View in EDS HTML Full Text PDF Full Text

Psychometric Evaluation of the Brief Resilience Scale and Multidimensional Scale of Perceived Social Support among Older Sexual Minority Women in the U.S.

Saved in:
Bibliographic Details
Title: Psychometric Evaluation of the Brief Resilience Scale and Multidimensional Scale of Perceived Social Support among Older Sexual Minority Women in the U.S.
Language: English
Authors: Jordan B. Westcott, Louis M. Rocconi
Source: Measurement and Evaluation in Counseling and Development. 2025 58(1):63-82.
Availability: Routledge. Available from: Taylor & Francis, Ltd. 530 Walnut Street Suite 850, Philadelphia, PA 19106. Tel: 800-354-1420; Tel: 215-625-8900; Fax: 215-207-0050; Web site: http://www.tandf.co.uk/journals
Peer Reviewed: Y
Page Count: 20
Publication Date: 2025
Document Type: Journal Articles
Reports - Research
Descriptors: Psychometrics, Resilience (Psychology), Social Support Groups, Measures (Individuals), Older Adults, Females, LGBTQ People, Disabilities, Factor Structure, Factor Analysis, Test Reliability, Construct Validity, Scores
DOI: 10.1080/07481756.2024.2397947
ISSN: 0748-1756
1947-6302
Abstract: Objective: This study sought to examine the factor structure, internal consistency, and measurement invariance of the Brief Resilience Scale (BRS) and the Multidimensional Scale of Perceived Social Support (MSPSS) among older sexual minority women with disabilities. Method: Participants (n = 208) consisted of sexual minority women aged 55 and older with disabilities recruited online. Results: CFA results showed a one-factor structure fit BRS data, RMSEA = 0.050 (90% CI [0.00, 0.107]), SRMR = 0.015, CFI = 0.990, TLI = 0.975, with good internal consistency (alpha = 0.863, omega = 0.864). For the MSPSS, CFA showed a three-factor structure fit the data, RMSEA = 0.080 (90% CI [0.064, 0.096]), SRMR = 0.037, CFI = 0.959, TLI = 0.947, with excellent internal consistency (alpha = 0.937, omega = 0.937). Full scalar invariance was achieved for both BRS and MSPSS when comparing single and partnered individuals. Additionally, full scalar invariance was also achieved for MSPSS when comparing individuals with a single and multiple health conditions. For BRS, partial scalar invariance was established across health conditions. Conclusions: The original factor structures of the BRS and MSPSS fit with good internal consistency, providing preliminary evidence of construct validity and reliability of scores on both scales in this population.
Abstractor: As Provided
Entry Date: 2025
Accession Number: EJ1457725
Database: ERIC
Full text is not displayed to guests.
FullText Links:
  – Type: pdflink
    Url: https://content.ebscohost.com/cds/retrieve?content=AQICAHj0k_4E0hTGH8RJwT4gCJyBsGNe_WN95AvKlDbXJGqwxwFfDG1oDKGvxeTuBPuf1PTVAAAA4zCB4AYJKoZIhvcNAQcGoIHSMIHPAgEAMIHJBgkqhkiG9w0BBwEwHgYJYIZIAWUDBAEuMBEEDIDLIdiLR7LouJLJZAIBEICBm6zsCiV9PSq5ePsi-hjqzIU6lDBrQBhpIieV3-qcWXL7sC9c4WikmM5TW23pwK8y2xDUdzZ6oX28cc0cI6N4B892BPuUYi0xaOoroHYylGlRvLZRA3lfnLeSTkDYQPgOtTeWtSe6sp4I1jQxERM_KHmJHPQJoKzjYym9KfXVu6z3xgdcsB8y-B6liljZH_Wl2qQn0dNrOsvc2x0d
Text:
  Availability: 1
  Value: <anid>AN0182245067;mev01jan.25;2025Jan17.04:41;v2.2.500</anid> <title id="AN0182245067-1">Psychometric Evaluation of the Brief Resilience Scale and Multidimensional Scale of Perceived Social Support Among Older Sexual Minority Women in the U.S </title> <p>Objective: This study sought to examine the factor structure, internal consistency, and measurement invariance of the Brief Resilience Scale (BRS) and the Multidimensional Scale of Perceived Social Support (MSPSS) among older sexual minority women with disabilities. Method: Participants (n = 208) consisted of sexual minority women aged 55 and older with disabilities recruited online. Results: CFA results showed a one-factor structure fit BRS data, RMSEA =.050 (90% CI [.00,.107]), SRMR =.015, CFI =.990, TLI =.975, with good internal consistency (alpha = 0.863, omega =.864). For the MSPSS, CFA showed a three-factor structure fit the data, RMSEA =.080 (90% CI [.064,.096]), SRMR =.037, CFI =.959, TLI =.947, with excellent internal consistency (alpha = 0.937, omega =.937). Full scalar invariance was achieved for both BRS and MSPSS when comparing single and partnered individuals. Additionally, full scalar invariance was also achieved for MSPSS when comparing individuals with a single and multiple health conditions. For BRS, partial scalar invariance was established across health conditions. Conclusions: The original factor structures of the BRS and MSPSS fit with good internal consistency, providing preliminary evidence of construct validity and reliability of scores on both scales in this population.</p> <p>SIGNIFICANCE STATEMENT: Older sexual minority women with disabilities are often excluded from research, limiting applications of evidence-supported mental health interventions. This study provides evidence that scores on resilience and social support measures are valid and reliable for this population, thereby expanding opportunities for research with this population to improve clinical practice.</p> <p>Keywords: Resilience; social support; confirmatory factor analysis; LGBTQ+; aging</p> <p>Lesbian, gay, bisexual, transgender, and queer/questioning (LGBTQ+) older adults (typically aged 65 or older) are underrepresented in health research (Fredriksen-Goldsen et al., [<reflink idref="bib28" id="ref1">28</reflink>]), despite evidence of disparate physical and mental health outcomes in comparison with cisgender, heterosexual older adults (Fredriksen-Goldsen et al., [<reflink idref="bib27" id="ref2">27</reflink>]). This is especially true in counseling research, where only ten articles in the past twenty-five years across counseling flagship journals have examined LGBTQ+ older adulthood (Fullen et al., [<reflink idref="bib33" id="ref3">33</reflink>]; Westcott, [<reflink idref="bib63" id="ref4">63</reflink>]). Further, subgroups of LGBTQ+ older adults who experience additional points of marginalization (e.g., on the basis of gender and disability status) receive even less attention, which masks health inequities within an already marginalized population.</p> <p>Gender and disability are clear areas of disparity. Disability refers to having health problems that require supports or modifications or that create limitations in major life domains, which may result from physical, mental, or emotional difficulties (Dispenza et al., [<reflink idref="bib20" id="ref5">20</reflink>]; Smart, [<reflink idref="bib56" id="ref6">56</reflink>]). These experiences are inclusive of chronic illnesses (Smart, [<reflink idref="bib56" id="ref7">56</reflink>]). Compared with both heterosexual older adults and sexual minority older men, older sexual minority women (SMW) experience poorer general health and report a greater number of serious health events (i.e. stroke, heart attack; Fredriksen-Goldsen et al., [<reflink idref="bib31" id="ref8">31</reflink>]). Additionally, nearly half of older SMW self-report having disabilities (Fredriksen-Goldsen et al., [<reflink idref="bib27" id="ref9">27</reflink>]), although very little research explores counseling needs of older SMW with disabilities. These omissions inhibit the development of culturally responsive mental health interventions for older SMW with disabilities, who have distinct health and mental health needs (Westcott, [<reflink idref="bib63" id="ref10">63</reflink>]). Therefore, even as the body of literature focused on LGBTQ+ older adults is growing (Fredriksen-Goldsen & Kim, [<reflink idref="bib29" id="ref11">29</reflink>]), subgroups like older SMW with disabilities also need further investigation to ensure that counselors can promote their growth and development.</p> <p>Among other challenges associated with conducting research with LGBTQ+ older adults, an important contributing factor is a lack of validated instruments for researchers to use with these populations (Fredriksen-Goldsen & Kim, [<reflink idref="bib29" id="ref12">29</reflink>]). Therefore, there remains a need to validate measures with this population for counseling researchers to advance research with older SMW with disabilities. Despite limited research focused on this population, there is evidence of health and mental health disparities among older SMW with disabilities. Older SMW are less likely to have a regular health care provider (Fredriksen-Goldsen et al., [<reflink idref="bib32" id="ref13">32</reflink>]) and more likely to have self-reported disabilities (Fredriksen-Goldsen et al., [<reflink idref="bib31" id="ref14">31</reflink>]) than other sexual minority or straight older adults. Older SMW are also more likely to report mental health diagnoses and serious psychological distress than other older people (Masa et al., [<reflink idref="bib49" id="ref15">49</reflink>]), although the intersection with disability is absent here. However, exploring strengths and protective factors, or characteristics that reduce the impact of risk factors on health outcomes (Substance Abuse & Mental Health Services Administration, [<reflink idref="bib58" id="ref16">58</reflink>]), is especially important in counseling, as we operate from holistic wellness frameworks and emphasize multidimensional wellbeing (Healey & Hays, [<reflink idref="bib37" id="ref17">37</reflink>]; Kaplan et al., [<reflink idref="bib40" id="ref18">40</reflink>]).</p> <hd id="AN0182245067-2">Protective Factors: Resilience and Social Support</hd> <p>Two key constructs that emerge as protective factors in both counseling and LGBTQ+ aging research are resilience (Fredriksen-Goldsen, [<reflink idref="bib26" id="ref19">26</reflink>]) and social support (Kim et al., [<reflink idref="bib42" id="ref20">42</reflink>]). Resilience, or the ability to withstand or recover from significant challenges in one's life (Colpitts & Gahagan, [<reflink idref="bib15" id="ref21">15</reflink>]; Smith et al., [<reflink idref="bib57" id="ref22">57</reflink>]), is a key construct in LGBTQ+ health frameworks (Colpitts & Gahagan, [<reflink idref="bib15" id="ref23">15</reflink>]). There is some evidence for resilience protecting against poorer outcomes for LGBTQ+ older adults (Bower et al., [<reflink idref="bib3" id="ref24">3</reflink>]), older SMW (Candrian et al., [<reflink idref="bib6" id="ref25">6</reflink>]), and people with disabilities (Alschuler et al., [<reflink idref="bib2" id="ref26">2</reflink>]), the intersectional identities of older SMW with disabilities remain unexplored. In one study, resilience mediated the negative impact of discrimination and victimization on LGBTQ+ older adults' general health (Fredriksen-Goldsen et al., [<reflink idref="bib31" id="ref27">31</reflink>]). Specific to older SMW, in a qualitative study of older lesbians, participants shared that resilience helped them channel fear and anger from discriminatory experiences into action (Candrian et al., [<reflink idref="bib6" id="ref28">6</reflink>]). Resilience accounted for variance in pain outcomes among people with chronic pain and physical disability above and beyond vulnerability factors (Alschuler et al., [<reflink idref="bib2" id="ref29">2</reflink>]), and greater resilience predicted decreases in depression in a longitudinal study (Silverman et al., [<reflink idref="bib55" id="ref30">55</reflink>]). Most recently, Westcott ([<reflink idref="bib63" id="ref31">63</reflink>]) found that resilience had a positive relationship with both physical and mental health in a sample of older SMW with disabilities.</p> <p>Another construct that appears to protect against poorer health outcomes is social support, or a subjective sense of support received from social relationships (Zimet et al., [<reflink idref="bib66" id="ref32">66</reflink>]). Similar to resilience, the paucity of research on the experiences of older SMW with disabilities requires exploration of related groups to contextualize how social support may manifest as a protective factor for mental health. In previous research with LGBTQ+ older adults, social support has been associated with better mental and physical health (Kim & Fredriksen-Goldsen, [<reflink idref="bib41" id="ref33">41</reflink>]). Social support and network size also served as protective factors in a model predicting health outcomes among sexual minority older adults broadly (Fredriksen-Goldsen et al., [<reflink idref="bib27" id="ref34">27</reflink>]). Perceived social support and social network have also been positively associated with mental health-related quality of life among LGBTQ+ older adults (Fredriksen-Goldsen et al., [<reflink idref="bib30" id="ref35">30</reflink>]). Westcott ([<reflink idref="bib63" id="ref36">63</reflink>]) also found that social support had a positive relationship with mental health in a sample of older SMW with disabilities.</p> <p>Given these findings, both social support and resilience may function as protective factors for older SMW with disabilities. These factors are especially relevant to counselors, as they appear to be mutable (e.g. Colpitts & Gahagan, [<reflink idref="bib15" id="ref37">15</reflink>]) and therefore amenable to counseling interventions. Furthermore, resilience and social support appear to be related. Social support is a distinct protective factor, yet it may also contribute to interpersonal resilience processes (Colpitts & Gahagan, [<reflink idref="bib15" id="ref38">15</reflink>]). This is empirically supported: Silverman et al. ([<reflink idref="bib55" id="ref39">55</reflink>]) found in a longitudinal study of people with disabilities that greater resilience at baseline predicted increased social functioning three years later. Similarly, social support and resilience both contributed to adaptive coping in a sample of older sexual minority people (Dejanipont et al., [<reflink idref="bib18" id="ref40">18</reflink>]). Again, less information is available about how these constructs function among older SMW with disabilities, but evidence from similar populations suggests that these constructs are distinct but related. To support future inquiry into relationships between these constructs and mental health variables among older SMW with disabilities and to measure the impact of counseling interventions on these factors, it is necessary to ensure that researchers can trust the scores produced by measures of resilience and social support in this population.</p> <hd id="AN0182245067-3">Measuring Resilience and Social Support</hd> <p>Although there is some evidence for both resilience and social support as protective factors for older SMW with disabilities, additional attention is needed on the methods researchers use to measure these constructs in these populations. A common tool for measuring resilience in counseling research is the Brief Resilience Scale (BRS; Smith et al., [<reflink idref="bib57" id="ref41">57</reflink>]). This six-item measure assesses respondents' perceived ability to recover from stressful situations. Its brevity and unidimensional structure are strengths for its application in counseling research. Smith et al. ([<reflink idref="bib57" id="ref42">57</reflink>]) originally validated the BRS in four samples: two consisted of undergraduate university students, one consisted of cardiac rehabilitation patients, and the last consisted of women who had fibromyalgia or were healthy controls. Although the fourth sample suggests that scores on the BRS may have adequate psychometric properties among women, it is less clear if scores from older SMW with disabilities would perform similarly. For example, Smith et al. described negative correlations with disability-related symptoms (e.g. pain, physical symptoms) as evidence of discriminant predictive validity (i.e. resilience would predict health outcomes) in the fourth sample, which included women with disabilities. In a sample of people with disabilities (e.g. older SMW with disabilities), it is less clear whether scores on the BRS adequately represent resilience.</p> <p>There is evidence that the BRS has acceptable psychometric properties in a range of diverse populations. For example, in a Spanish translation of the BRS tested with undergraduate students, the internal consistency reliability alpha was.84, although this study did not assess factor structure (Lenz et al., [<reflink idref="bib46" id="ref43">46</reflink>]). Specific to LGBTQ+ populations, Eadeh et al. ([<reflink idref="bib21" id="ref44">21</reflink>]) found acceptable internal consistency (Cronbach's alpha =.865) of BRS scores in a sample of 155 LGBTQ+ adults in the United States (U.S.). Similarly, scores on the BRS had acceptable internal consistency (Cronbach's alpha =.89) in an Australian sample of 895 LGBTQ+ adults aged 60 and older (Lyons et al., [<reflink idref="bib47" id="ref45">47</reflink>]). However, none of these studies explored the factor structure or other indicators of construct validity. Thus, there remains a need to explore the construct validity and factor structure of the BRS among older SMW with disabilities.</p> <p>Similarly, a popular measure of social support is the Multidimensional Scale of Perceived Social Support (MSPSS; Zimet et al., [<reflink idref="bib66" id="ref46">66</reflink>]), a 12-item instrument assessing perceived social support across three dimensions (friends, family, and significant other). Zimet et al. ([<reflink idref="bib66" id="ref47">66</reflink>]) originally validated the MSPSS with undergraduate university students enrolled in an introductory psychology course. There is evidence that scores on the MSPSS are valid and reliable for LGBTQ+ populations broadly. In a sample of 239 sexual minority Filipino people, the MSPSS had excellent internal consistency on each of its subscales (Cronbach's alpha of 0.91 for family, 0.92 for friends, and 0.96 for significant other), as well as acceptable overall internal consistency (Cronbach's alpha = 0.85; Reyes et al., [<reflink idref="bib52" id="ref48">52</reflink>]). Kler et al. ([<reflink idref="bib43" id="ref49">43</reflink>]) found that the original three-factor solution fit for a sample of 267 LGBTQ+ people of color, as well as evidence for convergent and divergent validity. Specific to LGBTQ+ older adults, a study of 113 LGBT adults aged 65 and older found that scores on the MSPSS had excellent internal consistency (Cronbach's alpha = 0.93), although they did not assess factor structure (Marmo et al., [<reflink idref="bib48" id="ref50">48</reflink>]). Notably, no research to date explores how this instrument performs with older SMW with disabilities, highlighting another limitation to conducting research with this population.</p> <p>There is further evidence that scores on the BRS and MSPSS may differ based on different group characteristics. For the BRS, because disability was used as evidence of less resilience in the initial validation (Smith et al., [<reflink idref="bib57" id="ref51">57</reflink>]), it is unclear whether more expansive experiences of disability (i.e. multiple disability types) might impact performance on the BRS compared with a single disability experience. Furthermore, there is evidence that women with two or more health conditions scored lower on the BRS than women with one health condition in a recent study (Kumar et al., [<reflink idref="bib44" id="ref52">44</reflink>]). Similarly, Kumar et al. ([<reflink idref="bib44" id="ref53">44</reflink>]) found that more single women report low resilience on the BRS than do partnered women. Although these statistics are unavailable for the study population given the dearth of research, it stands to reason that these effects could impact performance of the BRS among older SMW with disabilities.</p> <p>Regarding the MSPSS, one of the subscales focuses on support from a significant other. Therefore, we theorized that its measurement may vary between partnered and unpartnered individuals, as definitions of social support from a significant other may vary with the presence of a romantic or sexual partner. For example, in a study with adolescents in Hong Kong, the significant other support subscale appeared to measure support from friends and family simultaneously (Cheng & Chan, [<reflink idref="bib10" id="ref54">10</reflink>]), illustrating how an individual's perceptions of the term significant other may influence scores on the significant other subscale and therefore MSPSS broadly. Further, we were unable to locate any research that explicitly examined how disability might impact the performance of the MSPSS, a significant oversight in the current body of literature. Given that there are associations between social support and disability (e.g. Emerson et al., [<reflink idref="bib23" id="ref55">23</reflink>]), we hypothesized that respondents with a single disability and those with multiple disabilities may be differentially impacted by disability and therefore perform differently on the MSPSS. Notably, most measurement invariance conducted with the MSPSS has been in validation studies across different languages and cultures (e.g. among Chilean older adults; Pérez-Villalobos et al., [<reflink idref="bib51" id="ref56">51</reflink>]) rather than looking at intersections of U.S. populations. Therefore, measurement invariance for the MSPSS is needed to continue to validate the instrument with intersectional populations, such as older SMW with disabilities, in the U.S.</p> <hd id="AN0182245067-4">Purpose of the Study</hd> <p>Because no studies have examined how these commonly used instruments perform with older SMW with disabilities, counseling researchers may not be able to measure protective factors for this population, limiting our ability to contribute to an evidence base for counseling interventions. Therefore, additional inquiry into the validity and reliability of scores on measures such as the BRS (Smith et al., [<reflink idref="bib57" id="ref57">57</reflink>]) and the MSPSS (Zimet et al., [<reflink idref="bib66" id="ref58">66</reflink>]) may help counseling researchers measure resilience and social support among older SMW with disabilities. Furthermore, conducting psychometric evaluation of both instruments together allows us to consider evidence of convergent validity, given that these constructs are related but distinct. This study was consequently guided by the following research questions:</p> <p></p> <ulist> <item> Will Confirmatory Factor Analysis (CFA) support the unidimensional factor structure of the BRS (Smith et al., [<reflink idref="bib57" id="ref59">57</reflink>]) in a sample of older SMW with disabilities? If not, what modifications may be necessary?</item> <p></p> <item> What is the internal consistency of the BRS (Smith et al., [<reflink idref="bib57" id="ref60">57</reflink>]) in a sample of older SMW with disabilities?</item> <p></p> <item> Will CFA support the three-factor structure (i.e. friends, family, significant other) of the MSPSS (Zimet et al., [<reflink idref="bib66" id="ref61">66</reflink>]) in a sample of older SMW with disabilities? If not, what modifications may be necessary?</item> <p></p> <item> What is the internal consistency of the MSPSS (Zimet et al., [<reflink idref="bib66" id="ref62">66</reflink>]) in a sample of older SMW with disabilities?</item> <p></p> <item> Are the measurement properties of the BRS and MSPSS invariant across single and partnered individuals as well as individuals with single and multiple health conditions?</item> </ulist> <hd id="AN0182245067-5">Method</hd> <p></p> <hd id="AN0182245067-6">Inclusion and Exclusion</hd> <p>Potential respondents were eligible to participate in the study if they met the following inclusion criteria: a) residing in the U.S.; b) identifying as a lesbian, bisexual woman, or otherwise non-heterosexual woman; c) aged 55 or older; and d) identifying as having a disability or health concern. This age cutoff was selected given the small proportion of adults over the age of 65 identifying as LGBTQ+ and is in line with previous studies of LGBTQ+ older adults (e.g. Fredriksen-Goldsen et al., [<reflink idref="bib28" id="ref63">28</reflink>]). Eligibility was assessed using the following dichotomous questions: (<reflink idref="bib1" id="ref64">1</reflink>) <emph>Do you live in the United States?</emph>; (<reflink idref="bib2" id="ref65">2</reflink>) <emph>Do you identify as a lesbian (woman or non-binary) or a woman who pursues romantic and sexual relationships with women?</emph>; and (<reflink idref="bib3" id="ref66">3</reflink>) Are you 55 years of age or older? If respondents answered "No" to any of these questions, they were directed to an ineligibility page and were unable to participate. We assessed eligibility for disability by asking potential respondents indicate whether they had any of the following disabilities or health problems: (<reflink idref="bib1" id="ref67">1</reflink>) Blind; (<reflink idref="bib2" id="ref68">2</reflink>) Hard of hearing; (<reflink idref="bib3" id="ref69">3</reflink>) deaf/Deaf; (<reflink idref="bib4" id="ref70">4</reflink>) A physical disability or condition that impacts mobility of your fingers, hands, or arms; (<reflink idref="bib5" id="ref71">5</reflink>) A physical disability or condition that impacts mobility of your feet or legs; (<reflink idref="bib6" id="ref72">6</reflink>) A speech, language, or communication disorder; (<reflink idref="bib7" id="ref73">7</reflink>) A cognitive, intellectual, or learning disability; (<reflink idref="bib8" id="ref74">8</reflink>) A psychological illness or disability; (<reflink idref="bib9" id="ref75">9</reflink>) Another disability (with an opportunity for a write-in response); or (<reflink idref="bib10" id="ref76">10</reflink>) None of the above. If respondents selected none of the above, they were directed to an ineligibility page and were unable to participate.</p> <p>We utilized these specific screening procedures to ensure that respondents met the study criteria while allowing for self-identification that honors their salient identities (e.g. Sharma et al., [<reflink idref="bib54" id="ref77">54</reflink>]). For example, sexual minority women, defined as lesbians or non-heterosexual women, includes non-male genders who identify as lesbians and are comfortable being identified as a sexual minority woman (Thorne et al., [<reflink idref="bib60" id="ref78">60</reflink>]). The second screening question ensured that respondents who identified as non-binary still considered themselves to be lesbians or sexual minority women. Similarly, this screening question allowed respondents to elect not to disclose information (e.g. sexual orientation) that might expose them to additional risk, which is aligned with best practices when collecting sensitive information related to sexual orientation or gender identity (Welfare et al., [<reflink idref="bib62" id="ref79">62</reflink>]; Westcott et al., [<reflink idref="bib64" id="ref80">64</reflink>]). This approach also allowed us to use a more restrictive disability identifier question for our screening item, whereas we used an inclusive self-report item for how respondents might define their experience with disability (see Measures).</p> <hd id="AN0182245067-7">Participant Characteristics</hd> <p>Respondents (<emph>n</emph><bold></bold>=<bold></bold>208) ranged from 55 to 83 years of age (<emph>M</emph><bold></bold><emph>=</emph><bold></bold>62.80, <emph>SD</emph><bold></bold>=<bold></bold>6.53). All respondents indicated that they experienced at least one disability or health concern (i.e. chronic illness, hearing difficulty, ambulatory difficulty, etc.), with almost half (<emph>n</emph><bold></bold>=<bold></bold>103, 49.5%) reporting one health condition. Other respondents indicated that they had two (<emph>n</emph><bold></bold>=<bold></bold>50, 24.0%), three (<emph>n</emph><bold></bold>=<bold></bold>26, 12.5%), four (<emph>n</emph><bold></bold>=<bold></bold>19, 9.1%), five (<emph>n</emph><bold></bold>=<bold></bold>5, 2.4%), six (<emph>n</emph><bold></bold>=<bold></bold>4, 1.9%), or seven (<emph>n</emph><bold></bold>=<bold></bold>1, 0.5%) types of health concerns. Although nearly all participants identified as women (<emph>n</emph><bold></bold>=<bold></bold>203, 97.6%), four (1.9%) respondents indicated that they identified as non-binary and one respondent selected a different gender and self-identified as a "transgender woman." Additional relevant sociodemographic information is summarized in Table 1.</p> <p>Table 1. Sociodemographic Characteristics of Participants.</p> <p> <ephtml> <table><thead><tr><td>Characteristic</td><td><italic>n</italic></td><td>%</td></tr></thead><tbody valign="top"><tr><td>Gender</td><td /><td /></tr><tr><td> Woman</td><td char=".">203</td><td char=".">97.6%</td></tr><tr><td> Non-binary</td><td char=".">4</td><td char=".">1.9%</td></tr><tr><td> A different gender</td><td char=".">1</td><td char=".">0.5%</td></tr><tr><td>Racial identity</td><td /><td /></tr><tr><td> American Indian or Alaska Native</td><td char=".">9</td><td char=".">4.3%</td></tr><tr><td> Asian or Asian American</td><td char=".">0</td><td char=".">0.0%</td></tr><tr><td> Black or African American</td><td char=".">22</td><td char=".">10.6%</td></tr><tr><td> Native Hawaiian or Other Pacific Islander</td><td char=".">1</td><td char=".">0.5%</td></tr><tr><td> Middle Eastern or North African</td><td char=".">1</td><td char=".">0.5%</td></tr><tr><td> White</td><td char=".">177</td><td char=".">85.1%</td></tr><tr><td> Biracial, multiracial, or mixed race</td><td char=".">3</td><td char=".">1.4%</td></tr><tr><td> A different race</td><td char=".">3</td><td char=".">1.4%</td></tr><tr><td> Prefer not to say</td><td char=".">3</td><td char=".">1.0%</td></tr><tr><td>Ethnic identity</td><td /><td /></tr><tr><td> Of Hispanic, Latino, or Spanish origin</td><td char=".">16</td><td char=".">7.7%</td></tr><tr><td> Not of Hispanic, Latino, or Spanish origin</td><td char=".">192</td><td char=".">92.3%</td></tr><tr><td>Sexual orientation</td><td /><td /></tr><tr><td> Lesbian</td><td char=".">139</td><td char=".">66.8%</td></tr><tr><td> Bisexual</td><td char=".">65</td><td char=".">31.3%</td></tr><tr><td> Pansexual</td><td char=".">7</td><td char=".">3.4%</td></tr><tr><td> Queer</td><td char=".">5</td><td char=".">2.4%</td></tr><tr><td> Asexual</td><td char=".">3</td><td char=".">1.4%</td></tr><tr><td> Prefer not to say</td><td char=".">1</td><td char=".">0.5%</td></tr><tr><td>Disability type</td><td /><td /></tr><tr><td> Hearing difficulty</td><td char=".">50</td><td char=".">24.0%</td></tr><tr><td> Vision difficulty</td><td char=".">25</td><td char=".">12.0%</td></tr><tr><td> Cognitive difficulty</td><td char=".">39</td><td char=".">18.8%</td></tr><tr><td> Ambulatory difficulty</td><td char=".">92</td><td char=".">44.2%</td></tr><tr><td> Self-care difficulty</td><td char=".">22</td><td char=".">10.6%</td></tr><tr><td> Independent living difficulty</td><td char=".">16</td><td char=".">7.7%</td></tr><tr><td> Chronic illness</td><td char=".">131</td><td char=".">63.0%</td></tr><tr><td> Different type of health problem</td><td char=".">33</td><td char=".">15.9%</td></tr><tr><td> Prefer not to say</td><td char=".">5</td><td char=".">2.4%</td></tr><tr><td>Current marital status</td><td /><td /></tr><tr><td> Single</td><td char=".">68</td><td char=".">32.7%</td></tr><tr><td> Partnered</td><td char=".">26</td><td char=".">12.5%</td></tr><tr><td> Married</td><td char=".">48</td><td char=".">23.1%</td></tr><tr><td> Divorced</td><td char=".">39</td><td char=".">18.8%</td></tr><tr><td> Widowed</td><td char=".">27</td><td char=".">13.0%</td></tr><tr><td>Highest level of education attained</td><td /><td /></tr><tr><td> Grade school/some high school</td><td char=".">3</td><td char=".">1.4%</td></tr><tr><td> High school/GED</td><td char=".">31</td><td char=".">14.9%</td></tr><tr><td> Some college/Associate's degree</td><td char=".">87</td><td char=".">41.8%</td></tr><tr><td> Trade school</td><td char=".">17</td><td char=".">8.2%</td></tr><tr><td> Bachelor's degree</td><td char=".">36</td><td char=".">17.3%</td></tr><tr><td> Master's degree</td><td char=".">29</td><td char=".">13.9%</td></tr><tr><td> Doctoral degree</td><td char=".">3</td><td char=".">1.4%</td></tr><tr><td> Another degree</td><td char=".">1</td><td char=".">0.5%</td></tr><tr><td> Prefer not to say</td><td char=".">1</td><td char=".">0.5%</td></tr><tr><td>Income</td><td /><td /></tr><tr><td> Less than $25,000</td><td char=".">83</td><td char=".">39.9%</td></tr><tr><td> $25,000 to $49,999</td><td char=".">68</td><td char=".">32.7%</td></tr><tr><td> $50,000 to $74,999</td><td char=".">26</td><td char=".">12.5%</td></tr><tr><td> $75,00 to $99,999</td><td char=".">12</td><td char=".">5.8%</td></tr><tr><td> $100,000 to $124,999</td><td char=".">6</td><td char=".">2.9%</td></tr><tr><td> $125,000 to $149,999</td><td char=".">6</td><td char=".">2.9%</td></tr><tr><td> $150,000 to $174,999</td><td char=".">3</td><td char=".">1.4%</td></tr><tr><td> $175,000 to $199,999</td><td char=".">1</td><td char=".">0.5%</td></tr><tr><td> More than $200,000</td><td char=".">2</td><td char=".">1.0%</td></tr><tr><td> Prefer not to say</td><td char=".">1</td><td char=".">0.5%</td></tr><tr><td>Health insurance source</td><td /><td /></tr><tr><td> None</td><td char=".">5</td><td char=".">2.4%</td></tr><tr><td> Insured through current/former employer</td><td char=".">20</td><td char=".">9.6%</td></tr><tr><td> Insured through spouse/partner</td><td char=".">13</td><td char=".">6.3%</td></tr><tr><td> Healthcare.gov insurance</td><td char=".">15</td><td char=".">7.2%</td></tr><tr><td> Insurance purchased through insurance company</td><td char=".">14</td><td char=".">6.7%</td></tr><tr><td> Medicare</td><td char=".">114</td><td char=".">54.8%</td></tr><tr><td> Medicaid</td><td char=".">72</td><td char=".">34.6%</td></tr><tr><td> TRICARE/other military health care</td><td char=".">6</td><td char=".">2.9%</td></tr><tr><td> VA</td><td char=".">7</td><td char=".">3.4%</td></tr><tr><td> Another type of insurance</td><td char=".">4</td><td char=".">1.9%</td></tr><tr><td> Prefer not to say</td><td char=".">2</td><td char=".">1.0%</td></tr></tbody></table> </ephtml> </p> <p>1 <emph>Note.</emph> Percentages may add up to more than 100% for some sociodemographic characteristics, as participants had the opportunity to select more than one response (e.g. racial identity, sexual orientation, insurance source). Percentages may not add up to 100% for other sociodemographic characteristics due to rounding errors.</p> <hd id="AN0182245067-8">Sampling Procedures</hd> <p>Following Institutional Review Board approval, we collected data using CloudResearch's Prime Panels. This managed research service has a panel of more than 50 million potential participants, 26% of whom are over the age of 65 (Chandler et al., [<reflink idref="bib8" id="ref81">8</reflink>]). The service uses Sentry, a pre-survey instrument employing behavioral and technological assessment to ensure high quality data (CloudResearch, [<reflink idref="bib14" id="ref82">14</reflink>].). The survey was distributed to CloudResearch members whose demographic data indicated they might meet the inclusion criteria indicated above. Potential respondents provided informed consent by reviewing a document and confirming consent prior to accessing the survey.</p> <hd id="AN0182245067-9">Measures</hd> <p>Participants responded to a battery of validated instruments and sociodemographic information. For a full list of administered measures, see Author, 2023c. Below, we detail the relevant measures and instrumentation for the present study.</p> <hd id="AN0182245067-10">Sociodemographic Information</hd> <p>Respondents provided information related to demographics, including gender, age, race/ethnicity, sexual orientation, disability or chronic illness type, marital status, educational attainment, annual household income, and health insurance coverage source. We adapted the demographic question related to disability type (i.e. "What type of disability or chronic illness do you have? Please select all that apply.") from the American Community Survey (ACS), which can estimate disability accurately for small groups of a given population (U.S. Census Bureau, [<reflink idref="bib61" id="ref83">61</reflink>]). This was distinct from the screening eligibility questions to allow more accurate self-identification of disability type. Because eligibility was confirmed in a prior screening, respondents were able to elect not to respond to any demographic questions.</p> <hd id="AN0182245067-11">Attention Check Items</hd> <p>We distributed three items throughout the survey as attention checks. These items included: "<emph>Please select option 2, or 'Disagree', for this item."; "Please select 'Strongly Agree' for this item."</emph>; and <emph>"Please select 'Strongly Disagree' for this item."</emph></p> <hd id="AN0182245067-12">Brief Resilience Scale</hd> <p>The BRS (Smith et al., [<reflink idref="bib57" id="ref84">57</reflink>]) is a 6-item scale measuring a person's ability to recover from adverse experiences. Initial validation provided evidence for internal consistency of scores on the BRS, with Cronbach's alpha ranging from.80 to.91 across four samples and test-retest reliability ranging from.62 to.69 across two samples. Smith et al. ([<reflink idref="bib57" id="ref85">57</reflink>]) used principal component analysis to show that a one-factor solution accounted for 55-67% of the variance in scores across four samples in their initial validation study, with loadings ranging for 0.68 to 0.91 for each item. They provided evidence of convergent validity by demonstrating positive correlations with other validated resilience measures and life purpose, as well as negative correlations with pessimism and alexithymia measures. They also demonstrated discriminant validity by demonstrating negative correlations between the BRS and anxiety, depression, negative affect, pain, and physical symptoms. Example items include: <emph>"I tend to bounce back quickly after hard times"</emph> and <emph>"I tend to take a long time to get over set-backs in my life".</emph> Three items are reverse scored, with larger scores indicating greater resilience.</p> <hd id="AN0182245067-13">Multidimensional Scale of Perceived Social Support</hd> <p>The MSPSS (Zimet et al., [<reflink idref="bib66" id="ref86">66</reflink>]) is a 12-item measure of social support. Scores on the MSPSS have demonstrated good internal consistency (Cronbach's alpha ranging from.84 to.92 across multiple samples) and test-retest reliability (<emph>r</emph> =.85) in previous samples. The MSPSS has three subscales, measuring social support received from different sources, including significant others (Cronbach's alpha =.91, test-retest <emph>r</emph> =.72), family (Cronbach's alpha =.87, test-retest <emph>r</emph> =.85), and friends (Cronbach's alpha =.85, test-retest <emph>r</emph> =.75). Zimet et al. ([<reflink idref="bib66" id="ref87">66</reflink>]) used exploratory factor analysis to extract three factors, using an oblique rotation, with high factor loadings: 0.74 to 0.92 for significant other, 0.81 to 0.84 for family, and 0.79 to 0.86 for friends. They additionally provided evidence of construct validity by demonstrating that the MSPSS was negatively correlated with depression symptoms (<emph>r</emph><bold></bold>=<bold></bold>−0.25, <emph>p</emph> <.01), as were associated subscales. Example items include: "<emph>There is a special person who is around when I am in need</emph>" (Significant Other subscale), "<emph>My friends really help me</emph>" (Friend subscale), and "<emph>I can talk about my problems with my family</emph>" (Family subscale). Higher scores indicate greater perceived social support.</p> <hd id="AN0182245067-14">Data Collection</hd> <p>Data were collected over a four-week period between September and October 2022. A total of 404 individuals met eligibility criteria based on their responses, and 218 total respondents completed the survey (53.96% response rate for eligible participants).</p> <hd id="AN0182245067-15">Data Diagnostics</hd> <p>Following confirmation that all recorded responses were eligible based on their survey responses, we examined attention check items and response patterns to identify cases to review for validity. In total, 10 responses indicated potential problems with data validity (e.g. missed two or more attention check items, large amounts of missing data, etc.), leading us to remove these cases and retain the remaining 208 responses for analysis. Because of CloudResearch's quality control procedures (i.e. Sentry screening), after removing cases with validity issues, very little missing data remained. Of the measures utilized in the present study, the sixth item on the BRS was missing one case response, with no other data missing. Therefore, the missing data was negligible (Enders, [<reflink idref="bib24" id="ref88">24</reflink>]), and we did not impute data. Models were estimated using a full information maximum likelihood approach for missing data and results did not change; therefore, we retained the original analysis.</p> <hd id="AN0182245067-16">Analytic Strategy</hd> <p>We utilized Confirmatory Factor Analysis (CFA) to assess the factor structure of the BRS and MSPSS. CFA is a psychometric technique used to examine the latent structure of a set of observed variables and provides construct validity evidence for an instrument (Brown, [<reflink idref="bib4" id="ref89">4</reflink>]). CFA models were estimated in R using the lavaan (Rosseel, [<reflink idref="bib53" id="ref90">53</reflink>]) and semTools (Jorgensen et al., [<reflink idref="bib38" id="ref91">38</reflink>]) packages. Response options for the BRS scale consisted of five options that ranged from strongly disagree to strongly agree and the response options for MSPSS consisted of seven options that ranged from very strongly disagree to very strongly agree. Given the ordinal nature of the data (i.e. Likert scale), we utilized a robust maximum likelihood (MLR) estimator. Lei and Shiverdecker ([<reflink idref="bib45" id="ref92">45</reflink>]) demonstrate the robustness of the MLR estimator for ordinal data with at least five categories and a sample size greater than 200. As recommended by Brown ([<reflink idref="bib4" id="ref93">4</reflink>]), and Gana and Broc ([<reflink idref="bib35" id="ref94">35</reflink>]), the following fit indices were employed to assess the overall fit of the model to the data including the comparative fit index (CFI), Tucker-Lewis Index (TLI), root mean square error of approximation (RMSEA), and standardized root mean square residual (SRMR). Guidelines proposed by Brown ([<reflink idref="bib4" id="ref95">4</reflink>]) and Gana and Broc ([<reflink idref="bib35" id="ref96">35</reflink>]) were used for interpreting these indices: CFI and TLI values ≥.90 indicate adequate fit, while values ≥.95 indicate good fit; RMSEA values ≤.08 suggest adequate fit, while values ≤.05 indicate good fit; and SRMR values ≤.08 imply adequate fit, while values of ≤.06 indicate good fit. Additionally, local fit was assessed by examining residual correlations and modification indices. Reliability was assessed using Cronbach's alpha and McDonald's omega. Although Cronbach's alpha is a traditional measure of internal consistency in counseling and psychology research (McNeish, [<reflink idref="bib50" id="ref97">50</reflink>]), McDonald's omega is a more robust measure of internal consistency because it does not assume equal factor loadings among items and therefore is a better general estimator of reliability (Hayes & Coutts, [<reflink idref="bib36" id="ref98">36</reflink>]). Therefore, we reported both values to provide better evidence of internal consistency. After examining the factor structure of the BRS and MSPSS, we examined the measurement invariance for the models across single and partnered individuals in the sample as well as individuals with a single health condition and individuals with multiple health conditions. Three measurement invariance models were tested: configural, metric, and scalar invariance. To establish configural invariance, an equivalent factor structure was specified for each group. To test for metric invariance, factor loadings were constrained to be equal across groups. Finally, to establish scalar invariance, both factor loadings and intercepts were constrained to be equal across groups. If full measurement invariance did not hold, partial invariance was tested. While nested models are typically assessed using a chi-squared difference test, research (Chen, [<reflink idref="bib9" id="ref99">9</reflink>]; Cheung & Rensvold, [<reflink idref="bib11" id="ref100">11</reflink>]) has demonstrated that the chi-squared difference test is sensitive to sample size when evaluating measurement invariance. Therefore, Chen ([<reflink idref="bib9" id="ref101">9</reflink>]) recommends utilizing changes in CFI and RMSEA to assess measurement invariance in addition to chi-squared difference test. According to Chen's guidelines, non-invariance is indicated by a CFI decrease ≥.010 and RMSEA increase ≥.015; Chen also recommends prioritizing ΔCFI as the complexity of a model can impact RMSEA.</p> <p>Prior to estimating our CFA models, we assessed the relationships between pairs of observed variables using scatterplots, which indicated linear relationships. Normality was evaluated using QQ-plots and Mardia's test of multivariate normality. Results indicated deviations from multivariate normality, likely due to the ordinal nature of our response options; therefore, we utilized a robust maximum likelihood estimator to correct for these violations. Outliers were examined <emph>via</emph> scatterplots and by calculating standardized z-scores for each item, with no extreme outliers (<emph>z</emph> > 3.29) detected.</p> <hd id="AN0182245067-17">Results</hd> <p>Descriptive statistics for the BRS and MSPSS can be found in Table 2. The BRS had a moderate, positive correlation with the MSPSS, <emph>r</emph><bold></bold>=<bold></bold>0.239, <emph>p</emph> <.001, CI<subs>95</subs> = (.106,.363). This correlation suggests the scores produced by both scales have convergent validity (Swank & Mullen, [<reflink idref="bib59" id="ref102">59</reflink>]) as related but distinct constructs, as indicated theoretically (Colpitts & Gahagan, [<reflink idref="bib15" id="ref103">15</reflink>]) and in past empirical research (e.g. Silverman et al., [<reflink idref="bib55" id="ref104">55</reflink>]). Below, we detail the findings from our primary analyses.</p> <p>Table 2. Descriptive Statistics for Study Variables.</p> <p> <ephtml> <table><thead><tr><td>Variable</td><td>Mean</td><td>SD</td><td>Variance</td><td>Min</td><td>Max</td><td>Skewness</td><td>Kurtosis</td><td>α (95% CI)</td><td>ω (95% CI)</td></tr></thead><tbody valign="top"><tr><td>BRS</td><td char=".">2.960</td><td char=".">.819</td><td char=".">.670</td><td char=".">1.000</td><td>5.000</td><td char=".">−.263</td><td char=".">−.332</td><td char=".">.863 (.818,.898)</td><td char=".">.864 (.812,.899)</td></tr><tr><td>MSPSS</td><td char=".">4.676</td><td char=".">1.403</td><td char=".">1.967</td><td char=".">1.000</td><td>7.000</td><td char=".">−.601</td><td char=".">.204</td><td char=".">.940 (.921,.954)</td><td char=".">.929 (.903,.952)</td></tr><tr><td>Subscales</td><td>Mean</td><td>SD</td><td>Variance</td><td>Min</td><td>Max</td><td>Skewness</td><td>Kurtosis</td><td /><td /></tr><tr><td>MSPSS Significant Other</td><td char=".">4.908</td><td>1.766</td><td char=".">3.118</td><td char=".">1.000</td><td>7.000</td><td char=".">−.561</td><td char=".">−.563</td><td char=".">.937 (.912,.956)</td><td char=".">.937 (.908,.956)</td></tr><tr><td>MSPSS Family</td><td char=".">4.387</td><td>1.786</td><td char=".">3.188</td><td char=".">1.000</td><td>7.000</td><td char=".">−.383</td><td char=".">−.747</td><td char=".">.962 (.948,.972)</td><td char=".">.962 (.949,.972)</td></tr><tr><td>MSPSS Friend</td><td char=".">4.734</td><td>1.465</td><td char=".">2.146</td><td char=".">1.000</td><td>7.000</td><td char=".">−.607</td><td char=".">.280</td><td char=".">.940 (.917,.957)</td><td char=".">.940 (.916,.957)</td></tr></tbody></table> </ephtml> </p> <hd id="AN0182245067-18">Brief Resilience Scale</hd> <p>A one-factor CFA model was fit to the BRS data in which all six items loaded onto one latent factor. Items 2, 4, and 6 were reverse coded. Using guidelines from Brown ([<reflink idref="bib4" id="ref105">4</reflink>]) and Gana and Broc ([<reflink idref="bib35" id="ref106">35</reflink>]), the results indicated less than ideal data-model fit: <sups>2</sups>(<reflink idref="bib9" id="ref107">9</reflink>) = 55.4, <emph>p</emph> <.001; CFI = 0.88, TLI = 0.80, RMSEA = 0.191, 90% CI<subs>RMSEA</subs> = (.145,.241), SRMR = 0.075. An examination of residuals and modification indices indicated the presence of method effects in which items that were positively and negatively worded had residual covariance not accounted for by the BRS factor (Brown, [<reflink idref="bib4" id="ref108">4</reflink>]). Similar method effects for the BRS were found by Fung ([<reflink idref="bib34" id="ref109">34</reflink>]) and Ye et al. ([<reflink idref="bib65" id="ref110">65</reflink>]) in samples of Chinese and Taiwan undergraduate students, respectively, as well as Chmitorz et al. ([<reflink idref="bib13" id="ref111">13</reflink>]) using the German version of the BRS. The CFA model was re-estimated with a method factor to account for the residual association between the positively worded items. A method factor is an additional latent variable included in the CFA model to account for residual associations between items. The method factor captures the shared variance among positively worded items due to their wording rather than the underlying construct. In this model, the method factor and BRS factor were not allowed to correlate, as per Eid ([<reflink idref="bib22" id="ref112">22</reflink>]). This ensures that the method factor only accounts for the method-related variance and does not interfere with the construct-related variance. Given the additional method factor, we took additional steps to ensure our model was empirically identified. Empirical identification was confirmed as the model converged with no errors, and the model chi-square remained consistent when using different items as unit loading indicators. These checks provide evidence that our model has a unique solution and is empirically identified. Results demonstrated good fit: <sups>2</sups>(<reflink idref="bib6" id="ref113">6</reflink>) = 9.09, <emph>p</emph><bold></bold>=<bold></bold>.169; CFI = 0.990, TLI = 0.975, RMSEA = 0.050, 90% CI<subs>RMSEA</subs> = (.000,.107), SRMR = 0.015. A likelihood ratio test also indicated a preference for the revised model: Δ<sups>2</sups>(<reflink idref="bib3" id="ref114">3</reflink>) = 33.77, <emph>p</emph> <.001. This model has identical fit to a model allowing error covariances between positively worded items. Local fit was examined using residual correlations and found no residual correlations greater than |0.10|. Standardized factor loadings and R-squared values are presented in Table 3. Factor loadings ranged from 0.481 for Item 1 ("I tend to bounce back quickly after hard times") to 0.863 for Item 6 ("I tend to take a long time to get over set-backs in my life"), with the negatively worded items loading higher on the BRS factor (i.e. loadings >.8). R-squared values (i.e. the proportion of variance in the item explained by the factor) for the items ranged from.41 to.75. The average variance extracted for the BRS factor was.51. Cronbach's alpha and McDonald's hierarchical omega internal consistency reliability measures were 0.863 and 0.846, respectively.</p> <p>Table 3. Standardized Factor Loadings Standard Errors, <emph>p-</emph>Value, R-Squared Values, Error Variances, and Error Covariances for Indicator Items and Factor Correlations.</p> <p> <ephtml> <table><thead><tr><td>Item</td><td>Factor Loading</td><td>Standard Error</td><td><italic>p</italic>-Value</td><td>R<sup>2</sup></td><td>Error Variance</td><td>Standard Error</td><td><italic>p</italic>-Value</td></tr></thead><tbody valign="top"><tr><td><italic>Brief Resilience Scale</italic></td><td /><td /><td /><td /><td /><td /><td /></tr><tr><td>I tend to bounce back quickly after hard times (item 1).</td><td char=".">0.481</td><td char=".">.087</td><td char="."><.001</td><td char=".">0.413</td><td char=".">.765</td><td char=".">.110</td><td char="."><.001</td></tr><tr><td>I have a hard time making it through stressful events (item 2).</td><td char=".">0.803</td><td char=".">.051</td><td char="."><.001</td><td char=".">0.644</td><td char=".">.354</td><td char=".">.057</td><td char="."><.001</td></tr><tr><td>It does not take me long to recover from a stressful event (item 3)</td><td char=".">0.549</td><td char=".">.084</td><td char="."><.001</td><td char=".">0.655</td><td char=".">.395</td><td char=".">.112</td><td char="."><.001</td></tr><tr><td>It is hard for me to snap back when something bad happens (item 4).</td><td char=".">0.858</td><td char=".">.051</td><td char="."><.001</td><td char=".">0.737</td><td char=".">.262</td><td char=".">.055</td><td char="."><.001</td></tr><tr><td>I usually come through difficult times with little trouble (item 5).</td><td char=".">0.534</td><td char=".">.086</td><td char="."><.001</td><td char=".">0.583</td><td char=".">.711</td><td char=".">.114</td><td char="."><.001</td></tr><tr><td>I tend to take a long time to get over set-backs in my life (item 6)</td><td char=".">0.863</td><td char=".">.056</td><td char="."><.001</td><td char=".">0.745</td><td char=".">.255</td><td char=".">.077</td><td char=".">.001</td></tr><tr><td><italic>Method Factor</italic></td><td /><td /><td /><td /><td /><td /><td /></tr><tr><td>item 1</td><td char=".">.426</td><td char=".">.116</td><td char="."><.001</td><td /><td /><td /><td /></tr><tr><td>item 3</td><td char=".">.594</td><td char=".">.107</td><td char="."><.001</td><td /><td /><td /><td /></tr><tr><td>item 5</td><td char=".">.546</td><td char=".">.119</td><td char="."><.001</td><td /><td /><td /><td /></tr><tr><td><italic>Significant Other</italic></td><td /><td /><td /><td /><td /><td /><td /></tr><tr><td>There is a special person who is around when I am in need.</td><td char=".">.837</td><td char=".">.050</td><td char="."><.001</td><td char=".">.704</td><td char=".">.294</td><td char=".">.067</td><td char="."><.001</td></tr><tr><td>There is a special person with whom I can share my joys and sorrows.</td><td char=".">.851</td><td char=".">.059</td><td char="."><.001</td><td char=".">.729</td><td char=".">.270</td><td char=".">.078</td><td char="."><.001</td></tr><tr><td>I have a special person who is a real source of comfort to me.</td><td char=".">.940</td><td char=".">.048</td><td char="."><.001</td><td char=".">.888</td><td char=".">.111</td><td char=".">.032</td><td char="."><.001</td></tr><tr><td>There is a special person in my life who cares about my feelings.</td><td char=".">.904</td><td char=".">.045</td><td char="."><.001</td><td char=".">.820</td><td char=".">.179</td><td char=".">.039</td><td char="."><.001</td></tr><tr><td><italic>Family</italic></td><td /><td /><td /><td /><td /><td /><td /></tr><tr><td>My family really tries to help me.</td><td char=".">.940</td><td char=".">.044</td><td char="."><.001</td><td char=".">.888</td><td char=".">.111</td><td char=".">.021</td><td char="."><.001</td></tr><tr><td>I get the emotional help and support I need from my family.</td><td char=".">.964</td><td char=".">.037</td><td char="."><.001</td><td char=".">.934</td><td char=".">.066</td><td char=".">.014</td><td char="."><.001</td></tr><tr><td>I can talk about my problems with my family.</td><td char=".">.887</td><td char=".">.046</td><td char="."><.001</td><td char=".">.790</td><td char=".">.209</td><td char=".">.039</td><td char="."><.001</td></tr><tr><td>My family is willing to help me make decisions.</td><td char=".">.911</td><td char=".">.046</td><td char="."><.001</td><td char=".">.834</td><td char=".">.165</td><td char=".">.029</td><td char="."><.001</td></tr><tr><td><italic>Friend</italic></td><td /><td /><td /><td /><td /><td /><td /></tr><tr><td>My friends really try to help me.</td><td char=".">.910</td><td char=".">.058</td><td char="."><.001</td><td char=".">.833</td><td char=".">.166</td><td char=".">.036</td><td char="."><.001</td></tr><tr><td>I can count on my friends when things go wrong.</td><td char=".">.919</td><td char=".">.056</td><td char="."><.001</td><td char=".">.849</td><td char=".">.151</td><td char=".">.024</td><td char="."><.001</td></tr><tr><td>I have friends with whom I can share my joys and sorrows.</td><td char=".">.884</td><td char=".">.057</td><td char="."><.001</td><td char=".">.785</td><td char=".">.214</td><td char=".">.044</td><td char="."><.001</td></tr><tr><td>I can talk about my problems with my friends.</td><td char=".">.853</td><td char=".">.059</td><td char="."><.001</td><td char=".">.731</td><td char=".">.268</td><td char=".">.062</td><td char="."><.001</td></tr><tr><td><italic>Factor Correlations</italic></td><td>Correlation</td><td>Standard Error</td><td><italic>p</italic>-value</td><td /><td /><td /><td /></tr><tr><td>Significant Other - Family</td><td char=".">.620</td><td char=".">.059</td><td char="."><.001</td><td /><td /><td /><td /></tr><tr><td>Significant Other - Friend</td><td char=".">.611</td><td char=".">.058</td><td char="."><.001</td><td /><td /><td /><td /></tr><tr><td>Family - Friend</td><td char=".">.550</td><td char=".">.065</td><td char="."><.001</td><td /><td /><td /><td /></tr></tbody></table> </ephtml> </p> <p>Measurement invariance was also examined across single and partnered individuals as well as individuals with a single health condition and individuals with multiple heath conditions. Results are presented in Table 4. The configural model for single/partnered, in which factor loadings and intercepts were free to vary across groups, displayed good data-model fit: <sups>2</sups>(<reflink idref="bib12" id="ref115">12</reflink>) = 12.66, <emph>p</emph> =.394; CFI = 0.998, TLI =.997, RMSEA = 0.023, 90% CI<subs>RMSEA</subs> = (.000,.085), SRMR =.016, which indicates the same factor structure holds across singled and partnered individuals. The metric invariance model, which constrained the factor loadings to be equal across groups, also displayed good data-model fit: <sups>2</sups>(<reflink idref="bib17" id="ref116">17</reflink>) = 15.12, <emph>p</emph> =.587; CFI = 1.00, TLI = 1.00, RMSEA = 0.000, 90% CI<subs>RMSEA</subs> = (.000,.079), SRMR =.039, and did not display significantly worse fit when compared with the configural model, Δ<sups>2</sups>(<reflink idref="bib5" id="ref117">5</reflink>) = 2.06, <emph>p</emph> =.840, ΔCFI =.001, ΔRMSEA = −0.023. Similarly, the scalar model, which constrained the factor loading and intercepts to be equal across groups, displayed good data-model fit, <sups>2</sups>(<reflink idref="bib22" id="ref118">22</reflink>) = 22.92, <emph>p</emph> =.407; CFI =.998, TLI =.998, RMSEA =.020, 90% CI<subs>RMSEA</subs> = (.000,.072), SRMR =.046, and was not significantly worse than the metric model, Δ2(<reflink idref="bib5" id="ref119">5</reflink>) = 8.48, <emph>p</emph> =.132, ΔCFI = −0.002, ΔRMSEA =.020. Thus, full scalar invariance was achieved for BRS across single and partnered individuals.</p> <p>Table 4. Model Comparisons for Validation of Measurement Invariance for BRS and MSPSS.</p> <p> <ephtml> <table><thead><tr><td>Model</td><td>χ<sup>2</sup></td><td>df</td><td>CFI</td><td>RMSEA</td><td>AIC</td><td>Δχ<sup>2</sup></td><td>Δdf</td><td>ΔCFI</td><td>ΔRMSEA</td></tr></thead><tbody valign="top"><tr><td><italic>BRS – Single/partnered</italic></td></tr><tr><td>Configural</td><td char=".">12.66</td><td char=".">12</td><td char=".">.998</td><td char=".">.023</td><td char=".">3134</td><td>—</td><td>—</td><td>—</td><td>—</td></tr><tr><td>Metric</td><td char=".">15.12</td><td char=".">17</td><td char=".">1.00</td><td char=".">.000</td><td char=".">3127</td><td char=".">2.06</td><td char=".">5</td><td char=".">.001</td><td char=".">−.023</td></tr><tr><td>Scalar</td><td char=".">22.92</td><td char=".">22</td><td char=".">.998</td><td char=".">.020</td><td char=".">3124</td><td char=".">8.48</td><td char=".">5</td><td char=".">−.002</td><td char=".">.020</td></tr><tr><td><italic>BRS – Single health condition/multiple health conditions</italic></td></tr><tr><td>Configural</td><td char=".">17.56</td><td char=".">12</td><td char=".">.986</td><td char=".">.062</td><td char=".">3107</td><td>—</td><td>—</td><td>—</td><td>—</td></tr><tr><td>Metric</td><td char=".">22.96</td><td char=".">17</td><td char=".">.987</td><td char=".">.050</td><td char=".">3012</td><td char=".">4.76</td><td char=".">5</td><td char=".">.001</td><td char=".">−.012</td></tr><tr><td>Full Scalar</td><td char=".">43.57**</td><td char=".">22</td><td char=".">.938</td><td char=".">.095</td><td char=".">3110</td><td char=".">24.40***</td><td char=".">5</td><td char=".">−.049</td><td char=".">.045</td></tr><tr><td>Partial Scalar (item 3)<sup>a</sup></td><td char=".">26.20</td><td char=".">21</td><td char=".">.988</td><td char=".">.042</td><td char=".">3098</td><td char=".">3.34</td><td char=".">4</td><td char=".">.001</td><td char=".">−.008</td></tr><tr><td><italic>MSPSS – Single/partnered</italic></td></tr><tr><td>Configural</td><td char=".">297.62***</td><td char=".">102</td><td char=".">.915</td><td char=".">.118</td><td char=".">7269</td><td>—</td><td>—</td><td>—</td><td>—</td></tr><tr><td>Metric</td><td char=".">300.24***</td><td char=".">111</td><td char=".">.918</td><td char=".">.111</td><td char=".">7254</td><td char=".">2.50</td><td char=".">9</td><td char=".">.002</td><td char=".">−.007</td></tr><tr><td>Scalar</td><td char=".">320.02***</td><td char=".">120</td><td char=".">.913</td><td char=".">.110</td><td char=".">7255</td><td char=".">16.64</td><td char=".">9</td><td char=".">−.004</td><td char=".">−.001</td></tr><tr><td><italic>MSPSS – Single health condition/multiple health conditions</italic></td></tr><tr><td>Configural</td><td char=".">233.66***</td><td char=".">102</td><td char=".">.957</td><td char=".">.086</td><td char=".">7338</td><td>—</td><td>—</td><td>—</td><td>—</td></tr><tr><td>Metric</td><td char=".">240.12***</td><td char=".">111</td><td char=".">.957</td><td char=".">.082</td><td char=".">7327</td><td char=".">6.31</td><td char=".">9</td><td char=".">.000</td><td char=".">−.004</td></tr><tr><td>Scalar</td><td char=".">252.05***</td><td char=".">120</td><td char=".">.955</td><td char=".">.081</td><td char=".">7321</td><td char=".">12.18</td><td char=".">9</td><td char=".">−.002</td><td char=".">−.001</td></tr></tbody></table> </ephtml> </p> <ulist> <item>2 Note. χ<sups>2</sups>, scaled robust maximum likelihood chi-square test; CFI, comparative fit index; RMSEA, root mean square error of approximation; AIC, Akaike information criterion; Δχ<sups>2</sups>, scaled.</item> <item>3 chi-squared difference test for MLR estimator; ΔCFI, change in CFI; ΔRMESA, change in RMESA; **<emph>p</emph> <.01, ***<emph>p</emph> <.001; <sups>a</sups>intercepts in item listed in parenthesis are free to vary between groups.</item> </ulist> <p>The configural model for single/multiple heath conditions displayed adequate data-model fit, <sups>2</sups>(<reflink idref="bib12" id="ref120">12</reflink>) = 17.56, <emph>p</emph> =.130; CFI = 0.986, TLI =.965, RMSEA = 0.062, 90% CI<subs>RMSEA</subs> = (.000,.124), SRMR =.022, as did the metric invariance model, <sups>2</sups>(<reflink idref="bib17" id="ref121">17</reflink>) = 22.96, <emph>p</emph> =.151; CFI =.987, TLI =.977, RMSEA = 0.050, 90% CI<subs>RMSEA</subs> = (.000,.105), SRMR =.065. Moreover, the metric model did not display significantly worse fit when compared with the configural model, Δ<sups>2</sups>(<reflink idref="bib5" id="ref122">5</reflink>) = 4.76, <emph>p</emph> =.446, ΔCFI =.001, ΔRMSEA = −0.012. The full scalar model had less than ideal data-model fit: <sups>2</sups>(<reflink idref="bib22" id="ref123">22</reflink>) = 43.57, <emph>p</emph> =.004; CFI =.938, TLI =.916, RMSEA =.095, 90% CI<subs>RMSEA</subs> = (.051,.137), SRMR =.092, and fit significantly worse than the metric model, Δ<sups>2</sups>(<reflink idref="bib5" id="ref124">5</reflink>) = 24.40, <emph>p</emph> <.001, ΔCFI = −0.049, ΔRMSEA =.045. Modification indices suggested item 3 (i.e. it does not take me long to recover from a stressful event) as the source of misfit, so the equality constraint on the item intercept was removed and the model was re-estimated. Results demonstrated adequate model fit <sups>2</sups>(<reflink idref="bib21" id="ref125">21</reflink>) = 26.20, <emph>p</emph> =.199; CFI =.988, TLI =.984, RMSEA =.042, 90% CI<subs>RMSEA</subs> = (.000,.096), SRMR =.069, and fit was not significantly worse than the metric model, Δ<sups>2</sups>(<reflink idref="bib4" id="ref126">4</reflink>) = 3.34, <emph>p</emph> =.502, ΔCFI =.001, ΔRMSEA = −0.008. Thus, partial scalar invariance was achieved.</p> <hd id="AN0182245067-19">Multidimensional Scale of Perceived Social Support</hd> <p>A three-factor CFA model was fit to the MSPSS data in which four items were loaded onto the first factor (Significant Other), four items loaded onto the second factor (Family), and four items loaded onto the third factor (Friend). The factors were allowed to correlate. Results indicated adequate data-model fit: <sups>2</sups>(<reflink idref="bib51" id="ref127">51</reflink>) = 118.81, <emph>p</emph> <.001; CFI = 0.959, TLI = 0.947, RMSEA = 0.080, 90% CI<subs>RMSEA</subs> = (.064,.096), SRMR = 0.037. Although an examination of residuals and modification indices indicated that model fit could be improved by including an error covariance between Item 1 ("There is a special person who is around when I am in need") and Item 2 ("There is a special person with whom I can share my joys and sorrows") in the Significant Other factor, we decided to retain the current model to maintain fidelity to the theoretical model being tested. Local fit was examined using residual correlations and found two residual correlation greater than |0.10| and none were greater than |.11|. Standardized factor loadings and R-squared values are presented in Table 3. Factor loadings for all items were strong. Loadings for Significant Other ranged from 0.837 to 0.940. Factor loading for Family ranged from 0.887 to 0.964, and loadings for Friend ranged from 0.853 to 0.919. R-squared values for the items were all greater than 0.70. The average variance extracted for each factor was 0.786 for Significant Other, 0.863 for Family, and 0.796 for Friend. Correlations among factors (Table 4) were 0.620 for Significant Other and Family, 0.611 for Significant Other and Friend, and 0.550 for Family and Friend. Cronbach's alpha and McDonald's coefficient omega internal consistency reliability measures were 0.937 and 0.937, respectively, for Significant Other; 0.962 and 0.962 for Family; and.940 and 0.940 for Friend.</p> <p>Standardized factor loadings, the average variance extracted, and correlations among factors can be used to provide evidence of convergent and discriminant validity (Fornell & Larcker, [<reflink idref="bib25" id="ref128">25</reflink>]). Fornell and Larcker ([<reflink idref="bib25" id="ref129">25</reflink>]) suggest that evidence for convergent validity can be obtained when a latent construct accounts for at least half of the variance in its associated indicators. Cheung and Wang ([<reflink idref="bib12" id="ref130">12</reflink>]) extend this idea and argue that both the average variance extracted and standardized factor loadings for all items should be greater than 0.50 to provide evidence for convergent validity. Given these criteria, all three factors for the MSPSS demonstrate adequate convergent validity. Cheung and Wang ([<reflink idref="bib12" id="ref131">12</reflink>]) argue that discriminant validity can be established if the correlation between any two factors is less than 0.70. This criterion ensures that while the factors are related, they are distinct from one another and measure different constructs. Correlations among the three factors of MSPSS were all less than 0.70 suggesting evidence for discriminant validity for the three MSPSS factors.</p> <p>Measurement invariance (Table 4) was also examined across two groups: single and partnered individuals, and individuals with a single health condition versus individuals with multiple heath conditions. For single/partnered individuals the configural fit was as follows: <sups>2</sups>(<reflink idref="bib102" id="ref132">102</reflink>) = 297.62, <emph>p</emph> <.001; CFI = 0.915, TLI =.890, RMSEA = 0.118, 90% CI<subs>RMSEA</subs> = (.101,.135), SRMR =.051. The configural model did not meet established guidelines for adequate data-model fit. Modification indices again suggested that model fit could be improved by including an error covariance between Item 1 ("There is a special person who is around when I am in need") and Item 2 ("There is a special person with whom I can share my joys and sorrows") in the Significant Other factor. However, we decided to proceeded with the measurement invariance investigation using the original model rather than make modifications in order to stay true to the original factor structure. The metric invariance model displayed similar data-model fit, <sups>2</sups>(<reflink idref="bib111" id="ref133">111</reflink>) = 300.24, <emph>p</emph> <.001; CFI =.918, TLI =.902, RMSEA = 0.111, 90% CI<subs>RMSEA</subs> = (.095,.128), SRMR =.054, and did not significantly differ from the configural model, Δ<sups>2</sups>(<reflink idref="bib9" id="ref134">9</reflink>) = 2.50, <emph>p</emph> =.981, ΔCFI =.002, ΔRMSEA = −0.007. The fit for the scalar invariance model was <sups>2</sups>(<reflink idref="bib120" id="ref135">120</reflink>) = 320.02, <emph>p</emph> <.001; CFI =.913, TLI =.904, RMSEA = 0.110, 90% CI<subs>RMSEA</subs> = (.094,.126), SRMR =.057. Model comparison tests indicated a similar fit to the data as the metric model: Δ<sups>2</sups>(<reflink idref="bib9" id="ref136">9</reflink>) = 16.64, <emph>p</emph> =.055, ΔCFI = −0.004, ΔRMSEA = −0.001.</p> <p>For individuals with single/multiple heath conditions, the configural fit was <sups>2</sups>(<reflink idref="bib102" id="ref137">102</reflink>) = 233.657, <emph>p</emph> <.001; CFI = 0.957, TLI =.944, RMSEA = 0.086, 90% CI<subs>RMSEA</subs> = (.068,.104), SRMR =.039. The metric model for the single/multiple health comparison was <sups>2</sups>(<reflink idref="bib111" id="ref138">111</reflink>) = 240.12, <emph>p</emph> <.001; CFI =.957, TLI =.949, RMSEA = 0.082, 90% CI<subs>RMSEA</subs> = (.064,.100), SRMR =.045, and did not display significantly worse fit: Δ<sups>2</sups>(<reflink idref="bib9" id="ref139">9</reflink>) = 6.31, <emph>p</emph> =.708, ΔCFI =.000, ΔRMSEA = −0.004. Finally, the scalar fit for single/multiple health conditions was, <sups>2</sups>(<reflink idref="bib120" id="ref140">120</reflink>) = 252.05, <emph>p</emph> <.001; CFI =.955, TLI =.951, RMSEA = 0.081, 90% CI<subs>RMSEA</subs> = (.063,.098), SRMR =.047 and did not display worse fit compared with the metric model: Δ<sups>2</sups>(<reflink idref="bib9" id="ref141">9</reflink>) = 12.18, <emph>p</emph> =.203, ΔCFI = −0.002, ΔRMSEA = −0.001. Thus, full scalar invariance was achieved for MSPSS across single and partnered individuals as well as individuals with single and multiple health conditions.</p> <hd id="AN0182245067-20">Discussion</hd> <p>The findings of this study supported a unidimensional factor structure for the BRS, after taking into account method effects attributable to item wording, and a three-factor structure for the MSPSS among older SMW with disabilities. Both scales also had good internal consistency (i.e. alpha and omega >.80) and exhibited factor structures analogous to those observed in other populations. Taken together, these findings suggest that the sample's scores on the BRS and MSPSS may have adequate reliability and validity, providing evidence to support using these scales in future research with older SMW with disabilities.</p> <p>The factor structure of the BRS in our sample suggests that scores on the BRS measure a unidimensional construct, focusing on one specific element of resilience (i.e. the ability to bounce back from adverse experiences; Smith et al., [<reflink idref="bib57" id="ref142">57</reflink>]). However, adequate model fit was only achieved after adjusting for effects due to item wording (i.e. accounting for reverse-scored items having negatively worded language). While the original BRS was proposed to be unidimensional, the scale was developed based solely on the results of an exploratory factor analysis. However, subsequent validation studies using CFA (e.g. Chmitorz et al., [<reflink idref="bib13" id="ref143">13</reflink>]; Ye et al., [<reflink idref="bib65" id="ref144">65</reflink>]) have suggested the BRS factor structure to include method effects. Other studies (e.g. Fung, [<reflink idref="bib34" id="ref145">34</reflink>]) have also found good fit with a two-factor model that groups positive and negative items into separated, correlated factors. To stay true to the original intent of the BRS (Smith et al., [<reflink idref="bib57" id="ref146">57</reflink>]), we estimated a model that included a method factor to account for item wording. Future research should find ways to reduce the method effects within the BRS. Model fit statistics were similar to the fit statistics reported by Chmitorz et al. ([<reflink idref="bib13" id="ref147">13</reflink>]), Fung ([<reflink idref="bib34" id="ref148">34</reflink>]), and Ye et al. ([<reflink idref="bib65" id="ref149">65</reflink>]). The factor loadings, while generally similar, tended to be slightly higher for the positively worded items than reported in the other studies. Both Cronbach's alpha (0.863) and McDonald's omega (0.864) indicated tentative strong reliability for scores on the BRS (Kalkbrenner, [<reflink idref="bib39" id="ref150">39</reflink>]). McDonald's omega may be a slightly better estimate of internal consistency on scores produced by the BRS for older SMW with disabilities due to method effects (Kalkbrenner, [<reflink idref="bib39" id="ref151">39</reflink>]), although both alpha and omega were so similar as to be virtually indistinguishable.</p> <p>The BRS showed satisfactory evidence for measurement invariance across single and partnered individuals as well as individuals with single and multiple health conditions. The configural model confirmed the factor structure in each group; while the metric model confirmed that group members interpreted the items in a similar way. Full scalar invariance held for the single/partner comparison indicating that latent scores across groups can be compared. Partial scalar invariance was achieved for the single/multiple health condition comparison. It is noteworthy that only one of the six items were found to be noninvariant. Dimitrov ([<reflink idref="bib19" id="ref152">19</reflink>]) argues partial invariance can be achieved if less than 20% of parameters are freed. However, attention should be paid to item 3 ("<emph>It does not take me long to recover from a stressful event</emph>"), as it was the only nonequivalent item across the health condition comparison.</p> <p>Similarly, MSPSS scores in our sample had excellent internal consistency and factorability with evidence of convergent and discriminant validity. Similar model fit statistics and factor loadings were found by De Maria et al. ([<reflink idref="bib17" id="ref153">17</reflink>]) who examined the MSPSS in a sample of patients with chronic diseases in Italy as well as Calderón et al. ([<reflink idref="bib5" id="ref154">5</reflink>]) when examining the MSPSS in a sample of cancer patients in Spain. Again, Cronbach's alpha and McDonald's omega were similar for the MSPSS and its subscales (e.g. alpha =.937 and omega =.908 for the total scale). However, because the MSPSS is not unidimensional and the sample size was relatively small, McDonald's omega may be a slightly better estimate of internal consistency of scores on the MSPSS for the current sample (Kalkbrenner, [<reflink idref="bib39" id="ref155">39</reflink>]).</p> <p>Full scalar invariance was achieved when comparing individuals with single and multiple health conditions. Thus, the MSPSS and BRS appear to operate in a similar manner for those with a single versus multiple health conditions. Although the measurement invariance findings for the MSPSS showed full scalar invariance across single/partnered individuals, conclusions should be interpreted with caution. The configural model for single/partnered individuals had less than ideal fit. Although, the fit could be improved to an acceptable level by modifying the model (e.g. adding an error covariance), we decided to proceed with the measurement invariance analysis staying true to the original factor structure. Additional research should be done examine the measurement invariance properties of the MSPSS in older SMW with disabilities. Until then, counselors and other users of the MSPSS should exercise caution when using scores on the MSPSS to compare single and partnered individuals. This may be due in part to the inclusion of the "significant other" subscale, which has been identified as potentially problematic with other subgroups (i.e. adolescents in Hong Kong; Cheng & Chan, [<reflink idref="bib10" id="ref156">10</reflink>]).</p> <hd id="AN0182245067-21">Implications</hd> <p>As counselors became eligible for reimbursement under the Medicare program on January 1, 2024 (Consolidated Appropriations Act, [<reflink idref="bib16" id="ref157">16</reflink>]), the counseling profession can advance health equity by serving the increasingly diverse population of older people in the U.S. (Administration on Aging, [<reflink idref="bib1" id="ref158">1</reflink>]), including older SMW with disabilities. In fact, by the second quarter of 2024, more than 36,000 licensed mental health counselors enrolled in the Medicare program as providers (Centers for Medicare and Medicaid Services, [<reflink idref="bib7" id="ref159">7</reflink>]). However, this opportunity also requires additional research to ensure that counseling professionals can design and implement evidence-based approaches to serving these populations. Therefore, our findings can support counseling research focused on understanding how these protective factors impact specific mental health outcomes, as well as how different counseling interventions change them, among older SMW with disabilities. Indeed, confirming that the original factor structure for both the BRS and MSPSS fit data from older SMW with disabilities, as well as that there was adequate internal consistency on the scores observed in this sample, provides evidence to support using these measures in future research with this population. That may also provide preliminary evidence to support using these to examine resilience and social support in other subgroups of LGBTQ+ older adults. This evidence can support research focused on unique mental health needs and evidence-supported interventions among older SMW with disabilities, which is especially important given recent Medicare reimbursement policy changes.</p> <p>Our findings might also support the use of the BRS and MSPSS in practice with older SMW with disabilities. Counselors working with LGBTQ+ older people are likely to encounter older SMW with disabilities, as nearly half of older SMW report having disabilities (Fredriksen-Goldsen et al., [<reflink idref="bib27" id="ref160">27</reflink>]). Therefore, our findings suggest that counselors may be able to use the BRS and MSPSS as a valid tool for measuring one facet of resilience and overall social support, as well as specific types of support, in their clients. Additionally, our findings suggest that these tools work equally well with older SMW with a single health condition and those with multiple health conditions. The BRS also performed equally well across single and partnered individuals, allowing counselors to use the scale with clients regardless of marital status. However, because results from the MSPSS should be interpreted with caution, counselors might consider differences based on how partnered and unpartnered clients may respond. This is particularly true for the significant other subscale, which is likely influenced by the presence of a partner in one's life. Counselors may consider a different instrument for unpartnered clients in this population. Although neither the BRS nor MSPSS has clinical cutoff scores, these instruments may be effective in progress monitoring if these protective factors are of salient clinical focus. Additionally, practitioners may use scores from our sample as one benchmark of mean scores, although this approach should be used with caution given that our sample is not representative of the general population of older SMW with disabilities. However, because no other benchmarks exist, our findings may provide an initial datapoint for consideration.</p> <hd id="AN0182245067-22">Limitations and Directions for Future Research</hd> <p>This study is not without limitations. Regarding the sample, there are no estimates of the prevalence or characteristics of older SMW with disabilities in the U.S. population. Therefore, it is possible that our sample was not representative of all older SMW with disabilities. Future inquiry should document these characteristics to ensure researchers can recruit representative samples. Furthermore, homogeneity in regard to race/ethnicity is a significant limitation in applying these measures with older SMW of color who have disabilities. There may also be sampling bias related to inclusion criteria. Disability identity was defined broadly to include all people who experience a chronic illness or disability, using a demographic question from the American Community Survey (U.S. Census Bureau, [<reflink idref="bib61" id="ref161">61</reflink>]). However, this definition may have excluded people who experience disabling conditions but do not identify as having a disability; similarly, this definition is inclusive yet does not differentiate between people who qualify for disability services through the government and those who do not. Similarly, because self-identification as a lesbian, bisexual, or otherwise non-heterosexual woman was an inclusion criteria, these findings cannot be applied to all women who have sexual or romantic relationships with people of the same gender. Instead, they should only be applied to older SMW who identify with that language.</p> <p>This study also has some methodological limitations to consider. As stated earlier, the model fit for single/partner comparison was less than ideal and more research should be done to examine whether the MSPSS operates similarly for single and partnered sexually minoritized women. The upper limit of the confidence interval for the RMSEA of the BRS CFA and measurement invariance comparison for single/multiple health conditions, exceeded the exceeded the commonly accepted threshold for adequate model fit (.05 to.08). However, it is important to note that the lower limit of the confidence interval fell well within the range of good fit. This wide confidence interval suggests that there may be uncertainty about the true RMSEA in the population. One possible reason for this wide interval may be sample size (<emph>n</emph><bold></bold>=<bold></bold>208), as smaller samples can lead to wide confidence intervals. Similarly, the more complex the model (e.g. number of parameters relative to the sample size) can also lead to wide confidence intervals. Future researchers should strive to collect larger sample sizes and to reduce the method effects within the BRS, both of which should help to narrow the confidence interval and provide a more precise estimate of the RMSEA.</p> <p>These findings and their limitations highlight important directions for future research. First, further validation of these measures is necessary, such as providing evidence of convergent and discriminant validity, test-retest reliability, and establishing evidence of acceptable psychometric properties with related populations (e.g. LGBTQ+ older adults broadly, LGBTQ+ older adults of color, gender minority older adults). Knowing that the factor structure remains consistent in SMW is a crucial first step toward examining the measurement invariance of these instruments across diverse populations. It is essential for future research to explore the measurement invariance of these instruments to enable accurate comparisons of resilience and support among various groups. Additional attention is needed to better validate other common instruments used in counseling research with LGBTQ+ older adults and associated subgroups, especially given recent Medicare policy changes (Consolidated Appropriations Act, [<reflink idref="bib16" id="ref162">16</reflink>]) that will allow a greater number of counselors to serve this population. Given documented limitations in measuring resilience (Colpitts & Gahagan, [<reflink idref="bib15" id="ref163">15</reflink>]) through existing instruments for LGBTQ+ populations and unique social networks among LGBTQ+ older adults (Kim et al., [<reflink idref="bib42" id="ref164">42</reflink>]), researchers may also consider developing new assessments that better capture these constructs among LGBTQ+ older adults.</p> <hd id="AN0182245067-23">Conclusion</hd> <p>With the inclusion of licensed professional counselors as Medicare providers (Consolidated Appropriations Act, [<reflink idref="bib16" id="ref165">16</reflink>]), counseling professionals will have new opportunities to work with older adults, including LGBTQ+ older people and related subgroups. Additional inquiry is needed to ensure that counselors understand health mechanisms among LGBTQ+ older adults and multiply marginalized subsections, such as older SMW with disabilities. However, LGBTQ+ aging is rarely explored in counseling and psychology research, in part due to the lack of validated instruments to measure relevant constructs with this population. This study provided an important first step in demonstrating psychometric validation of two commonly used instruments in counseling research. The results of CFA, measurement invariance, and measures of internal consistency suggest that scores on the BRS (Smith et al., [<reflink idref="bib57" id="ref166">57</reflink>]) and the MSPSS (Zimet et al., [<reflink idref="bib66" id="ref167">66</reflink>]) among older SMW with disabilities are valid and reliable, providing initial evidence of acceptable use of these instruments in this population. These instruments offer counseling researchers tools to evaluate these constructs as potential protective factors for a range of mental health outcomes for older SMW with disabilities.</p> <hd id="AN0182245067-24">Acknowledgments</hd> <p>The authors would like to acknowledge the Association for Assessment and Research in Counseling for supporting this scholarship through their Sponsored Scholarship Program. The authors would also like to acknowledge Drs. Matthew Fullen, Laura Welfare, Tameka Grimes, David Kniola, and Christian Chan for their support on the broader research project.</p> <hd id="AN0182245067-25">Disclosure statement</hd> <p>No potential conflict of interest was reported by the authors.</p> <ref id="AN0182245067-26"> <title> References </title> <blist> <bibl id="bib1" idref="ref64" type="bt">1</bibl> <bibtext> Administration on Aging. (2022). 2021 profile of older Americans. https://acl.gov/sites/default/files/Profile%20of%20OA/2021%20Profile%20of%20OA/2021ProfileOlderAmericans_508.pdf</bibtext> </blist> <blist> <bibl id="bib2" idref="ref26" type="bt">2</bibl> <bibtext> Alschuler, K. N., Kratz, A. L., & Ehde, D. M. (2016). Resilience and vulnerability in individuals with chronic pain and physical disability. Rehabilitation Psychology, 61 (1), 7 – 18. https://doi.org/10.1037/rep0000055</bibtext> </blist> <blist> <bibl id="bib3" idref="ref24" type="bt">3</bibl> <bibtext> Bower, K. L., Lewis, D. C., Bermúdez, J. M., & Singh, A. A. (2021). Narratives of generativity and resilience among LGBT older adults: Leaving positive legacies despite social stigma and collective trauma. Journal of Homosexuality, 68 (2), 230 – 251. https://doi.org/10.1080/00918369.2019.1648082</bibtext> </blist> <blist> <bibl id="bib4" idref="ref70" type="bt">4</bibl> <bibtext> Brown, T. A. (2015). Confirmatory factor analysis for applied research (2nd ed.). Guilford Press.</bibtext> </blist> <blist> <bibl id="bib5" idref="ref71" type="bt">5</bibl> <bibtext> Calderón, C., Ferrando, P. J., Lorenzo-Seva, U., Gómez-Sánchez, D., Fernández-Montes, A., Palacín-Lois, M., Antoñanzas-Basa, M., Rogado, J., Manzano-Fernández, A., Ferreira, E., Asensio-Martínez, E., & Jiménez-Fonseca, P. (2021). Multidimensional scale of perceived social support (MSPSS) in cancer patients: Psychometric properties and measurement invariance. Psicothema, 33 (1), 131 – 138. https://doi.org/10.7334/psicothema2020.263</bibtext> </blist> <blist> <bibl id="bib6" idref="ref25" type="bt">6</bibl> <bibtext> Candrian, C., Wright, L., Callwood, A., Condolora, J., Kuka, K., Sills, J., Thiele, P., Tropeano, L., Kline, D., & Mullen, R. (2022). In her own words: Understanding the health needs of older lesbians. Journal of Women & Aging, 34 (5), 596 – 604. https://doi.org/10.1016/j.jaging.2012.09.001</bibtext> </blist> <blist> <bibl id="bib7" idref="ref73" type="bt">7</bibl> <bibtext> Centers for Medicare and Medicaid Services. (2024). Medicare fee-for-service public provider enrollment. Data.cms.gov. Retrieved August 1, 2024, from https://data.cms.gov/provider-characteristics/medicare-provider-supplier-enrollment/medicare-fee-for-service-public-provider-enrollment/</bibtext> </blist> <blist> <bibl id="bib8" idref="ref74" type="bt">8</bibl> <bibtext> Chandler, J., Rosenzweig, C., Moss, A. J., Robinson, J., & Litman, L. (2019). Online panels in social science research: Expanding sampling methods beyond Mechanical Turk. Behavior Research Methods, 51 (5), 2022 – 2038. https://doi.org/10.3758/s13428-019-01273-7</bibtext> </blist> <blist> <bibl id="bib9" idref="ref75" type="bt">9</bibl> <bibtext> Chen, F. F. (2007). Sensitivity of goodness of fit indexes to lack of measurement invariance. Structural Equation Modeling: A Multidisciplinary Journal, 14 (3), 464 – 504. https://doi.org/10.1080/10705510701301834</bibtext> </blist> <blist> <bibtext> Cheng, S.-T., & Chan, A. C. M. (2004). The multidimensional scale of perceived social support: Dimensionality and age and gender differences in adolescents. Personality and Individual Differences, 37 (7), 1359 – 1369. https://doi.org/10.1016/j.paid.2004.01.006</bibtext> </blist> <blist> <bibtext> Cheung, G. W., & Rensvold, R. B. (2002). Evaluating goodness-of-fit indexes for testing measurement invariance. Structural Equation Modeling: A Multidisciplinary Journal, 9 (2), 233 – 255. https://doi.org/10.1207/S15328007SEM0902_5</bibtext> </blist> <blist> <bibtext> Cheung, G. W., & Wang, C. (2017). Current approaches for assessing convergent and discriminant validity with SEM: Issues and solutions. Academy of Management Proceedings, 2017 (1), 12706. https://doi.org/10.5465/AMBPP.2017.12706abstract</bibtext> </blist> <blist> <bibtext> Chmitorz, A., Wenzel, M., Stieglitz, R. D., Kunzler, A., Bagusat, C., Helmreich, I., Gerlicher, A., Kampa, M., Kubiak, T., Kalisch, R., Lieb, K., & Tüscher, O. (2018). Population-based validation of a German version of the brief resilience scale. PLoS One, 13 (2), e0192761. https://doi.org/10.1371/journal.pone.0192761</bibtext> </blist> <blist> <bibtext> CloudResearch (n.d). Prime panels. Retrieved from https://<ulink href="http://www.cloudresearch.com/products/prime-panels/">www.cloudresearch.com/products/prime-panels/</ulink></bibtext> </blist> <blist> <bibtext> Colpitts, E., & Gahagan, J. (2016). The utility of resilience as a conceptual framework for understanding and measuring LGBTQ health. International Journal for Equity in Health, 15 (1), 60. https://doi.org/10.1186/s12939-016-0349-1</bibtext> </blist> <blist> <bibtext> Consolidated Appropriations Act. (2023). Pub. L. 117-328, 136 Stat. 4459.</bibtext> </blist> <blist> <bibtext> De Maria, M., Vellone, E., Durante, A., Biagioli, V., & Matarese, M. (2018). Psychometric evaluation of the multidimensional scale of perceived social support (MSPSS) in people with chronic diseases. Annali Dell'Istituto Superiore di Sanita, 54 (4), 308 – 315. https://doi.org/10.4415/ANN_18_04_07</bibtext> </blist> <blist> <bibtext> Dejanipont, B., Wang, C., Jenkins, S., Ta, A., Prybutok, V., & Vosvick, M. (2023). Stressor appraisals and moderating role of forgiveness, social support, and resilience as adaptive coping in stress and depression among older sexual minorities. Sexuality Research and Social Policy. Advance online publication. https://doi.org/10.1007/s13178-023-00831-1</bibtext> </blist> <blist> <bibtext> Dimitrov, D. M. (2010). Testing for factorial invariance in the context of construct validation. Measurement and Evaluation in Counseling and Development, 43 (2), 121 – 149. https://doi.org/10.1177/0748175610373459</bibtext> </blist> <blist> <bibtext> Dispenza, F., Varney, M., & Golubovic, N. (2017). Psychological and counseling practices with sexual and gender minority persons living with chronic illnesses and disabilities (CID). Psychology of Sexual Orientation and Gender Diversity, 4 (1), 137 – 142. https://doi.org/10.1037/sgd0000212</bibtext> </blist> <blist> <bibtext> Eadeh, H.-M., Breaux, R., Boyd-Rogers, C., Priest, J. B., & Nikolas, M. A. (2023). Self-regulation in gender and sexual orientation diverse adults: Exploring patterns of risk and resilience using a person-centered approach. Journal of LGBTQ Issues in Counseling, 17 (1), 22 – 39. https://doi.org/10.1080/26924951.2022.2093310</bibtext> </blist> <blist> <bibtext> Eid, M. A. (2000). Multitrait-multimethod model with minimal assumptions. Psychometrika, 65 (2), 241 – 261. https://doi.org/10.1007/BF02294377</bibtext> </blist> <blist> <bibtext> Emerson, E., Fortune, N., Llewellyn, G., & Stancliffe, R. (2021). Loneliness, social support, social isolation and wellbeing among working age adults with and without disability: Cross-sectional study. Disability and Health Journal, 14 (1), e100965. https://doi.org/10.1016/j.dhjo.2020.100965</bibtext> </blist> <blist> <bibtext> Enders, C. K. (2010). Applied missing data analysis. The Guilford Press.</bibtext> </blist> <blist> <bibtext> Fornell, C., & Larcker, D. F. (1981). Evaluating structural equation models with unobservable variables and measurement error. Journal of Marketing Research, 18 (1), 39 – 50. https://doi.org/10.1177/002224378101800104</bibtext> </blist> <blist> <bibtext> Fredriksen-Goldsen, K. I. (2014). Despite disparities, most LGBT elders are aging well. Aging Today: The Bimonthly Newspaper of the American Society on Aging, 35 (3). https://<ulink href="http://www.ncbi.nlm.nih.gov/pmc/articles/PMC4243168/">www.ncbi.nlm.nih.gov/pmc/articles/PMC4243168/</ulink></bibtext> </blist> <blist> <bibtext> Fredriksen-Goldsen, K. I., Emlet, C. A., Kim, H.-J., Muraco, A., Erosheva, E. A., Goldsen, J., & Hoy-Ellis, C. P. (2013). The physical and mental health of lesbian, gay, and bisexual (LGB) older adults: The role of key health indicators and protective factors. The Gerontologist, 53 (4), 664 – 675. https://doi.org/10.1093/geront/gns123</bibtext> </blist> <blist> <bibtext> Fredriksen-Goldsen, K. I., Jen, S., & Muraco, A. (2019). Iridescent life course: LGBTQ aging research and blueprint for the future—A systematic review. Gerontology, 65 (3), 253 – 274. https://doi.org/10.1159/000493559</bibtext> </blist> <blist> <bibtext> Fredriksen-Goldsen, K. I., & Kim, H. (2017). The science of conducting research with LGBT older adults–An introduction to aging with pride: National Health, Aging, and Sexuality/Gender Study (NHAS). The Gerontologist, 57 (Suppl 1), S1 – S14. https://doi.org/10.1093/geront/gnw212</bibtext> </blist> <blist> <bibtext> Fredriksen-Goldsen, K. I., Kim, H.-J., Shiu, C., Goldsen, J., & Emlet, C. A. (2015). Successful aging among LGBT older adults: Physical and mental health-related quality of life by age group. The Gerontologist, 55 (1), 154 – 168. https://doi.org/10.1093/geront/gnu081</bibtext> </blist> <blist> <bibtext> Fredriksen-Goldsen, K. I., Kim, H.-J., Shui, C., & Bryan, A. E. B. (2017). Chronic health conditions and key health indicators among lesbian, gay, and bisexual older US adults, 2013-2014. American Journal of Public Health, 107 (8), 1332 – 1338. https://doi.org/10.2105/AJPH.2017.303922</bibtext> </blist> <blist> <bibtext> Fredriksen-Goldsen, K. I., Romanelli, M., Jung, H. H., & Kim, H.-J. (2024). Health, economic, and social disparities among lesbian, gay, bisexual, and sexually diverse adults: Results from a population-based study. Behavioral Medicine (Washington, D.C.), 50 (2), 141 – 152. https://doi.org/10.1080/08964289.2022.2153787</bibtext> </blist> <blist> <bibtext> Fullen, M. C., Gorby, S., Chan, C. D., Dobmeier, R. A., & Jordan, J. (2019). The current state of gerontological counseling research: A 26-year content analysis. Journal of Counseling & Development, 97 (4), 387 – 397. https://doi.org/10.1002/jcad.12287</bibtext> </blist> <blist> <bibtext> Fung, S. F. (2020). Validity of the brief resilience scale and brief resilient coping scale in a Chinese sample. International Journal of Environmental Research and Public Health, 17 (4), 1265. https://doi.org/10.3390/ijerph17041265</bibtext> </blist> <blist> <bibtext> Gana, K., & Broc, G. (2019). Structural equation modeling with Lavaan. Wiley.</bibtext> </blist> <blist> <bibtext> Hayes, A. F., & Coutts, J. J. (2020). Use omega rather than Cronbach's alpha for estimating reliability. But.... Communication Methods and Measures, 14 (1), 1 – 24. https://doi.org/10.1080/19312458.2020.1718629</bibtext> </blist> <blist> <bibtext> Healey, A. C., & Hays, D. G. (2012). A discriminant analysis of gender and counselor professional identity development. Journal of Counseling & Development, 90 (1), 55 – 62. https://doi.org/10.1111/j.1556-6676.2012.00008.x</bibtext> </blist> <blist> <bibtext> Jorgensen, T. D., Pornprasertmanit, S., Schoemann, A. M., & Rosseel, Y. (2022). semTools: Useful tools for structural equation modeling. R package version 0.5-6. https://CRAN.R-project.org/package=semTools</bibtext> </blist> <blist> <bibtext> Kalkbrenner, M. T. (2024). Choosing between Cronbach's coefficient alpha, McDonald's coefficient omega, and coefficient H : Confidence intervals and the advantages and drawbacks of interpretive guidelines. Measurement and Evaluation in Counseling and Development, 57 (2), 93 – 105. https://doi.org/10.1080/07481756.2023.2283637</bibtext> </blist> <blist> <bibtext> Kaplan, D. M., Tarvydas, V. M., & Gladding, S. T. (2014). 20/20: A vision for the future of counseling: The new consensus definition of counseling. Journal of Counseling & Development, 92 (3), 366 – 372. https://doi.org/10.1002/j.1556-6676.2014.00164.x</bibtext> </blist> <blist> <bibtext> Kim, H. J., & Fredriksen-Goldsen, K. I. (2017). Disparities in mental health quality of life between Hispanic and non-Hispanic White LGB midlife and older adults and the influence of lifetime discrimination, social connectedness, socioeconomic status, and perceived stress. Research on Aging, 39 (9), 991 – 1012. https://doi.org/10.1177/0164027516650003</bibtext> </blist> <blist> <bibtext> Kim, H. J., Fredriksen-Goldsen, K. I., Bryan, A. E. B., & Muraco, A. (2017). Social network types and mental health among LGBT older adults. The Gerontologist, 57 (suppl 1), S84 – S94. https://doi.org/10.1093/geront/gnw169</bibtext> </blist> <blist> <bibtext> Kler, S., Arora, S., & Le, T. P. (2022). Validation of the multidimensional scale of perceived social support (MSPSS) among LGBTQ+ people of color. Measurement and Evaluation in Counseling and Development, 56 (4), 283 – 296. https://doi.org/10.1080/07481756.2022.2109174</bibtext> </blist> <blist> <bibtext> Kumar, S., Lee, N. K., Pinkerton, E., Wroblewski, K. E., Lengyel, E., & Tobin, M. (2022). Resilience: A mediator of the negative effects of pandemic-related stress on women's mental health in the USA. Archives of Women's Mental Health, 25 (1), 137 – 146. https://doi.org/10.1007/s00737-021-01184-7</bibtext> </blist> <blist> <bibtext> Lei, P., & Shiverdecker, L. (2020). Performance of estimators for confirmatory factor analysis of ordinal variables with missing data. Structural Equation Modeling: A Multidisciplinary Journal, 27 (4), 584 – 601. https://doi.org/10.1080/10705511.2019.1680292</bibtext> </blist> <blist> <bibtext> Lenz, A. S., Gómez Soler, I., Dell'Aquilla, J., & Uribe, P. M. (2017). Translation and cross-cultural adaptation of assessments for use in counseling research. Measurement and Evaluation in Counseling and Development, 50 (4), 224 – 231. https://doi.org/10.1080/07481756.2017.1320947</bibtext> </blist> <blist> <bibtext> Lyons, A., Alba, B., Waling, A., Minichiello, V., Hughes, M., Barrett, C., Fredriksen-Goldsen, K. I., Savage, T., & Edmonds, S. (2022). Assessing the combined effect of ageism and sexuality-related stigma on the mental health and well-being of older lesbian and gay adults. Aging & Mental Health, 26 (7), 1460 – 1469. https://doi.org/10.1080/13607863.2021.1978927</bibtext> </blist> <blist> <bibtext> Marmo, S., Pardasani, M., & Vincent, D. (2021). Senior centers and LGBTQ participants: Engaging older adults virtually in a pandemic. Journal of Gerontological Social Work, 64 (8), 864 – 884. https://doi.org/10.1080/01634372.2021.1937431</bibtext> </blist> <blist> <bibtext> Masa, R., Inoue, M., Prieto, L., Baruah, D., Nosrat, S., Mehak, S., & Operario, D. (2024). Mental health of older adults by sexual minority status: Evidence from the 2021 National Health Interview Survey. Journal of Applied Gerontology: The Official Journal of the Southern Gerontological Society, 43 (3), 276 – 286. https://doi.org/10.1177/07334648231203838</bibtext> </blist> <blist> <bibtext> McNeish, D. (2018). Thanks coefficient alpha, we'll take it from here. Psychological Methods, 23 (3), 412 – 433. https://doi.org/10.1037/met0000144</bibtext> </blist> <blist> <bibtext> Pérez-Villalobos, C., Briede-Westermeyer, J. C., Schilling-Norman, M. J., & Contreras-Espinoza, S. (2021). Multidimensional scale of perceived social support: Evidence of validity and reliability in a Chilean adaptation for older adults. BMC Geriatrics, 21 (1), 461. https://doi.org/10.1186/s12877-021-02404-6</bibtext> </blist> <blist> <bibtext> Reyes, M. E. S., Bautista, N. B., Betos, G. R. A., Martin, K. I. S., Sapio, S. T. N., Pacquing, M. C. T., & Kliatchko, J. M. R. (2022). In/out of the closet: Perceived social support and outness among LGB youth. Sexuality & Culture, 27 (1), 290 – 309. https://doi.org/10.1007/s12119-022-10013-7</bibtext> </blist> <blist> <bibtext> Rosseel, Y. (2012). lavaan: An R package for structural equation modeling. Journal of Statistical Software, 48 (2), 1 – 36. https://doi.org/10.18637/jss.v048.i02</bibtext> </blist> <blist> <bibtext> Sharma, J., McDonald, C. P., Bledsoe, K. G., Grad, R. I., Jenkins, K. D., Moran, D., O'Hara, C., & Pester, D. (2021). Intersectionality in research: Call for inclusive, decolonized, and culturally sensitive research designs in counselor education. Counseling Outcome Research and Evaluation, 12 (2), 63 – 72. https://doi.org/10.1080/21501378.2021.1922075</bibtext> </blist> <blist> <bibtext> Silverman, A. M., Molton, I. R., Alschuler, K. N., Ehde, D. M., & Jensen, M. P. (2015). Resilience predicts functional outcomes in people aging with disability: A longitudinal investigation. Archives of Physical Medicine and Rehabilitation, 96 (7), 1262 – 1268. https://doi.org/10.1016/j.apmr.2015.02.023</bibtext> </blist> <blist> <bibtext> Smart, J. (2019). Disability definitions, diagnoses, and practice implications: An introduction for counselors. Routledge.</bibtext> </blist> <blist> <bibtext> Smith, B. W., Dalen, J., Wiggins, K., Tooley, E., Christopher, P., & Bernard, J. (2008). The brief resilience scale: Assessing the ability to bounce back. International Journal of Behavioral Medicine, 15 (3), 194 – 200. https://doi.org/10.1080/10705500802222972</bibtext> </blist> <blist> <bibtext> Substance Abuse and Mental Health Services Administration. (2019). Risk and protective factors. https://<ulink href="http://www.samhsa.gov/sites/default/files/20190718-samhsa-risk-protective-factors.pdf">www.samhsa.gov/sites/default/files/20190718-samhsa-risk-protective-factors.pdf</ulink></bibtext> </blist> <blist> <bibtext> Swank, J. M., & Mullen, P. M. (2017). Evaluating evidence for conceptually related constructs using bivariate correlations. Measurement and Evaluation in Counseling and Development, 50 (4), 270 – 274. https://doi.org/10.1080/07481756.2017.1339562</bibtext> </blist> <blist> <bibtext> Thorne, N., Kam-Tuck Yip, A., Bouman, W. P., Marshall, E., & Arcelus, J. (2020). The terminology of identities between, outside, and beyond the gender binary—A systematic review. In Joz, M., Nieder, T., & Bouman, W. (Eds.), Non-binary and genderqueer genders. Routledge. https://doi.org/10.4324/9781003015888</bibtext> </blist> <blist> <bibtext> U.S. Census Bureau. (2008). How disability data are collected from the American Community Survey. https://<ulink href="http://www.census.gov/topics/health/disability/guidance/data-collection-acs.html">www.census.gov/topics/health/disability/guidance/data-collection-acs.html</ulink></bibtext> </blist> <blist> <bibtext> Welfare, L. E., Westcott, J. B., McCarty, F., & Smith, H. (2023). Collecting gender identities in counseling research: A content analysis. Journal of LGBTQ Issues in Counseling, 17 (4), 324 – 340. https://doi.org/10.1080/26924951.2023.2247961</bibtext> </blist> <blist> <bibtext> Westcott, J. B. (2023). [ Access to health care at the margins: Implications for older sexual minority women with disabilities ]. [Unpublished doctoral dissertation]. Virginia Tech.</bibtext> </blist> <blist> <bibtext> Westcott, J. B., Epstein, D., Wiley, B., Westcott, J. M., Welfare, L. E., & Catalano, C. (2023). Collecting sexual orientation in counseling research: Implications for counselor education. Counselor Education and Supervision, 62 (4), 384 – 397. https://doi.org/10.1002/ceas.12285</bibtext> </blist> <blist> <bibtext> Ye, Y. C., Wu, C. H., Huang, T. Y., & Yang, C. T. (2022). The difference between the Connor-Davidson Resilience Scale and the Brief Resilience Scale when assessing resilience: Confirmatory factor analysis and predictive effects. Global Mental Health (Cambridge, England), 9, 339 – 346. https://doi.org/10.1017/gmh.2022.38</bibtext> </blist> <blist> <bibtext> Zimet, G. D., Dahlem, N. W., Zimet, S. G., & Farley, G. K. (1988). The multidimensional scale of perceived social support. Journal of Personality Assessment, 52 (1), 30 – 41. https://doi.org/10.1207/s15327752jpa5201_2</bibtext> </blist> </ref> <aug> <p>By Jordan B. Westcott and Louis M. Rocconi</p> <p>Reported by Author; Author</p> <p></p> <p>Dr. Jordan B. Westcott , NCC, is an Assistant Professor of Counselor Education in the Department of Counseling, Human Development, and Family Science at the University of Tennessee, Knoxville. Her research aims to advance access to high-quality, effective, and culturally responsive mental health services for marginalized populations, especially LGBTQ+ communities, older people, and intersections therein. She also studies inclusive research practices for these populations.</p> <p>Dr. Louis M. Rocconi is an Associate Professor and the Program Coordinator for the Evaluation, Statistics, & Methodology PhD program in the Educational Leadership and Policy Studies Department at the University of Tennessee, Knoxville. He applies quantitative methods to solve substantive issues in educational research, and he primarily studies program evaluation and assessment in higher education and methodological issues in educational research.</p> </aug> <nolink nlid="nl1" bibid="bib28" firstref="ref1"></nolink> <nolink nlid="nl2" bibid="bib27" firstref="ref2"></nolink> <nolink nlid="nl3" bibid="bib33" firstref="ref3"></nolink> <nolink nlid="nl4" bibid="bib63" firstref="ref4"></nolink> <nolink nlid="nl5" bibid="bib20" firstref="ref5"></nolink> <nolink nlid="nl6" bibid="bib56" firstref="ref6"></nolink> <nolink nlid="nl7" bibid="bib31" firstref="ref8"></nolink> <nolink nlid="nl8" bibid="bib29" firstref="ref11"></nolink> <nolink nlid="nl9" bibid="bib32" firstref="ref13"></nolink> <nolink nlid="nl10" bibid="bib49" firstref="ref15"></nolink> <nolink nlid="nl11" bibid="bib58" firstref="ref16"></nolink> <nolink nlid="nl12" bibid="bib37" firstref="ref17"></nolink> <nolink nlid="nl13" bibid="bib40" firstref="ref18"></nolink> <nolink nlid="nl14" bibid="bib26" firstref="ref19"></nolink> <nolink nlid="nl15" bibid="bib42" firstref="ref20"></nolink> <nolink nlid="nl16" bibid="bib15" firstref="ref21"></nolink> <nolink nlid="nl17" bibid="bib57" firstref="ref22"></nolink> <nolink nlid="nl18" bibid="bib55" firstref="ref30"></nolink> <nolink nlid="nl19" bibid="bib66" firstref="ref32"></nolink> <nolink nlid="nl20" bibid="bib41" firstref="ref33"></nolink> <nolink nlid="nl21" bibid="bib30" firstref="ref35"></nolink> <nolink nlid="nl22" bibid="bib18" firstref="ref40"></nolink> <nolink nlid="nl23" bibid="bib46" firstref="ref43"></nolink> <nolink nlid="nl24" bibid="bib21" firstref="ref44"></nolink> <nolink nlid="nl25" bibid="bib47" firstref="ref45"></nolink> <nolink nlid="nl26" bibid="bib52" firstref="ref48"></nolink> <nolink nlid="nl27" bibid="bib43" firstref="ref49"></nolink> <nolink nlid="nl28" bibid="bib48" firstref="ref50"></nolink> <nolink nlid="nl29" bibid="bib44" firstref="ref52"></nolink> <nolink nlid="nl30" bibid="bib10" firstref="ref54"></nolink> <nolink nlid="nl31" bibid="bib23" firstref="ref55"></nolink> <nolink nlid="nl32" bibid="bib51" firstref="ref56"></nolink> <nolink nlid="nl33" bibid="bib54" firstref="ref77"></nolink> <nolink nlid="nl34" bibid="bib60" firstref="ref78"></nolink> <nolink nlid="nl35" bibid="bib62" firstref="ref79"></nolink> <nolink nlid="nl36" bibid="bib64" firstref="ref80"></nolink> <nolink nlid="nl37" bibid="bib14" firstref="ref82"></nolink> <nolink nlid="nl38" bibid="bib61" firstref="ref83"></nolink> <nolink nlid="nl39" bibid="bib24" firstref="ref88"></nolink> <nolink nlid="nl40" bibid="bib53" firstref="ref90"></nolink> <nolink nlid="nl41" bibid="bib38" firstref="ref91"></nolink> <nolink nlid="nl42" bibid="bib45" firstref="ref92"></nolink> <nolink nlid="nl43" bibid="bib35" firstref="ref94"></nolink> <nolink nlid="nl44" bibid="bib50" firstref="ref97"></nolink> <nolink nlid="nl45" bibid="bib36" firstref="ref98"></nolink> <nolink nlid="nl46" bibid="bib11" firstref="ref100"></nolink> <nolink nlid="nl47" bibid="bib59" firstref="ref102"></nolink> <nolink nlid="nl48" bibid="bib34" firstref="ref109"></nolink> <nolink nlid="nl49" bibid="bib65" firstref="ref110"></nolink> <nolink nlid="nl50" bibid="bib13" firstref="ref111"></nolink> <nolink nlid="nl51" bibid="bib22" firstref="ref112"></nolink> <nolink nlid="nl52" bibid="bib12" firstref="ref115"></nolink> <nolink nlid="nl53" bibid="bib17" firstref="ref116"></nolink> <nolink nlid="nl54" bibid="bib25" firstref="ref128"></nolink> <nolink nlid="nl55" bibid="bib102" firstref="ref132"></nolink> <nolink nlid="nl56" bibid="bib111" firstref="ref133"></nolink> <nolink nlid="nl57" bibid="bib120" firstref="ref135"></nolink> <nolink nlid="nl58" bibid="bib39" firstref="ref150"></nolink> <nolink nlid="nl59" bibid="bib19" firstref="ref152"></nolink> <nolink nlid="nl60" bibid="bib16" firstref="ref157"></nolink>
Header DbId: eric
DbLabel: ERIC
An: EJ1457725
AccessLevel: 3
PubType: Academic Journal
PubTypeId: academicJournal
PreciseRelevancyScore: 0
IllustrationInfo
Items – Name: Title
  Label: Title
  Group: Ti
  Data: Psychometric Evaluation of the Brief Resilience Scale and Multidimensional Scale of Perceived Social Support among Older Sexual Minority Women in the U.S.
– Name: Language
  Label: Language
  Group: Lang
  Data: English
– Name: Author
  Label: Authors
  Group: Au
  Data: <searchLink fieldCode="AR" term="%22Jordan+B%2E+Westcott%22">Jordan B. Westcott</searchLink><br /><searchLink fieldCode="AR" term="%22Louis+M%2E+Rocconi%22">Louis M. Rocconi</searchLink>
– Name: TitleSource
  Label: Source
  Group: Src
  Data: <searchLink fieldCode="SO" term="%22Measurement+and+Evaluation+in+Counseling+and+Development%22"><i>Measurement and Evaluation in Counseling and Development</i></searchLink>. 2025 58(1):63-82.
– Name: Avail
  Label: Availability
  Group: Avail
  Data: Routledge. Available from: Taylor & Francis, Ltd. 530 Walnut Street Suite 850, Philadelphia, PA 19106. Tel: 800-354-1420; Tel: 215-625-8900; Fax: 215-207-0050; Web site: http://www.tandf.co.uk/journals
– Name: PeerReviewed
  Label: Peer Reviewed
  Group: SrcInfo
  Data: Y
– Name: Pages
  Label: Page Count
  Group: Src
  Data: 20
– Name: DatePubCY
  Label: Publication Date
  Group: Date
  Data: 2025
– Name: TypeDocument
  Label: Document Type
  Group: TypDoc
  Data: Journal Articles<br />Reports - Research
– Name: Subject
  Label: Descriptors
  Group: Su
  Data: <searchLink fieldCode="DE" term="%22Psychometrics%22">Psychometrics</searchLink><br /><searchLink fieldCode="DE" term="%22Resilience+%28Psychology%29%22">Resilience (Psychology)</searchLink><br /><searchLink fieldCode="DE" term="%22Social+Support+Groups%22">Social Support Groups</searchLink><br /><searchLink fieldCode="DE" term="%22Measures+%28Individuals%29%22">Measures (Individuals)</searchLink><br /><searchLink fieldCode="DE" term="%22Older+Adults%22">Older Adults</searchLink><br /><searchLink fieldCode="DE" term="%22Females%22">Females</searchLink><br /><searchLink fieldCode="DE" term="%22LGBTQ+People%22">LGBTQ People</searchLink><br /><searchLink fieldCode="DE" term="%22Disabilities%22">Disabilities</searchLink><br /><searchLink fieldCode="DE" term="%22Factor+Structure%22">Factor Structure</searchLink><br /><searchLink fieldCode="DE" term="%22Factor+Analysis%22">Factor Analysis</searchLink><br /><searchLink fieldCode="DE" term="%22Test+Reliability%22">Test Reliability</searchLink><br /><searchLink fieldCode="DE" term="%22Construct+Validity%22">Construct Validity</searchLink><br /><searchLink fieldCode="DE" term="%22Scores%22">Scores</searchLink>
– Name: DOI
  Label: DOI
  Group: ID
  Data: 10.1080/07481756.2024.2397947
– Name: ISSN
  Label: ISSN
  Group: ISSN
  Data: 0748-1756<br />1947-6302
– Name: Abstract
  Label: Abstract
  Group: Ab
  Data: Objective: This study sought to examine the factor structure, internal consistency, and measurement invariance of the Brief Resilience Scale (BRS) and the Multidimensional Scale of Perceived Social Support (MSPSS) among older sexual minority women with disabilities. Method: Participants (n = 208) consisted of sexual minority women aged 55 and older with disabilities recruited online. Results: CFA results showed a one-factor structure fit BRS data, RMSEA = 0.050 (90% CI [0.00, 0.107]), SRMR = 0.015, CFI = 0.990, TLI = 0.975, with good internal consistency (alpha = 0.863, omega = 0.864). For the MSPSS, CFA showed a three-factor structure fit the data, RMSEA = 0.080 (90% CI [0.064, 0.096]), SRMR = 0.037, CFI = 0.959, TLI = 0.947, with excellent internal consistency (alpha = 0.937, omega = 0.937). Full scalar invariance was achieved for both BRS and MSPSS when comparing single and partnered individuals. Additionally, full scalar invariance was also achieved for MSPSS when comparing individuals with a single and multiple health conditions. For BRS, partial scalar invariance was established across health conditions. Conclusions: The original factor structures of the BRS and MSPSS fit with good internal consistency, providing preliminary evidence of construct validity and reliability of scores on both scales in this population.
– Name: AbstractInfo
  Label: Abstractor
  Group: Ab
  Data: As Provided
– Name: DateEntry
  Label: Entry Date
  Group: Date
  Data: 2025
– Name: AN
  Label: Accession Number
  Group: ID
  Data: EJ1457725
PLink https://search.ebscohost.com/login.aspx?direct=true&site=eds-live&db=eric&AN=EJ1457725
RecordInfo BibRecord:
  BibEntity:
    Identifiers:
      – Type: doi
        Value: 10.1080/07481756.2024.2397947
    Languages:
      – Text: English
    PhysicalDescription:
      Pagination:
        PageCount: 20
        StartPage: 63
    Subjects:
      – SubjectFull: Psychometrics
        Type: general
      – SubjectFull: Resilience (Psychology)
        Type: general
      – SubjectFull: Social Support Groups
        Type: general
      – SubjectFull: Measures (Individuals)
        Type: general
      – SubjectFull: Older Adults
        Type: general
      – SubjectFull: Females
        Type: general
      – SubjectFull: LGBTQ People
        Type: general
      – SubjectFull: Disabilities
        Type: general
      – SubjectFull: Factor Structure
        Type: general
      – SubjectFull: Factor Analysis
        Type: general
      – SubjectFull: Test Reliability
        Type: general
      – SubjectFull: Construct Validity
        Type: general
      – SubjectFull: Scores
        Type: general
    Titles:
      – TitleFull: Psychometric Evaluation of the Brief Resilience Scale and Multidimensional Scale of Perceived Social Support among Older Sexual Minority Women in the U.S.
        Type: main
  BibRelationships:
    HasContributorRelationships:
      – PersonEntity:
          Name:
            NameFull: Jordan B. Westcott
      – PersonEntity:
          Name:
            NameFull: Louis M. Rocconi
    IsPartOfRelationships:
      – BibEntity:
          Dates:
            – D: 01
              M: 01
              Type: published
              Y: 2025
          Identifiers:
            – Type: issn-print
              Value: 0748-1756
            – Type: issn-electronic
              Value: 1947-6302
          Numbering:
            – Type: volume
              Value: 58
            – Type: issue
              Value: 1
          Titles:
            – TitleFull: Measurement and Evaluation in Counseling and Development
              Type: main
ResultId 1