Joumal of Counseling Psychology 2014, Vol. 61, No. I, 169-175

© 2013 American Psychological Associatíon 0022-0167/14/$12.00 DOI: 10.1037/a0034593

BRffiF REPORT

Locus of Control, Minority Stress, and Psychological Distress Among Lesbian, Gay, and Bisexual Individuals Larry W. Carter II and Debra Mollen

Nathan Grant Smith

Texas Woman's University

University of Houston

Within the framework of minority stress theory, lesbian, gay, and bisexual (LGB) individuals are conceptualized as members of a minority group defined by sexual orientation. Two of the component processes of minority stress hypothesized by Meyer (2003), intemalized heterosexism and the experience of prejudice events, were examined in the current study. Both intemalized heterosexism and the experience of prejudice events have been associated with increased psychological distress in LGB populations. Researchers have also observed a relationship between extemal locus of control and increased psychological distress in general population samples. The current study explored whether locus of control served as a moderator in the relationship between the overall psychological distress of LGB individuals and both intemalized heterosexism and the experience of workplace-based prejudice events {n = 165). Results indicated that locus of control served as a moderator in the relationship between experience of workplace-based prejudice events and overall psychological distress but not for the relationship between intemalized heterosexism and distress. Keywords: minority stress, intemalized heterosexism, prejudice, locus of control

Research indicates that lesbian, gay, and bisexual (LGB) individuals are particularly vulnerable to certain stress-sensitive psychological disorders and behaviors (Cochran & Mays, 2009). These include an increased risk for mood and anxiety disorders (Cochran, Sullivan, & Mays, 2003), alcoholism and substance use (Cochran, Ackerman, Mays, & Ross, 2004), high-risk sexual behavior in gay men (Dudley, Rostosky, Korfhage, & Zimmerman, 2004), and suicidality (Remafedi, 2002). These findings have been interpreted as the manifestations of the negative impact of societal prejudice against LGB individuals (Garnets & Kimmel, 2003) and have been conceptualized using the minority stress hypothesis (Meyer 1995, 2003). Meyer (1995) initially operationalized LGB-related minority stress as composed of three processes; intemalized homophobia, perceived stigma, and the experience of prejudice events. Meyer (2003) later expanded this operationalization to include a fourth process: identity concealment. As a result of Meyer's attempts to operationalize LGB-related minority stress, a large body of re-

search emerged examining each of these processes and their association with negative psychological outcomes. The majority of this research has focused on two of these four processes: intemalized homophobia (e.g., Szymanski & Chung, 2002; Szymanski, Kashubeck-West, & Meyer, 2008) and the experience of prejudice events (e.g., Herek, 2009; Mays & Cochran, 2001). Consistent with the extant research, the current study limits its examination of minority stress to these two processes, intemalized homophobia (hereafter referred to as intemalized heterosexism) and the experience of prejudice events, whose association with negative psychological outcome is well documented within the minority stress literature.

Internalized Heterosexism and Prejndice Events Meyer and Dean (1998) defined intemalized heterosexism as the decreased self-regard and intrapsychic confiict that result from LGB individuals' inward directing of negative societal attitudes toward same-sex attraction. Research has linked intemalized heterosexism to delays in LGB identity development (Peterson & Gerrity, 2006; Rowen & Malcolm, 2003), increased risk of mood and anxiety disorders (Igartua, Gill, & Montoro, 2003), increased risk of alcohol and substance use (Diplacido, 1998), high-risk sexual behavior in men (Meyer & Dean, 1995), and reduced quality of romantic relationships (Szymanski et al., 2008). As members of a minority group, LGB individuals contend with experiences of sexual orientation-related prejudice and discrimination. Research suggests that both everyday experiences of discrimination (Swim, Pearson, & Johnston, 2007) and more severe forms of sexual orientation-related bias crime (Herek, 2009) are

This article was published Online First November 4, 2013. Larry W. Carter II and Debra Mollen, Department of Psychology and Philosophy, Texas Woman's University; Nathan Grant Smith, Department of Educational Psychology, University of Houston. This article is an adaptation of Larry W. Carter IF s doctoral dissertation. We thank Sally Stabb, lenelle Fitch, and Shannon Scott for their invaluable guidance and feedback on an earlier version of this article. Correspondence conceming this article should be addressed to Larry W. Carter II, 850 Crawford Parkway #3114, Portsmouth, VA 23704. E-mail: Larry CarterPhD @ gmail.com 169

CARTER, MOLLEN, AND SMITH

170

widespread. Existing research has linked these experiences to increased levels of depression (Herek, Gillis, & Cogan, 1999), anxiety (Swim, Johnston, & Pearson, 2009), and increased prevalence of psychiatric disorders (Mays & Cochran, 2001). Waldo (1999) cogently identified the workplace as an ideal place to study experiences of prejudice events. Subsequent research has supported its utility. Consistent with Waldo's (1999) observation that experiences of workplace-related prejudice events were related to psychological distress, health-related problems, and decreased job satisfaction. Smith and Ingram (2004) found that experiences of workplace heterosexism were positively related to levels of depression and psychological symptoms. Further, Lyons, Brenner, and Fassinger (2005) found that LGB employees' experiences with workplace-based heterosexism accounted for 13% of the variance in job satisfaction. Although data suggest a strong link between these minority stress processes and increased psychological distress, results have not been universally consistent. In their review of the research, Szymanski et al. (2008) noted that these inconsistencies might refiect the operation of a moderating or mediating variable or variables. Additional research has since supported the utility of examining third-variable moderating or mediating effects (e.g., Kashubeck-West & Szymanski, 2008; Span & Derby, 2009). These findings illustrate that even with regard to outcomes that have been consistently shown to be associated with minority stress, there is value in examining third-variable relationships. Given the importance of minority stress in the larger arena of stress and coping, one noteworthy finding within the literature is the moderating role that perception of control can play with regard to the impact of Stressors on overall psychological distress (Hay & Diehl, 2010; Neupert, Almeida, & Charles, 2007). As such, examining perceptions of control within the context of minority stress is a promising next area of inquiry.

Locus of Control Rotter (1990) defined locus of control as the degree to which individuals believe the outcomes of their behavior are determined by either intemal or extemal factors. Numerous general population studies in the United States and other Westem cultures have consistently linked locus of control with psychological distress. For example, an external locus of control has been associated with increased depression (Presson & Benassi, 1996), increased anxiety (Molinari & Khanna, 1981), and decreased psychological adjustment (Rotter, 1966). LGB-specific research examining locus of control has associated an intemal locus of control with increased self-esteem (Anderson, 1998), increased life satisfaction (Isikoff, 1983), and more positive gay identity development (Binks, 1993). Data emerging from existing research on the relationships among locus of control, intemalized heterosexism, experiences of workplace-based prejudice events, and overall psychological distress of LGB individuals indicate that such an examination may offer a unique contribution to the understanding of the relationships among these variables. The results of several studies provide evidence that perception of control may infiuence the relationship between psychological distress and health Stressors, coping strategies, stressful experiences, and experiences of racism and discrimination. Ong, Bergeman, and Bisconti (2005) found that perception of control moderated the relationship between daily stress

and anxiety, whereas Hay and Diehl (2010) found that perception of control moderated the relationship between health Stressors and physical symptoms. Likewise, Elfstöm and Kreuter (2006) found that intemality of locus of control predicted number of coping strategies, which in tum predicted psychological adjustment. Additional research has indicated that an increased perception of control may serve as a buffer against the negative impact of both stressful experiences (Hay & Diehl, 2010; Roddenberry & Renk, 2010) and experiences of prejudice and discrimination (Cadinu, Maass, Lombardo, & Frigerio, 2006; Pieterse & Carter, 2010). To date, the moderating role of locus of control in the relationship between minority stress and psychological outcome in an LGB sample has not yet been examined. However, the results of research examining locus of control as it applies to LGB individuals have generally associated an increasingly intemal locus of control with improved psychological adjustment (Anderson, 1998; Binks, 1993; Isikoff, 1983).

Hypotheses On the basis of the above literature, we hypothesized that locus of control would serve as a moderator in the relationship between (a) intemalized heterosexism and overall psychological distress and between (b) the experience of workplace-based prejudice events and overall psychological distress of LGB individuals.

Method Participants and Procedure Participants were recmited through a variety of means using electronic, print, and traditional paper flyer modalities. Participant recmitment material included a link to an online data collection. In retum for their participation, participants had the opportunity to register for a chance to win a drawing for a $100 incentive. Participant data are presented in Table 1. In addition, the range, mean, and standard deviation for participants' age were also calculated. Participants' ages in the current sample ranged from 18 to 82 years old (M = 37.89, SD = 12.21). Eligibility requirements for participation in the original study required that participants (a) were at least 18 years of age; (b) identified as having either a male or a female gender identity congment with their biological sex; and (c) identified their sexual orientation as lesbian, gay, or bisexual. Participants who did not meet these eligibility requirements were excluded from final data analysis. A total of 234 individuals accessed the study. Of these 234 individuals, 27 participants accessed the study but did not provide any data. This resulted in a data set of 207 participants. Of these 207 participants, 20.3% (n = 42) were excluded from final data analysis. Of these 42 individuals, .5% (n = 1) of these 207 participants did not meet the sexual orientation criterion, 1.9% in = 4) did not meet the gender identity criterion, 2.4% (n = 5) provided a gender/sexual orientation combination (e.g., biological males who self-identified as lesbian) that yielded their data uninterpretable, and an additional 15.5% in = 32) were excluded from final data analysis due to a large percentage of missing data in their responses. Specifically, 26 participants were excluded due to a lack of responses to over 40% of the survey questions, not allowing for the calculation of overall scores for several of the scales.

LOCUS OF CONTROL AND PSYCHOLOGICAL DISTRESS

Table 1 Frequencies and Percentages of Categorical Demographic Variables Variable Biological sex Female Male Sexual orientation Lesbian Gay Bisexual Racial/ethnic identity African American Asian or Pacific Islander Biracial or multiracial European American Hispanic or Latino/a Native American Other Highest education level completed High school Some college Bachelor's degree Master's degree Doctoral-level degree Individual income level Less than $15,000 $15,000-$29,999 $30,000-$44,999 $45,00O-$74,999 $75,000 or more Relationship status In a romantic relationship Not in a romantic relationship Geographic region Northeast Midwest Mid-Atlantic Southeast Southwest Northwest Outside the United States Type of community Rural Suburban Urban

Measures

N

%

65

39.40 60.60

100 54 92 19 10 6 3 125 17 3

171

32.70 55.80 11.50

1

6.10 - 3.60 1.80 75.80 10.30 1.80 0.60

16 50 38 38 22

9.80 30.50 23.20 23.20 13.40

40 24 42 38 19

24.50 14.70 25.80 23.30 11.70

89 75

54.30 45.70

10 26 3 11 109 3 3

6.10 15.80 1.80 6.70 66.10 1.80 1.80

17 78 69

10.40 47.60 42.10

Note. Frequencies not equaling 165 and percentages not equaling 100% reflect missing data.

An additional six participants were excluded because, although they responded to a majority of the survey questions, they did not provide enough responses to receive an overall score for all available scales. The resulting sample consisted of 165 participants. Data analysis was conducted using the resulting sample of 79.7% (n = 165) of the original 207 participants who voluntarily provided informed consent, met all the inclusion criteria for participation, and provided an adequate proportion of responses to survey items. Of the 165 participants who were included in data analysis, 46.7% (n = 77) of participants were recmited as a result of having received an e-mail announcement, and 27.3% (n = 45) were recruited as a result of printed fiyers. Electronic and printed advertisements accounted for 23.0% (n = 38) and 2.4% (n = 4) of the obtained sample, whereas 0.6% (n = 1) did not respond to the item regarding manner of recruitment.

Lesbian and Gay Identity Scale-Homonegativity subscale. The Lesbian and Gay Identity Scale-Homonegativity subscale (LGIS-IH; Mohr & Fassinger, 2000) is a five-item, self-report measure used to assess intemalized heterosexism along a 7-point Likert-type scale ranging from 1 (Disagree Strongly) to 7 (Agree Strongly). Participant responses were used to calculate a single overall mean score ranging from 1 to 7, with higher scores indicating a higher level of intemalized heterosexism. Intemal consistency reliability was measured by a calculation of Cronbach's alpha, which was observed to be .86 in the current study. Workplace Heterosexist Experiences Questionnaire. The Workplace Heterosexist Experiences Questionnaire (WHEQ; Waldo, 1999) is a 22-item, self-report measure used to assess experiences of sexual orientation-based harassment and discrimination in the workplace along a 5-point Likert-type scale ranging from 0 (Never) to 4 (Most of the Time). Participant responses were used to calculate a single summed score ranging from 0 to 88, with higher scores indicating more frequent experiences of heterosexism. Following Smith and Ingram (2004), Waldo's original referent time frame of 24 months for the experience of heterosexist behavior was reduced to 12 months to better refiect the immediacy captured by other measures used in this study. Intemal consistency reliability was measured via calculation of Cronbach's alpha, which was observed to be .93 in the current study. The I-E scale. The I-E scale (Rotter, 1966) is a 29-item, self-report measure used to assess locus of control. Each item presents participants with two statements and asks them to choose which one of the statements better describes what they believe to be true. Participant responses were scored either a 1 or a 0 for each item and then summed into a single overall score ranging from 0 to 23, with higher scores indicating a more extemal locus of control. Internal consistency reliability was measured via calculation of Cronbach's alpha, which was observed to be .76 in the current study. Hopkins Symptom Checklist. The Hopkins Symptom Checklist (Derogatis, Lipman, Rickels, Uhlenhuth, & Covi, 1974) is a 58-item, self-report measure used to assess overall psychological distress along a 4-point Likert-type scale ranging from 1 (Not at All) to 4 (Extremely). Participant responses were used to calculate a single overall mean score ranging from 1 to 4, with higher scores representing increased psychological distress. Because the current study used a sexual minority sample, one question regarding participant level of comfort with the opposite sex was not presented to participants. L. R. Derogatis (personal communication. May 11, 2006) indicated that the removal of this item would not significantly alter the psychometric properties of the instrument. Intemal consistency reliability was measured by a calculation of Cronbach's alpha, which was observed to be .96 in the current study.

Results The first step in data analysis consisted of addressing missing data points within the data set. For this purpose, multiple imputation (MI; Subramanian, 2011) was used. Following MI, each missing data point was replaced by ra > 1 plausible values drawn from the predictive distribution of missing data under the appropriate data model and missing data mechanism. The result was m

CARTER, MOLLEN, AND SMITH

172

completed data sets. Using Rubin's (1987) mie, each ofthe m data sets was analyzed separately and then combined in order to produce an overall estimate and confidence interval that incorporated the uncertainty introduced by missing data. Using the MI method ensured a valid statistical inference in replacement of a nonresponse by a participant and produced unbiased parameter estimates under reasonable assumptions. In order to implement MI for the current study, each tnissing data point was imputed five times. MI was undertaken on the data set including 165 participants, rather than the larger data set because inclusion of the 32 participants with larger amounts of missing data resulted in the data not being missing at random, a requirement for performing MI. Following the replacement of missing data using the MI method, data were tested for outliers using the criterion of greater than 3 times the interquartile range. Extreme outliers were found on both the LGIS-IH (three cases) and WHEQ (five cases), which were removed prior to conducting analyses. Means, ranges, standard deviations, and correlations of the variables of interest are presented in Table 2. As can be seen, higher scores on intemalized heterosexism, experience of workplace-based prejudice events, and locus of control (i.e., a more extemal locus of control) were all associated with greater psychological distress. In addition, more extemal locus of control was positively associated with psychological distress and intemalized heterosexism but not significantly associated with experience of workplace-based prejudice events. To test the hypothesis that locus of control would moderate the relationship between intemalized heterosexism and overall psychological distress, a hierarchical regression equation was computed, with the two main effects (i.e., locus of control and internalized heterosexism) entered as a block in the first step and the Locus of Control X Internalized Heterosexism interaction entered in the second step. Both locus of control and intemalized heterosexism were centered around the mean to avoid multicoUinearity. The model containing the two main effects was significant, F(2, 159) = 18.43, p < .001, R^ = .188. The model adding the interaction to the two rnain effects was not significantly different than the model with only the main effects, AF(1, 158) = 0.223, p = .637, A/?^ = .001. Therefore, there was no evidence of moderation. The regression coefficients for the final step are displayed in Table 3. To test the hypothesis that locus of control would moderate the relationship between experience of workplace-based prejudice

events and overall psychological distress, a hierarchical regression equation was computed, with the two main effects (i.e., locus of control and experience of workplace-based prejudice events) entered as a block in the first step and the Locus of Control X Experience of Workplace-Based Prejudice Events interaction entered in the second step. Both locus of control and experience of workplace-based prejudice events were centered around the mean to avoid multicoUinearity. The model containing the two main effects was significant, F(2, 159) = 27.83, p < .001, R^ = .262. The model adding the interaction to the two main effects was significantly different than the model with only the main effects, Af(l, 156) = 4.787, p < .05, AR^ = .022. As such, there was evidence of moderation. The regression coefficients for the final step are displayed in Table 3. The interaction was probed for significance using the methods outhned by Cohen, Cohen, West, and Aiken (2003). The slope of the line predicting overall psychological distress from experience of workplace-based prejudice events was tested at average, low ( - 1 SD), and high (+1 SD) levels of locus of control. The slope of the line was not significant for low levels of locus of control (ß = .003, t = .532, p = .600). However, the slopes of the lines for average and high levels of locus of control were significant (ß = .010,/ = 2.865, p < . 0 5 ; and ß = .017, í = 3.742, p < .001, respectively). As such, the hypothesis that locus of control would moderate the relationship between experience of workplace-based prejudice events and overall psychological distress was supported. The interaction is displayed in Figure 1.

Discussion Results of this examination revealed that locus of control served as a moderator in the relationship between the experience of workplace-based prejudice events and psychological distress. For participants with lower scores on locus of control (i.e., more intemal locus of control), there was not a significant relationship between experience of prejudice events and psychological distress. However, this relationship was significant for participants with average and high scores on locus of control (i.e., more extemal locus of control). In contrast, results revealed that locus of control did not serve a similar role as a moderator in the relationship between intemalized heterosexism and psychological distress. These results are consistent with other research examining the role of locus of control in the relationship between stress and outcomes. In samples not specifically examining LGB individuals.

Table 2 Pearson Product-Moment Correlations for the Hopkins Symptom Checklist Overall, Lesbian and Gay Identity Scale-Homonegativity Subscale, Workplace Heterosexist Experiences Questionnaire, and I-E Scale Variable 1. Psychological distress 2. Internalized heterosexism 3. Experience of prejudice events 4. Locus of control

M SD Range

**>

Locus of control, minority stress, and psychological distress among lesbian, gay, and bisexual individuals.

Within the framework of minority stress theory, lesbian, gay, and bisexual (LGB) individuals are conceptualized as members of a minority group defined...
6MB Sizes 0 Downloads 0 Views