Abstract
There is consensus that providers who work with transgender and gender nonconforming (TGNC) older adults should use the language preferred by the older adult; however, assessment tools to document professional training efforts are lacking. This study presents data on a new TGNC Language Self-Efficacy Scale, using responses collected from employees and volunteers of area agencies on aging (AAA) across the United States (N = 155). Exploratory factor analysis supports a two-factor model: Interactions and Information subscales significantly correlated with knowledge of TGNC terminology, genderism and transphobia, gender role beliefs, beliefs about gender and gender identity, and attitudes toward lesbians and gay men. Future directions include additional measurement development work and replication of these findings outside of AAAs.
The proportion of older adults is rapidly increasing in the United States, with approximately 20% of citizens projected to be 65 years or older by the year 2040 (Administration on Aging, 2016). Transgender and gender nonconforming (TGNC) older adults tend to be an unconsidered minority among the aging population. Any estimation, however, is unlikely to reflect the actual number of TGNC older adults due to the fear of persecution that maintains these individuals’ perceived need for anonymity (Witten, 2003, 2009); This risk of discrimination, faced by all TGNC individuals extends to older TGNC adults seeking aging-focused services. Social services (e.g., case management for at-risk seniors, transportation, and meals programs), including those services offered by area agencies on aging (AAA), have thus far focused little on TGNC aging issues (Witten, 2014). Although the gerontology literature reflects a genuine focus on aging lesbians and gay men, there remains a dearth of information on issues facing TGNC older adults and those who aim to serve them (Fredriksen-Goldsen, Hoy-Ellis, Goldsen, Emlet, & Hooyman, 2014; Porter et al., 2016).
Contained within the key competencies and strategies for providing culturally competent service with lesbian, gay, bisexual, and transgender (LGBT) older adults is the assertion that language is inclusive and used appropriately (Burch, 2008; Fredriksen-Goldsen et al., 2014). Using inaccurate language (i.e., use of incorrect pronouns and name) may communicate negative attitudes or beliefs toward those with whom one is working (Shelton & Delgado-Romero, 2011). TGNC individuals report feeling stigmatized when misgendered (McLemore, 2015), and older TGNC adults may attribute discrimination to gender identity more so than age (White Hughto & Reisner, 2016). Relevant professional competencies include information about gender identity and sexual identity, along with understanding these as distinct concepts that are intertwined. In addition, it is important for providers to know the differences between sex and gender as well as the meaning of the terms “transgender” and “gender nonconforming” (Fredriksen-Goldsen et al., 2014). Understanding these terms and using them appropriately communicates understanding and respect for older TGNC adults and are appropriate targets for continuing education programs within agencies.
There is no information on whether aging-focused professionals have the requisite knowledge of terminology and self-efficacy for using gender affirming language (Fredriksen-Goldsen et al., 2014; Porter et al., 2016). Self-efficacy, as defined by Bandura, is the level of confidence one possesses that one can perform a given behavior in a specific domain (Bandura, 1977, 1982, 1997). Self-efficacy affects one’s ability to persist at a specific behavior, despite obstacles. These beliefs may derive from vicarious learning experiences, during which an individual observes another individual successfully perform the given behavior (Bandura, 1977, 1982, 1997).
Related to the lack of research on whether professionals are competent in working with TGNC older adults, there are currently no measures aimed at assessing their levels of confidence for affirmative interactions. Such measures are important in both needs assessments and continuing education efforts. The current study describes preliminary support for a measure of self-efficacy in using TGNC-related language by employees and volunteers at AAA across the United States. The goals of this study were to explore factor structure, reliability, and preliminary validity for this new scale.
Method
Participants
Participants were employees or volunteers in AAAs across the United States (N = 155). The current study was conducted within the context of a larger study of educational approaches for providers within AAAs. All assessment and intervention tasks were conducted online, which allowed for nationwide recruitment. The project was advertised through direct emails to administrators and directors of AAAs. Emails to AAA directors and administrators included the invitation to participate in a “free professional education opportunity” and a brief description of the study as an attempt to evaluate “different educational strategies for how to provide more affirmative services to older transgender and gender nonconforming adults.” All participants were asked how many TGNC individuals they believe they interact with in professional and personal contexts, most (98.06%) responded to these questions. Overall, 65.13% indicated contact with TGNC individuals in professional contexts, and 67.11% reported contact in personal contexts.
The participants represented 32 states, ranged in age from 21 to 84 years (M = 45.19 years, SD = 13.89), and the majority identified as women (85.2%), White/Caucasian (82.6%), and Christian (62.6%). Participants were primarily full-time AAA employees (87.1%) with an average of 7.05 years at their current agency. Most participants reported their highest level of education as college (45.2%) or graduate school (44.5%). Participants represented a range of disciplinary backgrounds: social work (29.0%), psychology (12.3%), business administration (10.3%), education (5.8%), medicine (3.2%), law (1.3%), and other disciplines (38.1%).
Procedure
Individuals who met study criteria (i.e., an employee or volunteer within the agency) and provided informed consent for the study then completed demographic/descriptive items and five measures in addition to the TGNC Language Self-Efficacy items. These additional self-report measures were administered to examine preliminary validity of the scale. This project was reviewed by the associated university’s institutional review board (IRB; 810395-2), and approval was given before initiation of data collection.
The TGNC Language Self-Efficacy items were developed for the current study by the first author using self-efficacy theory and aimed to assess participants’ confidence in using affirmative language while conversing with TGNC older adults in a professional setting. The instructions and scaling were based on Bandura’s (2006) recommendations for the measurement of self-efficacy (0-100 scale where higher scores indicated greater belief in their ability to perform the behavior). Ten items were generated both to correspond with components of the online intervention and to reflect domains of knowledge being targeted. These items were reviewed and slightly revised by a panel of three content experts who are published in the area of gender and sexual diversity. The items were also piloted with a small group of mental health professionals to evaluate clarity and ease of response.
Participants also completed the 10-item Knowledge of TGNC Terminology Scale (Israel, Harkness, Delucio, Ledbetter, & Avellar, 2014), which measured objective knowledge of TGNC terminology. Participants were asked to match each term (e.g., gender identity) to its respective definition. One point was provided for each correct answer and the total number of points earned reflected the participant’s total score with a maximum possible score of 10. This scale demonstrated excellent internal consistency in the current study (α = .81).
The 17-item Genderism/Transphobia subscale (e.g., People are either men or women) of the Genderism and Transphobia Scale–Revised (GTS-R; Tebbe, Moradi, & Ege, 2014) assessed negative attitudes toward TGNC individuals and demonstrated excellent internal consistency in the current study (α = .95). Items were rated on a 7-point Likert-type scale with response choices ranging from 1 (strongly disagree) to 7 (strongly agree). Item ratings were averaged to produce GTS-R subscale and overall scale scores with higher scores indicating greater anti-TGNC prejudice.
The Gender Role Beliefs Scale (GRBS; Kerr & Holden, 1996) contained 20 items (e.g., The initiative in courtship should usually come from the man) measuring gender role ideologies (1 = strongly agree, 7 = strongly disagree). Higher scores indicate more feminist gender role beliefs. In the current study, this 20-item scale demonstrated excellent internal consistency (α = .89).
Traditional Beliefs about Gender and Gender Identity Scale (Dasgupta & Rivera, 2006) comprised two subscales: Traditional Beliefs About Gender (TBG; It is important that men appear masculine and that women appear feminine; eight items) and Traditional Beliefs About Gender Identity (TBI; e.g., If a member of my sex made a sexual advance toward me I would feel angry; seven items). The former included eight items that focused on the degree to which individuals endorsed traditional prescriptive gender norms in various life domains (e.g., professional life and physical appearance). The latter included items that measured the degree to which people were invested in emphasizing their heterosexual identity to others and to themselves. These items were rated on a 7-point Likert-type scale (1 = strongly disagree, 7 = strongly agree). Higher scores indicated more traditional beliefs about gender roles and gender identity. This 15-item measure demonstrated excellent internal consistency (α = .88) in the current study. The Gender subscale (α = .84) and Gender Identity subscale (α = .83) also demonstrated good internal consistency.
Attitudes Toward Lesbians and Gay Men–Short Form (Herek, 1988) consisted of five items used to assess attitudes toward lesbians (ATL-S; for example, Lesbians just cannot fit into our society) and five items used to assess attitudes toward gay men (ATG-S; for example, Male homosexuality is a perversion). The items were rated on a 9-point Likert-type scale (1 = strongly disagree, 9 = strongly agree) and four items were reverse-scored. Scoring was accomplished through averaging scores across items for each subscale and these scores could be combined in a single ATLG-S score. This 10-item measure demonstrated excellent internal consistency (α = .90) in this study. The subscales were also examined independently and also demonstrated good internal consistency: ATL-S subscale (α = .71) and ATG-S (α = .89).
Results
An exploratory factor analysis was performed on the 10 TGNC language self-efficacy items using a principal component extraction method and oblimin rotation to allow for correlated factors. The Kaiser–Meyer–Olkin measure of sampling adequacy was 0.83 indicating that the present data were suitable for principal components analysis. Similarly, Bartlett’s test of sphericity was significant (p < .001) indicating sufficient correlation between the variables to proceed to analysis.
Using the Kaiser–Guttman retention criterion of eigenvalues greater than 1.0, a two-factor solution provided the clearest extraction. These two factors accounted for 70.91% of the total variance. Table 1 presents the 10 items, their factor correlations, and communality estimates. Communalities were fairly high for each of the 10 items, with a range of 0.55 to 0.79; thus, all items were retained.
Summary of Items and Factor Loadings for Oblimin Two-Factor Solution for the TGNC Language Self-Efficacy Scale (N = 155).
Note. Boldface indicates highest factor loadings. TGNC = transgender and gender nonconforming.
Factor 1: self-efficacy for interactions (eigenvalue = 5.57) accounted for 55.74% of the variance and had six items; Factor 2: self-efficacy for information (eigenvalue = 1.52) accounted for 15.17% of the variance and had four items. The rationale used in naming these two factors was guided in part by the recommendations of Comrey and Lee (1992) and Rummel (1970) in which sorted factor weights in excess of .65 were used to “drive” the process of labeling and interpreting each factor. The present two-factor model was deemed an appropriate solution because of its conceptual clarity and ease of interpretability.
The Self-Efficacy for Interactions subscale consisted of 6 items and demonstrated strong internal consistency as a pretest measure (α = .83) and posttest measure (α = .89). Results of the Self-Efficacy for Interactions subscale were approximately leptokurtic and negatively skewed. There were two univariate outliers (z > 3.29) representing roughly 1% of the sample. These individuals were not removed from analyses due to small sample size. Instead, mean value substitution was utilized to replace the individuals’ scores. Because normality was not improved through the use of transformations, no transformation was performed prior to the main analyses. The Self-Efficacy for Information subscale consisted of four items and demonstrated good internal consistency as a pretest measure (α = .78) and posttest measure (α = .87). Results of this subscale were approximately leptokurtic and evenly skewed. No outliers were identified. Because normality was not improved through the use of transformations, no transformation was performed.
The Pearson’s product–moment correlations for the variables used in this study are presented in Table 2. As can be seen from this table, both the Interactions and Information subscales of the TGNC Language Self-Efficacy Scale were significantly correlated with all other variables, with comparable associations with most variables.
Means, Standard Deviations, and Intercorrelations for Scores on Outcome and Secondary Measures.
Note. TGNC = transgender and gender nonconforming.
Correlations significant at p < .05. **Correlations significant at p < .01.
The TGNC Language Self-Efficacy Scale: Interactions demonstrated a moderate positive correlation with the Knowledge of TGNC Terminology Scale; individuals reporting lower levels of self-efficacy for affirmative interactions reported more anti-TGNC attitudes and more traditional beliefs about gender and gender identity. The Self-Efficacy: Interactions scale demonstrated a moderate positive correlation with the GRBS (r = .28), indicating that individuals who reported greater self-efficacy also reported more feminist beliefs. The Self-Efficacy: Interactions scale demonstrated a small negative correlation with the ATLG-S (r = −.20) suggesting that individuals who reported lower self-efficacy tended to report more negative attitudes toward lesbians and gay men.
The Self-Efficacy for Information subscale demonstrated a moderate positive correlation with Knowledge of TGNC Terminology (r = .57) suggesting individuals who reported higher levels of knowledge-based self-efficacy also demonstrated higher objective knowledge of TGNC terminology. The Self-Efficacy for Information subscale also demonstrated negative moderate correlations with the GTS-R (r = −.38) and the Traditional Beliefs about Gender and Gender Identity Scale (r = −.44). Individuals reporting a lower level of self-efficacy for TGNC-related knowledge reported more anti-TGNC attitudes and more traditional beliefs about gender and gender identity.
Discussion
This study described initial development of the TGNC Language Self-Efficacy Scale among employees and volunteers within AAAs across 32 states in the United States. Results of an exploratory factor analysis suggest a two-factor solution that elucidates self-efficacy for interactions with TGNC older adults and self-efficacy for understanding TGNC-related terminology. The pattern of intercorrelations with relevant constructs suggests that the general score may be useful for administrators interested in assessing levels of confidence among their staff, particularly when administrators are interested in supporting their staff in implementing new, more affirmative behaviors. The subscale related to knowledge may also be used on its own for a limited focus on service providers’ understanding of TGNC terminology.
Several limitations of the current scale are notable. We began with a small pool of items that were related to a specific professional development intervention and that were all retained following the principal components analysis. These items may not capture the true range of behaviors required for professional competency with TGNC older adults. Developing a larger pool of items to examine along with those in the current scale may allow a more refined selection of normally distributed questions. Furthermore, measurement development work is needed to provide additional support for construct validity. In addition, the data for this study come from a relatively small number (N = 155) of AAA employees and volunteers who read the study description and consented to participate in a project related to “how to provide more affirmative services to older transgender and gender nonconforming adults.” The sample was predominantly White and female, with limited variability on a number of demographic and professional variables. This sample is thus quite vulnerable to selection bias, with the assumption that culturally competent staff may be more comfortable with the study question and thus more likely to participate. Such an interpretation is compatible with the high average confidence ratings for some of the self-efficacy items. The skewed distribution of the Self-Efficacy for Interactions subscale creates the potential of ceiling effects; additional item development is warranted to increase the range of likely responses and identify items with lower average confidence ratings. The Self-Efficacy for Interactions questions may include items that are overly similar in degree of difficulty (i.e., generating similar average confidence ratings). It is also the case that replication and expansion to a larger sample with greater variability in attitudes, knowledge, and experience with TGNC issues may show different response distributions.
Future directions should include utilizing a sample of additional AAAs as well as aging-focused providers outside of AAAs to both aid in further scale development and associated professional development interventions. Furthermore, previous research suggests providers’ attitudes toward sexual minorities (i.e., lesbians and gay men) may predict their intentions to provide affirmative care (Warren, Steffen, & Wayland, 2015). The current study aimed to not confound sexual orientation and gender identity. The correlations in attitudes toward sexual minorities and TGNC individuals suggest, however, that anti-TGNC attitudes and self-efficacy for affirmative interactions may interrelate over time. Future research can investigate whether professional training in affirmative interaction styles also leads to improved attitudes toward TGNC older adults. Aging-focused professionals and older TGNC adults will benefit from continued development and refinement of empirically supported measures and evidence-based professional interventions that support affirmative communication.
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This project was funded via a dissertation award from the Department of Psychological Sciences, University of Missouri-St. Louis.