Professional Documents
Culture Documents
Bru 1234
Bru 1234
Bru 1234
Cirleen DeBlaere
University of Florida
Lehigh University
Bonnie Moradi
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
University of Florida
The present study explored whether 3 existing measures of workplace constructs germane to the
experiences of sexual minority people could be modified to improve their applicability with transgender
individuals. To this end, the Workplace Heterosexist Experiences Questionnaire (WHEQ; C. R. Waldo,
1999); the Lesbian, Gay, Bisexual, and Transgendered Climate Inventory (LGBTCI; B. J. Liddle, D. A.
Luzzo, A. L. Hauenstein, & K. Schuck, 2004); and the Workplace Sexual Identity Management Measure
(WSIMM; M. Z. Anderson, J. M. Croteau, Y. B. Chung, & T. M. DiStefano, 2001) were modified to
explicitly address the experiences of transgender individuals. Data from a sample of 263 transgender
individuals were used to evaluate the psychometric properties of the modified measures. Analyses of the
structures of the modified measures (Transgender Forms [TF]) suggested an alternative 2-factor structure
for the WHEQ-TF, but provided support for the previously observed unidimensional structure for the
LGBTCI-TF, and a slightly modified 3-factor structure for the WSIMM-TF. Cronbachs alpha reliability
coefficients for scale or subscale items across the 3 measures were acceptable. Criterion-related validity
was evident in theoretically consistent patterns of correlations between scores on the 3 modified measures
and scores on indicators of job satisfaction and outness. These data provide preliminary support for
transgender-specific versions of measures of 3 key constructs in the sexual minority vocational behavior
research.
Keywords: transgender, workplace harassment, workplace climate, heterosexism, gender identity management
We aimed to address this gap in the present study by examining whether existing measures of the three key constructs
germane to the workplace experiences of sexual minority populationsperceived discrimination, perceived workplace climate, and workplace sexual identity management could be
modified to improve their applicability with transgender individuals. Consistent with Worthington and Whittakers (2006)
suggestion to build on existing measures when appropriate, in
this study, the Workplace Heterosexist Experiences Questionnaire (WHEQ; Waldo, 1999); the Lesbian, Gay, Bisexual, and
Transgendered Climate Inventory (LGBTCI; Liddle, Luzzo,
Hauenstein, & Schuck, 2004); and the Workplace Sexual Identity Management Measure (WSIMM; Anderson, Croteau,
Chung, & DiStefano, 2001) were used as a basis for modification to explicitly address the experiences of transgender individuals. These measures were selected due to their prior use and
psychometric evaluation as operationalizations of the underlying constructs of interest with sexual minority populations. The
WHEQ and WSIMM have been used extensively with LGB
samples, and the LGBTCI has been used with sexual minority
samples that included a few transgender people. In the present
study, factor structure, reliability, and validity evidence for the
modified measures are evaluated with a sample of transgender
individuals to test and extend the applicability of these measures to this understudied sexual minority population.
TRANSGENDER WORKPLACE
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
61
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
62
To assess internal consistency reliability, Cronbachs alpha coefficients are computed for scale or subscale items. We expect that
values will be in the good or excellent range according to Ponterotto and Ruckdeschels (2007) matrix for interpreting Cronbachs alphas for items on the WHEQ-TF (Hypothesis 2a),
LGBTCI-TF (Hypothesis 2b), and WSIMM-TF (Hypothesis 2c).
Finally, to assess criterion-related validity, correlations of
WHEQ-TF, LGBTCI-TF, and WSIMM-TF scores with job satisfaction and level of outness are evaluated. We focus on work,
coworker, and supervisor job satisfaction following prior research
and to allow for the possibility of nuances in interpersonal versus
work-related aspects of job satisfaction (e.g., Badgett et al., 2007;
Waldo, 1999). On the basis of previous research with LGB populations, we hypothesize that job satisfaction correlates negatively
with perceived workplace discrimination (i.e., WHEQ-TF; Hypothesis 3a) and concealment of transgender identity (i.e.,
WSIMM-TF Covering; Hypothesis 3b) and correlates positively
with perceptions of supportive workplace climate (i.e., LGBTCITF; Hypothesis 3c) and disclosure of transgender identity (i.e.,
WSIMM-TF Explicitly Out and Implicitly Out; Hypothesis 3d and
3e). Similarly, we hypothesize that outness correlates negatively
with perceived workplace discrimination (i.e., WHEQ-TF; 4a) and
concealment of transgender identity (i.e., WSIMM-TF Covering;
Hypothesis 4b) and correlates positively with perceptions of supportive workplace climate (i.e., LGBTCI-TF; Hypothesis 4c) and
disclosure of transgender identity (i.e., WSIMM-TF Explicitly Out
and Implicitly Out; Hypothesis 4d and 4e). In addition to considering statistical significance, we use Cohens (1992) benchmarks
for small (r .10), medium (r .30), and large (r .50) effect
sizes to interpret validity coefficients.
Method
Participants. Data from 263 transgender participants were
analyzed in this study. Participants ranged in age from 18 to 68
years old (M 38.28, SD 13.54, Mdn 36.50). Approximately
86% of the sample identified as White, 7% as multiracial, 2% as
Asian American/Pacific Islander, 1% as Hispanic/Latina/o, 1% as
Native American, less than 1% as African American/Black, and
3% as other races or ethnicities. About 7% of the sample identified
as exclusively lesbian or gay, 16% as mostly lesbian or gay, 26%
as bisexual, 14% as mostly heterosexual, 11% as exclusively
heterosexual, 2% as asexual, and 24% as another self-specified
sexual orientation (e.g., queer, pansexual). Approximately 37% of
participants had some college education, 35% had a college degree, 19% had a professional degree, 9% had a high school
diploma, and less than 1% had some high school education. About
41% of participants identified as working class, 39% as middle
class, 13% as upper-middle class, 7% as lower class, and less than
1% as upper class. Participants were also asked to report their
employment status (full time, part time, or not employed) and to
report their job title. About 68% reported full-time and 25%
reported part-time employment; 7% checked the not employed
option but were retained because they provided current job titles
(e.g., laborer, writer, kitchen designer) that suggested that they
were self-employed. Participants reported residing in 37 of the 50
United States; most resided in California (8%), Washington (8%),
Florida (7%), and Michigan (6%); some (7%) resided in Canada.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
TRANSGENDER WORKPLACE
63
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
64
negative (but nonsignificant) correlations with level of and satisfaction with identity disclosure (Anderson et al., 2001).
Outness as transgender. The 10-item Outness Inventory (OI;
Mohr & Fassinger, 2000) assesses the degree to which respondents sexual orientation is known or talked about with different
people in their lives (e.g., parents, siblings, friends, coworkers).
For the present study, the measure was modified to assess transgender identity outness on a 7-point continuum (from 1 person
definitely does not know about you being transgender to 7
person definitely knows about you being transgender, and it is
openly talked about). Item ratings are averaged to yield an overall
score, with higher scores indicating greater levels of outness. OI
items yielded Cronbachs alphas of .87 in samples of LGB people
(Balsam & Mohr, 2007) and bisexual people, including 2% who
identified as transgender (Brewster & Moradi, 2010). In terms of
validity, LGB peoples OI scores were found to correlate positively with involvement in LGB communities (Balsam & Mohr,
2007; Mohr & Fassinger, 2000).
Job satisfaction. Job satisfaction was measured using the
Work, Coworker, and Supervisor subscales of the Job Descriptive
Index (JDI; P. C. Smith, Kendal, & Hulin, 1969; as modified by
Roznoski, 1989). The JDI presents respondents with 18 adjectives
(e.g., boring, fascinating) and the options yes, I dont know (?),
and no to rate the relevance of these adjectives to their work,
coworkers, and supervisors. Following Hanischs (1992) recommendations, negative responses (yes to a negatively valenced
item, no to a positively valenced item) are coded as 0, positive
responses (yes to a positive item, no to a negative item) are
coded as 3, and I dont know responses are coded as 1. Ratings
for the Work, Coworker, and Supervisor items are averaged, with
higher scores indicating higher satisfaction in that domain. In a
sample of LGB individuals, Waldo (1999) found Cronbachs alphas of .90, .91, and .92 for the Work, Coworker, and Supervisor
items, respectively. In that same study, JDI Work, Coworker, and
Supervisor scores correlated negatively with job withdrawal (i.e.,
desire and intention to quit).
Results
Factor structure: Hypotheses 1a1c. We used AMOS 18.0
with maximum likelihood estimation to conduct confirmatory factor analyses (CFAs) to evaluate the fit of the hypothesized models
with the present data. The sample size of 263 exceeded guidelines
for deriving meaningful and interpretable models and fit indices
(e.g., minimum sample size of 200, 510 cases per parameter;
Quintana & Maxwell, 1999). In terms of multivariate normality,
Mardias normalized estimate of multivariate kurtosis values were
95.80, 18.311, and 12.57 for the WHEQ-TF, LGBTCI-TF, and
WSIMM-TF data, respectively, though, it has been noted that
statistically significant multivariate nonnormality may be too conservative a criterion to rely on (Tabachnick & Fidell, 2001).
Maximum likelihood estimation procedures are robust to moderate
multivariate nonnormality, particularly when the data approximate
univariate normality (e.g., Muthen & Kaplan, 1985; Weston &
Gore, 2006). Indeed, we inspected the largest multivariate outliers
(indicated by Mahalanobis distances) and found that removal of
these cases had minimal impact on fit index values and parameter
estimates; thus, all participants were retained. Furthermore, item
distributions across measures met guidelines for univariate nor-
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
TRANSGENDER WORKPLACE
between the two factors was .60 (p .001). Details for this
structure, along with item rearrangements relative to the hypothesized model, are presented in Table 1.
The LGBTCI-TF. To test Hypothesis 1b regarding the unidimensional structure of LGBTCI-TF data, items were constrained
to load onto one Supportive Climate factor as found by Liddle and
colleagues (2004). The initial CFA suggested that this model
provided adequate fit to the data, 2(120, N 263) 574.91, p
.001, CFI .90, SRMR .05, RMSEA .10, 90% CI [.09, .10]);
loadings ranged from .42 to .90. Although the fit of this model was
adequate according to less stringent criteria (e.g., Weston & Gore,
2006), we considered that there may be systematic variance associated with the reverse-coded items that could be modeled and
improve model fit. Thus, following approaches for modeling
method variance (e.g., Wei, Russell, Mallinckrot, & Vogel, 2007),
we included two orthogonal factors representing the variance associated with direction of item coding (i.e., positively coded items
loaded onto one factor, and reverse-coded items loaded onto a
separate factor). The resultant model of a single Supportive Climate factor accounting for method variance provided excellent fit
to the data, 2(150, N 263) 291.31, p .001, CFI .96,
SRMR .03, RMSEA .06, 90% CI [.05, .07]). Standardized
regression weights for items onto the climate factor ranged from
.37 to .91 (see Table 2).
The WSIMM-TF. To evaluate Hypothesis 1c, following the
three-factor model outlined by Anderson et al. (2001), items were
65
set to load on three latent and correlated factors reflecting Covering, Explicitly Out, and Implicitly Out. This initial model provided
a poor fit to the data, 2(51, N 263) 226.90, p .001, CFI
.84, SRMR .13, RMSEA .12, 90% CI [.10, .13]); loadings for
Covering, Explicitly Out, and Implicitly Out subscales ranged
from .52 to .67, from .54 to .88, and from .37 to .80, respectively.
We examined the modification indices to determine areas of misspecification. The largest modification index (i.e., 90.15) suggested that Item 8 (i.e., Openly associate with coworkers known
to be transgender and let others think that I am transgender too, if
they want to), which was constrained to load onto the Implicitly
Out factor, should be allowed to load on the Explicitly Out factor
as well. This modification resulted in good model fit, 2(50, N
263) 105.73, p .001, CFI .95, SRMR .06, RMSEA
.07, 90% CI [.05, .08]), and Item 8 loaded significantly on both
Explicitly Out (loading .68) and Implicitly Out (loading .21)
factors. Indeed, inspection of the item indicates that it is a doublebarreled item tapping both Explicitly Out (i.e., openly associate
with transgender coworkers) and Implicitly Out (i.e., let others
think I was transgender). Given the relative magnitude of these
loadings and to provide a parsimonious solution, we conducted
another CFA with Item 8 loading only on the Explicitly Out factor.
This model resulted in minimal changes to the fit indices and good
fit to the data, 2(51, N 263) 119.51, p .001, CFI .94,
SRMR .06, RMSEA .07, 90% CI [.06, .09]). As indicated in
Table 3, loadings for Covering, Explicitly Out, and Implicitly Out
Table 1
Items and Principle Axis Factor Loadings for the Transgender Form of the Workplace
Heterosexist Experiences Questionnaire
Factor loading
Abbreviated item
3.
4.
7.
5.
9.
6.
13.
8.
10.
20.
14.
15.
11.
1.
12.
18.
19.
21.
16.
17.
22.
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
...
.86
.84
.83
.82
.82
.70
.69
.64
.59
.58
.54
.53
.48
.37
.36
.12
.20
.02
.16
.29
.23
.13
.09
.06
.08
.02
.00
.03
.12
.15
.25
.15
.11
.04
.24
.01
.98
.98
.89
.71
.56
.51
Note. Factor loadings from the pattern matrix are reported (Tabachnick & Fidell, 2001). Factor loadings .30
are in bold. Factor 1 Direct Discrimination; Factor 2 Indirect Discrimination. The correlation between the
two factors was r .60 (p .001). Item numbers reflect the order in which the items appeared in the survey.
As described in the Results section, Item 2 (. . . made transphobic remarks in general) was removed in this
analysis. This structure broadly parallels Waldos (1999) two factors of Direct and Indirect discrimination, with
somewhat different item composition for the two factors: Items 12 and 14 loaded on the Indirect factor in
Waldos structure but loaded on the Direct factor in the present sample, and Item 17 loaded on the Direct factor
in Waldos structure but loaded on the Indirect factor in the present sample.
66
Table 2
Items and Factor Loadings for the Transgender Form of the Lesbian, Gay, Bisexual, and Transgendered Climate Inventory
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
Item
17.
1.
4.
7.
11.
16.
20.
12.
2.
6.
13.
14.
19.
10.
3.
9.
5.
15.
8.
18.
Note.
Factor
loading
Uniqueness
.91
.88
.88
.81
.81
.80
.80
.74
.74
.73
.73
.70
.70
.67
.17
.23
.23
.34
.34
.36
.36
.45
.45
.47
.47
.51
.51
.55
.62
.62
.62
.61
.43
.37
.62
.62
.62
.63
.82
.86
Item numbers reflect the order in which the items appeared in the survey. (R) Item is reverse scored.
subscales ranged from .51 to .68, from .56 to .87, and from .69 to
.79, respectively. Correlations between Covering and Explicitly
Out, Covering and Implicitly Out, and Implicitly Out and Explicitly Out were .60 (p .001), .19 (p .05), and .23 (p .01),
respectively.
Reliability: Hypotheses 2a2c. We calculated Cronbachs
alphas for WHEQ-TF, LGBTCI-TF, and WSIMM-TF scale or
subscale items (based on the present factor analysis results). As
indicated in Table 4, for the WHEQ-TF, Cronbachs alphas were
.92 for Direct and .91 for Indirect items. Cronbachs alpha was .96
for LGBTCI-TF items. For the WSIMM-TF, Cronbachs alphas
were .66 for Covering, .80 for Implicitly Out, and .86 for Explicitly
Table 3
Items, Factor Loadings, and Uniquenesses for the Transgender Form of the Workplace Sexual Identity Management Measure
Factor loading
Abbreviated item
12.
5.
9.
3.
4.
2.
11.
6.
7.
1.
8.
10.
Uniqueness
.87
.82
.79
.68
.56
.54
.66
.69
.74
.38
.41
.52
.24
.33
.38
.54
.69
.68
.58
.56
.51
.79
.77
.69
Note. Factor 1 Covering; Factor 2 Implicitly Out; Factor 3 Explicitly Out. The correlation between Factors 1 and 2 was r .19 (p .05),
between Factors 1 and 3 was r .60 (p .001), and between Factors 2 and 3 was r .23 (p .01). Item numbers reflect the order in which the items
appeared in the survey.
TRANSGENDER WORKPLACE
67
Table 4
Intercorrelations, Descriptive Statistics, and Internal Consistency Reliabilities
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
.60
10
.12
.14
.21
.58
14
2.21
0.96
.86
.44
.58
.13
03
2.07
0.74
.92
.45
.20
03
2.17
0.79
.92
.19
03
1.76
0.64
.84
17
4.27
1.39
.84
.64
.17
.16
.11
.32
.10
.21
.39
.08
.48
.13
.45
.44
.46
.31
.03
04
0.49
0.67
.92
.30
.40
.25
.30
04
0.89
1.11
.91
.38
.43
.36
.54
14
2.69
0.81
.96
.18
.17
.17
.31
14
1.36
0.52
.66
.44
.25
.03
.13
.01
.02
14
2.42
0.95
.80
Note. WHEQ-TF Transgender Form of the Workplace Heterosexist Experiences Questionnaire; LGBTCI-TF Transgender Form of the Lesbian, Gay,
Bisexual, and Transgendered Climate Inventory; WSIMM-TF Transgender Form of the Workplace Sexual Identity Management Measure; JDI Job
Descriptive Index; OI Outness Inventory.
Discussion
In the present study, measures that assess sexual minority peoples perceived experiences of workplace discrimination (WHEQ;
Waldo, 1999), perceptions of supportive workplace climates
(LGBTCI; Liddle et al., 2004), and use of identity management
strategies (WSIMM; Anderson et al., 2001) were modified to
address the experiences of transgender individuals, resulting in the
transgender forms of each of these measures: WHEQ-TF,
LGBTCI-TF, WSIMM-TF. The present analyses of factor struc-
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
68
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
TRANSGENDER WORKPLACE
individuals who are familiar with the Internet and eager to participate in transgender-related research. Another consideration with
Internet surveys is the issue of individuals who respond to a few
items and then drop out. Indeed, as described previously, this
occurred in the present study, but it is possible that some of these
submissions were from individuals who checked the survey to
assess their interest and then returned to participate at a later date.
Due to the anonymity of data collection, we could not determine
the proportion of such cases or whether there were demographic
(or other) differences between survey completers and noncompleters. Moreover, in the present study, the majority of participants
reported having had some college education or beyond (91%),
identified as either working class or middle class (80%), and
identified as White (86%). Given the paucity of nationally representative data on the transgender population, it is unknown
whether the demographics of our sample parallel or diverge from
characteristics of the broader transgender population. Future studies are needed to test the generalizability of our findings with
transgender individuals from more diverse socioeconomic, racial,
and ethnic backgrounds.
Another consideration is the broadness and fluidity of the term
transgender. In describing our inclusion criteria to participants, we
provided a description of this term that we hoped would not only
communicate a shared understanding but also be inclusive enough
to respect participants self-determination with regard to identifying as transgender (i.e., be somewhere on the transgender spectrum, for example, genderqueer, M-to-F, F-to-M, androgynous).
Although all participants affirmed that they identified within this
transgender umbrella, it is possible that individuals with other
understandings of transgender, or those who fall within this umbrella but disidentify with the terms we used, may not have
participated. This underscores the importance (and challenge) of
balancing clarity and inclusiveness in sampling transgender individuals.
An additional direction for future research is to examine the
replicability of the present psychometric findings, particularly
because a few content and structural modifications were made to
the measures assessed in this study. Future evaluations of criterionrelated validity could also include dimensions of job satisfaction
beyond those included in the present study (e.g., satisfaction with
pay and promotion). We hope that the present study provides
useful groundwork for such efforts and for further research and
practice to address the workplace experiences, functioning, and
needs of transgender populations.
References
American Psychological Association, Task Force on Gender Identity and
Gender Variance. (2009). Report of the Task Force on Gender Identity
and Gender Variance. Washington, DC: Author.
Anderson, M. Z., Croteau, J. M., Chung, Y. B., & DiStefano, T. M. (2001).
Developing an assessment of sexual identity management for lesbian
and gay workers. Journal of Career Assessment, 9, 243260. doi:
10.1177/106907270100900303
Badgett, M. V. L., Lau, H., Sears, B., & Ho, D. (2007). Bias in the workplace:
Consistent evidence of sexual orientation and gender identity discrimination. Retrieved from http://www.law.ucla.edu/williamsinstitute
Balsam, K. F., & Mohr, J. J. (2007). Adaptation to sexual orientation
stigma: A comparison of bisexual and lesbian/gay adults. Journal of
Counseling Psychology, 54, 306 319. doi:10.1037/0022-0167.54.3.306
69
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
70
Quintana, S. M., & Maxwell, S. E. (1999). Implications of recent developments in structural equation modeling for counseling psychology. The
Counseling Psychologist, 27, 485527. doi:10.1177/0011000099274002
Roznowski, M. (1989). Examination of the measurement properties of the
Job Descriptive Index with experimental items. Journal of Applied
Psychology, 74, 805 814. doi:10.1037/0021-9010.74.5.805
Schafer, J. L. (1997). Analysis of incomplete multivariate data. London,
England: Chapman & Hall.
Schilt, K. (2006). Just one of the guys?: How transmen make gender visible
at work. Gender and Society, 20, 465 490. doi:10.1177/
0891243206288077
Schilt, K., & Connell, C. (2007). Do workplace gender transitions make
trouble? Gender, Work and Organization, 14, 596 618. doi:10.1111/
j.1468-0432.2007.00373.x
Schlomer, G. L., Bauman, S., & Card, N. A. (2010). Best practices for
missing data management in counseling psychology research. Journal of
Counseling Psychology, 57, 110. doi:10.1037/a0018082
Selvidge, M. M., Matthews, C. R., & Bridges, S. K. (2008). The relationship of minority stress and flexible coping to psychological well being in
lesbian and bisexual women. Journal of Homosexuality, 55, 450 470.
doi:10.1080/00918360802345255
Smith, N. G., & Ingram, K. M. (2004). Workplace heterosexism and
adjustment among lesbian, gay, and bisexual individuals: The role of
unsupportive social interactions. Journal of Counseling Psychology, 51,
57 67. doi:10.1037/0022-0167.51.1.57
Smith, P. C., Kendall, L., & Hulin, C. L. (1969). The measurement of
satisfaction in work and retirement. Chicago, IL: Rand McNally.
Tabachnick, B. G., & Fidell, L. S. (2001). Using multivariate statistics.
Boston, MA: Pearson.
Waldo, C. R. (1999). Working in a majority context: A structural model of
heterosexism as minority stress in the workplace. Journal of Counseling
Psychology, 46, 218 232. doi:10.1037/0022-0167.46.2.218
Wei, M., Russell, D. W., Mallinckrodt, B., & Vogel, D. L. (2007). The
experiences in close relationship scale (ECR)-short form: Reliability,
validity, and factor structure. Journal of Personality Assessment, 88,
187204. doi:10.1080/00223890701268041
Weston, R., & Gore, P. A. (2006). A brief guide to structural equation
modeling. The Counseling Psychologist, 34, 719 751. doi:10.1177/
0011000006286345
Worthington, R. L., & Whittaker, T. A. (2006). Scale development
research: A content analysis and recommendations for best practices.
The Counseling Psychologist, 34, 806 838. doi:10.1177/
0011000006288127