor analysis (CFA) using the multiple groups method (Nunnally & Bernstein, 1994) was used to establish factorial validity. This method allows the researcher to examine the correlation between each individual item and each domain in the scale. This in turn allows the researcher to confirm or disconfirm a priori hypotheses about factor loadings. The number of domains was fixed at two and were titled "belief" and "behavior." As a general rule, item correlations are considered moderately high when they load on their intended domain at or about .60 (Nunnally & Bernstein). As discussed by several authors, validating a scale is an ongoing process and does not end on completion of the aforementioned types of analyses.
Of the 3,000 surveys sent to members of NASW and APA, 488 were returned completed, for a response rate of 16.3 percent. Despite the low response rate, this study is one of the largest conducted to examine homophobia in social workers and psychologists. The majority of respondents were women (74 percent), married (69 percent), heterosexual (86 percent), Democrats (69 percent), and white (92 percent). NASW members responded at a slightly higher rate than APA members, with 53 percent of the sample indicating NASW membership. Respondents were almost evenly split between those whose highest degree was a master's (49 percent) or a doctorate (48 percent). Most respondents (59 percent) reported mental health as their primary area of practice. (Additional information about the sample can be found in Table 1).
The majority of respondents reported having at least one gay or lesbian friend (87 percent) and at least one gay or lesbian client (65 percent); only 38 percent reported having a gay or lesbian family member (Table 2). Although most (73 percent) of the respondents reported attending a workshop that included content on gay and lesbian issues, fewer than half (47 percent) reported attending a workshop with a focus on gay and lesbian issues. Respondents' feelings about gay men and lesbians were very positive and appeared to be similar, as indicated by a mean score of 82.19 (SD = 17.85) on the lesbian feeling thermometer and a mean score of 81.27 (SD = 17.52) on the gay male feeling thermometer. Additional information about characteristics of the sample can be found in Crisp (2002).
Reliability and Validity
The two domains of the initial 80-item GAP were examined for internal consistency reliability. Cronbach's alpha was .93 for the belief domain and .95 for the behavior domain. Analysis of the factor loading for each item identified 17 belief items and 29 behavior items that loaded on their respective domain at or above this level. On the basis of the factor analysis, 15 items in each domain were retained for the final version of the scale. (Additional information about the rationale for retaining or deleting specific items can be found in Crisp ).
Final Version of the Gay Affirmative Practice Scale (GAP)
Reliability. The final version of the GAP consists of two 15-item domains (see the Appendix) with an overall Cronbach's alpha of .95. Cronbach's alpha is .93 for the belief domain and .94 for the behavior domain. Both domains well exceed Nunnally's (1978) minimum criteria of at least .70 to demonstrate internal consistency. Based on reliability standards set by Springer, Abell, and Nugent (2002), the reliability for both domains is "very good."
The SEM of 1.91 for the belief domain and 2.71 for the behavior domain provide evidence of the scale's reliability. Both SEM scores meet Hudson's (1999) recommendation that the SEM should be less than 5 percent (6.0 for each of these two domains) of the possible range of scores.
Validity. Evidence of validity was demonstrated using several methods. CFA revealed that each item loads on its intended domain at .60 or greater, providing support for factorial validity. Evidence of convergent construct validity was obtained by examining Pearson's r correlations between scores on the belief domain and scores on the HATH and between the behavior domain and the ATLG. The correlation between the belief domain of the GAP and the HATH was .624 (p = .000); the correlation between the behavior domain and the ATLG was .466 (p = .000).The scale is thus a strong first attempt at assessing gay affirmative practice. Both correlations were significant at the .001 level, are in the expected direction, and fall within the acceptable range of greater than or equal to .40 (Downie & Heath, 1967). Overall, there is evidence that the two domains of the GAP correlate at least adequately with the instruments with which they are expected to correlate, thus providing evidence of convergent construct validity. Evidence of discriminant construct validity was obtained by examining the correlation between the SDS and the entire 30-item GAP scale. The correlation between these two instruments was .021 and was nonsignificant (p = .691). This finding provides strong evidence that the GAP does not measure socially desirable responses and is evidence of its discriminant construct validity. The aforementioned reliability and validity analyses collectively suggest that the GAP measures gay affirmative practice.
A chief limitation of this study is the low response rate. The poor response rate may be partially attributable to a delay in the mail that caused many questionnaires to be received by respondents after the requested return date of February 15, 2002. On their completed questionnaires, 92 of the 488 respondents indicated that they received the packet after the response date. Several other factors such as a lack of time to complete the survey, lack of interest in the topic, or the length of the survey may have contributed to the low response rate. Whatever the cause of the response rate, concern arises that the nonresponders may hold different views from the responders. This concern limits the generalizability of the current study because it is not known to what degree the pool of respondents is representative of social workers and psychologists as a group.
The use of memberships lists from NASW and APA to obtain the sample also limits the study's generalizability because members of these organizations may hold different views than those who are not members of NASW and APA. In addition, social workers' and psychologists' views may not represent the views of many other helping professionals such as nurses and counselors. Generalizability is also limited by the high percent of respondents who identified as Democrat, Caucasian/white, female, married, and heterosexual.
Another limitation is that the study did not examine known-groups criterion validity to determine whether the GAP can distinguish between those who would reasonably be expected to have higher scores on the scale and those who would be reasonably expected to have lower scores on the scale. Conducting such a validity study would further reinforce claims that the scale is a valid measure of gay affirmative practice. Nevertheless, this validation study of the GAP was based on a large number of respondents, and the GAP appears to have sufficient reliability and validity.
UTILITY OF THE GAP
The key issue with any measurement instrument is its utility (Springer, 1997).The GAP has many uses for social workers and other helping professionals. First, as a rapid assessment instrument, the GAP can be quickly and easily administered and scored by a variety of helping professionals in a brief amount of time. Second, practitioners can use the GAP as a self-assessment instrument to evaluate the degree to which they practice affirmatively with gay and lesbian individuals. Third, the scale can be used to assess the effectiveness of different types of educational interventions on practitioners' work with gay and lesbian individuals. Such studies might consist of training on gay and lesbian issues or treatment approaches given to one group of practitioners and withheld from another group and using the GAP to assess the impact of the training. In addition, following test-retest reliability studies, the GAP may be administered to individuals before and after different training methods and content on gay and lesbian issues to evaluate the magnitude of change in each group. Fourth, the GAP can be used to evaluate claims by students and other helping professionals that despite holding antigay attitudes, they can practice affirmatively with gay and lesbian individuals. Although many remain skeptical about these claims, this scale may be used to evaluate such claims and to further identify factors that affect clinicians' practice and beliefs about practice with gay and lesbian individuals.
IMPLICATIONS FOR PRACTICE, EDUCATION, AND RESEARCH
Clinicians who want to improve their practice with gay and lesbian individuals have had few tools with which to evaluate their beliefs and practice with this population. This self-evaluation process is consistent with the ethics of social work as a profession, which encourage competence and respect for diversity. In addition, by evaluating their practice with gay and lesbian individuals, social workers demonstrate a commitment to culturally competent practice with members of this population, consistent with the increasing emphasis in the profession on cultural competence with diverse groups.
The moderate relationship between the domains of this scale and measures of homophobia may suggest that educational efforts targeting attitudes toward gay and lesbian individuals are insufficient to ensure gay affirmative practice. Although attitudes may be an important component of affirmative practice with gay and lesbian individuals, they may not be sufficient to ensure affirmative practice with such clients (Oles et al., 1999). It may be equally, if not more, important to educate students and practitioners about components of gay affirmative practice and ways in which they can apply this model to the many settings in which they practice. As suggested by Van Den Bergh and Crisp (2004), social work education about gay and lesbian individuals might thus address knowledge, attitudes, and skills in practice with gay men and lesbians rather than focusing almost exclusively on the relationship between attitudes and behavior.
For the past 30 years, research has focused largely on social workers' attitudes toward gay and lesbian individuals. Although this research has contributed to the knowledge base, it is time to move the focus from attitudes to an examination of beliefs and behaviors in practice with gay and lesbian individuals. In doing so, the research can move from an assumed relationship between homophobia and practice to one that is empirically tested. The development and validation of the GAP is one step in that direction.
The development and validation of this scale may also encourage others to embark on similar studies. This scale can be used in validation studies for related measures of affirmative practice and culturally competent practice with gay and lesbian individuals. The development of additional measures may facilitate research in this area and increase knowledge about practice with gay and lesbian individuals.
The GAP was developed in an attempt to bridge the gap between attitudes and behaviors in practice with gay and lesbian individuals, and by doing so gain insight into the relationship between the two. Although this study provides initial evidence of the reliability and validity of the GAP, the findings are of little practical significance unless they benefit gay and lesbian individuals who use social work and other clinical services. Gay affirmative practice is increasingly accepted as the model from which to approach practice with gay and lesbian individuals and is consistent with many social work values. Clinicians whose practice is based on this model convey support and affirmation for gay and lesbian individuals' identities, support their right to self-identify, assist them with the challenges of living in an oppressive and homophobic world, help them express positive feelings about gay and lesbian identities, and explicitly reject the use of reparative therapies. Given research that shows gay men and lesbians are more likely than heterosexual individuals to use therapeutic services (Rudolph, 1988), it is important that clinicians have measures by which to evaluate their competence with gay and lesbian individuals and be trained to treat gay men and lesbians affirmatively. The GAP provides an initial means by which to assess practice and, along with other resources on gay affirmative practice, can be used to improve practice with gay and lesbian individuals.
GAY AFFIRMATIVE PRACTICE SCALE (GAP) [c] 2002 Catherine Crisp, PhD
This questionnaire is designed to measure clinicians' beliefs about treatment with gay and lesbian clients and their behaviors in clinical settings with these clients. There are no right or wrong answers. Please answer every question as honestly as possible.
Please rate how strongly with you agree or disagree with each statement about treatment with gay and lesbian clients on the basis of the following scale:
SA = Strongly agree
A = Agree
N = Neither agree nor disagree
D = Disagree
SD = Strongly disagree
1. In their practice with gay/lesbian clients, practitioners should support the diverse makeup of their families.
2. Practitioners should verbalize respect for the lifestyles of gay/lesbian clients.
3. Practitioners should make an effort to learn about diversity within the gay/lesbian community.
4. Practitioners should be knowledgeable about gay/lesbian resources.
5. Practitioners should educate themselves about gay/lesbian lifestyles.
6. Practitioners should help gay/lesbian clients develop positive identities as gay/lesbian individuals.
7. Practitioners should challenge misinformation about gay/lesbian clients.
8. Practitioners should use professional development opportunities to improve their practice with gay/lesbian clients.
9. Practitioners should encourage gay/lesbian clients to create networks that support them as gay/lesbian individuals.
10. Practitioners should be knowledgeable about issues unique to gay/lesbian couples.
11. Practitioners should acquire knowledge necessary for effective practice with gay/lesbian clients.
12. Practitioners should work to develop skills necessary for effective practice with gay/lesbian clients.
13. Practitioners should work to develop attitudes necessary for effective practice with gay/lesbian clients.
14. Practitioners should help clients reduce shame about homosexual feelings.
15. Discrimination creates problems that gay/lesbian clients may need to address in treatment.
Please rate how frequently you engage in each of the behaviors with gay and lesbian clients on the basis of the following scale:
A = Always
U = Usually
S = Sometimes
R = Rarely
N = Never
16. I help clients reduce shame about homosexual feelings.
17. I help gay/lesbian clients address problems created by societal prejudice.
18. I inform clients about gay affirmative resources in the community.
19. I acknowledge to clients the impact of living in a homophobic society.
20. I respond to a client's sexual orientation when it is relevant to treatment.
21. I help gay/lesbian clients overcome religious oppression they have experienced based on their sexual orientation.
22. I provide interventions that facilitate the safety of gay/lesbian clients.
23. I verbalize that a gay/lesbian orientation is as healthy as a heterosexual orientation.
24. I demonstrate comfort about gay/lesbian issues to gay/lesbian clients.
25. I help clients identify their internalized homophobia.
26. I educate myself about gay/lesbian concerns.
27. I am open-minded when tailoring treatment for gay/lesbian clients.
28. I create a climate that allows for voluntary self-identification by gay/lesbian clients.
29. I discuss sexual orientation in a non-threatening manner with clients.
30. I facilitate appropriate expression of anger by gay/lesbian clients about oppression they have experienced.
Scoring instructions: Using the chart below, please give each answer the indicated number of points. After all questions have been answered, add up the total number points. Higher scores reflect more affirmative practice with gay and lesbian clients.
Items 1-15 Items 16-30 Points Strongly agree Always 5 Agree Usually 4 Neither agree nor disagree Sometimes 3 Disagree Rarely 2 Strongly disagree Never 1
Original manuscript received February 20, 2004
Final revision received November 16, 2004
Accepted April 11, 2005
American Academy of Pediatrics. (n.d.).Just the facts about sexual orientation and youth: A primer for principals, educators, and school personnel. Retrieved January 26, 2004, from http://www.apa.org/pi/ lgbc/facts.pdf
American Psychiatric Association. (1998). Gay, lesbian, and bisexual issues fact sheet: APA fact sheet pertinent to gay and lesbian issues. Retrieved January 26, 2004, from http: //www.psych.org/public_info/ gaylesbianbisexualissues22701.pdf
Appleby, G. A., & Anastas, J. W. (Eds.). (1998). Not just a passing phase: Social work with gay, lesbian, and bisexual people. New York: Columbia University Press.
Berkman, C., & Zinberg, G. (1997). Homophobia and heterosexism in social workers. Social Work, 42, 319-332.
Brown, L. S. (1996). Preventing heterosexism and bias in psychotherapy and counseling. In E. Rothblum & L. Bond (Eds.), Preventing heterosexism and homophobia (pp. 36-58). Thousand Oaks, CA: Sage Publications.
Crisp, C. (2002). Beyond homophobia: Development and validation of the Gay Affirmative Practice Scale (GAP). Dissertation Abstracts International, 64, 07A. (UMI No. 3099441)
Crocker, L., & Algina, J. (1986). Introduction to classical and modern test theory. New York: Holt, Rinehart & Winston.
Crowne, D. P., & Marlowe, D. (1960). A new scale of social desirability independent of psychopathology. Journal of Consulting Psychology, 24, 349-354.
Davies, D. (1996). Towards a model of gay affirmative therapy. In D. Davies & C. Neal (Eds.), Pink therapy: A guide for counselors and therapists working with lesbian, gay and bisexual clients (pp. 24-40). Philadelphia: Open University Press.
DeCrescenzo, T. A. (1984). Homophobia: A study of the attitudes of mental health professionals toward homosexuality. In R. Schoenberg, R. Goldberg, & D. Shore (Eds.), With compassion toward some: Homosexuality and social work in America (pp. 115-136). New York: Harrington Park Press.
DeVellis, R. F. (1991). Scale development: Theory and applications. Newbury Park, CA: Sage Publications.
Downie, N. M., & Heath, R. W. (1967). Basic statistical methods (2nd ed.). New York: Harper & Row.
Fischer, D. G., & Fick, C. (1993). Measuring social desirability: Short forms of the Marlowe-Crowne Social Desirability Scale. Educational and Psychological Measurement, 53, 417-424.
Fishbein, M., & Ajzen, I. (1975). Belief, attitude, intention, and behavior: An introduction to theory and research. Reading, MA: Addison-Wesley.
Hardman, K. (1997). Social worker's attitudes to lesbian clients. British Journal of Social Work, 27, 545-563.
Harris, M., Nightengale, J., & Owen, N. (1995). Health care professionals' experience, knowledge, and attitudes concerning homosexuality. Journal of Gay and Lesbian Social Services, 2, 91-107.
Herek, G. M. (1988). Heterosexuals' attitudes toward lesbians and gay men: Correlates and gender differences. Journal of Sex Research, 25, 451-477.
Hudson, W. W. (1985). Indexes and scales. In R. M. Grinnell, Jr. (Ed.), Social work research and evaluation (2nd ed., pp. 185-205). Itasca, IL: F. E. Peacock.
Hudson, W. W. (1999). Measuring personal and social problems: Methods for scale development. Unpublished manuscript.
Hudson, W. W., & Ricketts, W. A. (1980).A strategy for the measurement of homophobia. Journal of Homosexuality, 5, 357-372.
Hunter, S., & Hickerson, J. C. (2003). Affirmative practice: Understanding and working with lesbian, gay, bisexual, and transgender persons. Washington, DC: NASW Press.
Hunter, S., Shannon, C., Knox, J., & Martin, J. I. (1998). Lesbian, gay, and bisexual youths and adults: Knowledge for human services practice. Thousand Oaks, CA: Sage Publications.
Larsen, K. S., Reed, M., & Hoffman, S. (1980).Attitudes of heterosexuals toward homosexuality: A Likert-type scale and construct validity. Journal of Sex Research, 16, 245-257.
Lynn, M. R. (1986). Determination and quantification of content validity. Nursing Research, 35, 382-385.
McHenry, S. S., & Johnson, J. W. (1993). Homophobia in the therapist and gay or lesbian client: Conscious and unconscious collusions in self-hate. Psychotherapy, 30(1), 141-151.
Messing, A. E., Schoenberg, R., & Stephens, R. K. (1984). Confronting homophobia in health care settings: Guidelines for social work practice. In R. Schoenberg, R. Goldberg, & D. Shore (Eds.), With compassion toward some: Homosexuality and social work in America (pp. 65-74). New York: Harrington Park Press.
National Association of Social Workers National Committee on Lesbian, Gay, and Bisexual Issues. (2000, January 21). Position statement: "Reparative" and "conversion" therapies for lesbians and gay men. Retrieved November 19, 2003, from http://www. naswdc.org/PRAC/LGB/reparative.html
Nunnally, J. C. (1978). Psychometric theory (2nd ed.). New York: McGraw-Hill.
Nunnally, J. C., & Bernstein, I. H. (1994). Psychometric theory (3rd ed.). New York: McGraw-Hill.
Oles, T. P., Black, B. M., & Cramer, E. P. (1999). From attitude change to effective practice: Exploring the relationship. Journal of Social Work Education, 35, 87-100.
Peterson, K. J. (1996). Preface: Developing the context: The impact of homophobia and heterosexism on the health care of gay and lesbian people. In K. J. Peterson (Ed.), Health care for lesbians and gay men: Confronting homophobia and sexism (pp. xix-xxii). New York: Harrington Park Press.
Rudolph, J. (1988). Counselors' attitudes toward homosexuality: A selective review of the literature. Journal of Counseling and Development, 67, 165-168.
Springer, D. W. (1997). Development of the Adolescent Concerns Evaluation (ACE) rapid assessment instrument: Detecting risk indicators of runaway behavior in adolescents (Doctoral dissertation, Florida State University, 1997). Dissertation Abstracts International, 61(8), 3355.
Springer, D. W., Abell, N., & Hudson, W. W. (2002). Creating and validating rapid assessment instruments for practice and research: Part 1. Research on Social Work Practice, 12, 408-439.
Springer, D. W., Abell, N., & Nugent, W. (2002). Creating and validating rapid assessment instruments for practice and research: Part 2. Research on Social Work Practice, 12, 768-795.
Strahn, R., & Gerbasi, K. C. (1972). Short, homogeneous versions of the Marlowe-Crowne Social Desirability Scale. Journal of Clinical Psychology, 28, 191-193.
Sue, D. W., Bernier, J. E., Durran, A., Feinberg, L., Pedersen, P., Smith, E. J., & Vasquez-Nuttal, E. (1982). Position paper: Cross cultural counseling competencies. Counseling Psychologist, 10(2), 45-52.
Tinsley, H. E. A., & Tinsley, D. J. (1987). Uses of factor analysis in counseling psychology research. Journal of Counseling Psychology, 34, 414-424.
Torabi, M. R., & Ding, K. (1998). Selected critical measurement and statistical issues in health education evaluation and research. International Electronic Journal of Health Education, 1, 26-38.
Tozer, E. E., & McClanahan, M. K. (1999).Treating the purple menace: Ethical considerations of conversion therapy and affirmative alternatives. Counseling Psychologist, 27, 722-742.
Van Den Bergh, N., & Crisp, C. (2004). Defining culturally competent practice with sexual minorities: Implications for social work education and practice. Journal of Social Work Education, 40, 221-238.
Waltz, C. F., & Bausell, R. B. (1981). Nursing research: Design, statistics, and computer analysis. Philadelphia: F. A. Davis.
Wisniewski, J. J., & Toomey, B. G. (1987). Are social workers homophobic? Social Work, 32, 454-455.
Catherine Crisp, PhD, is assistant professor, School of Social Welfare, University of Kansas, 1545 Lilac Lane, Lawrence, KS 66049; e-mail: email@example.com.
Table 1: Characteristics of Social Workers Completing GAP Survey (N = 488) n % Relationship status Single 47 9.6 Married 338 69.3 Divorced 36 7.4 Widowed 16 3.3 Living with long-term partner 22 4.5 Long-term relationship but not living together 7 1.4 Sexual orientation Heterosexual 420 86.1 Bisexual 17 3.5 Gay/lesbian 48 9.8 Race/ethnicity African American/black 12 2.5 Asian American 5 1.0 Caucasian/white 449 92.0 Hispanic/Latino 5 1.0 Mexican American 4 0.8 Native American 1 0.2 Puerto Rican 5 1.0 Other 3 0.6 Current religious affiliation Baptist 15 3.1 Catholic 79 16.2 Episcopal 27 5.5 Fundamentalist 4 0.8 Lutheran 8 1.6 Methodist 30 6.1 Presbyterian 29 5.9 Conservative Jewish 8 1.6 Orthodox Jewish 13 2.7 Reform Jewish 43 8.8 Jewish NOS 9 1.8 Other 107 21.9 None 107 21.9 Current political party Democrat 337 69.1 Republican 42 8.6 Independent 57 11.7 Green 5 1.0 Libertarian 1 0.2 Reform 2 0.4 Other 8 1.6 None 29 5.9 Note: NOS = not otherwise specified. Table 2: Respondent Contact with and Training about Gay Men and Lesbians (N = 488) n Minimum Maximum Range M SD Number of gay or lesbian friends 453 0 50 50 5.71 7.18 Number of gay or lesbian family members 466 0 10 10 0.61 1.04 Number of gay or lesbian clients 455 0 75 75 2.93 6.34 Percent of clients who are gay or lesbian 436 0 100 100 7.42 12.99 Number of workshops with specific focus on gay or lesbian issues 467 0 40 40 1.66 3.43 Number of workshops with content on gay or lesbian issues 413 0 50 50 4.25 6.50