Testing the moderating role of victimization and microaggressions on the relationship between human-animal interaction and psychological adjustment among LGBTQ+ emerging adults

46 Human-animal interaction (HAI) is associated with positive psychological adjustment. Although 47 these benefits are hypothesized to be most pronounced for individuals who experience adversity 48 and compromised social relationships, such as LGBTQ+ (lesbian, gay, bisexual, transgender, 49 queer, and other sexual/gender minority identities) individuals, this hypothesis has not been 50 tested. The current, cross-sectional study examined whether the strength of the relationship 51 between emotional comfort from companion animals and self-esteem and personal hardiness 52 varies as a function of exposure to LGBTQ+ interpersonal stressors (i.e., victimization, 53 microaggressions). Our sample included 155 LGBTQ+ emerging adults who lived with a dog 54 and/or cat in the past year ( M age = 19.34 years, SD = 1.12 years). To test the hypothesis, we 55 conducted simple and multiple moderation analyses. We found evidence that the magnitude of 56 the association between comfort from companion animals and personal hardiness was greater for 57 those who experienced high levels of interpersonal microaggressions. Similarly, victimization 58 moderated the relation between comfort from companion animals and self-esteem. Including 59 victimization and interpersonal microaggressions in the same model resulted in only one 60 significant interaction effect: the relation between comfort from companion animals and self- 61 esteem was positive at high levels of victimization and negative at low levels of victimization. 62 Our results suggest that among LGBTQ+ emerging adults, the benefits of HAI on self-esteem 63 were only present when high levels of victimization were reported. Future research should 64 continue to examine factors that may influence the benefits and risks associated with HAI to 65 identify for whom and under what circumstances HAI is beneficial.

We have no known conflict of interests to disclose. The current study tested the hypothesis that the association between HAI and key aspects 183 of positive psychological adjustment (self-esteem, personal hardiness) in LGBTQ+ emerging 184 adults is moderated by experiences of adversity. Specifically, we tested whether and the degree 185 to which the associations between comfort from companion animals and positive psychological 186 adjustment is moderated by victimization and interpersonal microaggressions in a sample of 187 LGBTQ+ emerging adults. Based on prior assertions that the benefits of HAI may be most 188 pronounced among those who have compromised human relationships and experience adversity, 189 we hypothesized that: (a) comfort from companion animals would be positively associated with 190 self-esteem and personal hardiness and (b) the magnitude of the association between comfort 191 from companion animals and self-esteem and personal hardiness would be greater among those 192 who report more experiences of victimization and interpersonal microaggressions. 193

Methods 194
Participants 195 Participants were recruited as part of an ongoing, longitudinal study of LGBTQ+ stress 196 and supports. Inclusion criteria for the overarching study were: being between the ages of 15 and 197 21 years, understanding spoken English, and self-identifying as LGBTQ+. In this paper, we 198 report on cross-sectional data from the first wave of data collection. Due to the limited number of 199 adolescents in our sample (n = 5), we restricted our sample for the current study to 155 LGBTQ+ 200 emerging adults between the ages of 18 and 21 (Mage = 19.34 years, SD = 1.12; 61.9% 201 racial/ethnic minority) who lived with a pet dog and/or cat in the past 12 months. Approximately 202 47% of participants endorsed a gender minority identity (e.g., transgender, non-binary) and 203 nearly all identified as a sexual minority (98.7%; e.g., asexual, gay, lesbian). Additionally, 46.5% 204 of our sample indicated they were the primary caretaker of the cat(s) and/or dog(s) with whom 205 they lived. More detailed demographic information is provided in Table 1. 206

Procedures 207
All procedures were approved by the first author's university institutional review board 208 (HM20014415). Recruitment and data collection took place from April 2019 to December 2020 209 in an urban, southeastern city of the U.S. Participants were recruited by posting flyers at five 210 local community partner agencies that provide youth with LGBTQ+ inclusive services, online 211 through social media, and through LGBTQ+ organizations' listservs. Participants were also 212 recruited at LGBTQ+ community events (e.g., PRIDE celebration events). Those interested in 213 participating contacted the study's project coordinators by phone or email and completed a 214 screening interview via phone call. Participants who met inclusion criteria then scheduled an 215 interview at a partner agency or at a private office at a local university. To begin the interview, a 216 research assistant described the study to the participants and completed the informed consent 217 process. Participants had the option of completing an online survey by either self-administration 218 using a laptop provided by the study staff member or by having the research team member 219 verbally administer the survey. All participants chose to self-administer the survey. All 220 interviews were conducted online via Zoom (version 5) beginning March 17, 2020, following the 221 onset of the COVID-19 pandemic to adhere to public health recommendations. Nearly 23% of 222 the interviews were conducted following this protocol. Although there were crisis protocols 223 established for in-person and virtual interviews to provide guidance if participants indicated 224 threat of harm to themselves or others, suicidal ideation, or extreme mental distress, no 225 participants demonstrated distress that required the use of these crisis protocols. Additionally, a 226 list of mental health and animal welfare resources was shared with all participants following the 227 completion of the interview. 228

Measures 229
Self-esteem 230 Self-esteem was measured using the 10-item Rosenberg Self-Esteem Scale (RSES; 231 Rosenberg et al., 1995). Participants responded to statements (e.g., "I take a positive attitude 232 toward myself," "I feel that I'm a person of worth, at least on an equal plane with others") on a 233 4-point Likert scale from strongly disagree (1) to strongly agree (4). Total scores were created by 234 summing each participant's responses ( = .88). 235

Personal Hardiness 236
Personal hardiness was assessed using the Courage to Challenge Scale (Smith & Gray, 237 2009). The Courage to Challenge Scale was developed for use with sexual and/or gender 238 minority populations to examine resilience, coping, and self-efficacy. Respondents answered 18 239 items (e.g., "Getting through tough times prepares me for future challenges," "Dealing with 240 difficult situations has helped me grow in positive ways") on a 7-point Likert scale ranging from 241 strongly disagree (1) to strongly agree (7). Responses to each item were averaged to create a total 242 score ( = .85). 243

Comfort from Companion Animals 244
Emotional comfort from companion animals was assessed using the Comfort from 245 Companion Animals Scale (CCAS; Zasloff, 1996). Respondents rated their agreement with 11 246 statements (e.g., "My pet is a source of constancy in my life," "My pet makes me feel needed") 247 on a 4-point Likert scale from strongly agree (1) to strongly disagree (4). A total score was 248 computed by summing the items ( = .92). 249

Victimization 250
Victimization related to one's gender identity and/or expression was measured using the

Analysis Plan 280
All analyses for the study were conducted using IBM SPSS Statistics (Version 26) and 281 PROCESS (Hayes, 2017). We conducted four simple moderation analyses (see Figure 1A) to 282 examine whether, and to what extent, the association between comfort from companion animals 283 and each psychological adjustment variable (personal hardiness, self-esteem) varied as a function 284 of LGBTQ-related interpersonal stressors (i.e., victimization and interpersonal 285 microaggressions). Additionally, we conducted additive multiple moderation models that 286 included both victimization and interpersonal microaggressions as moderators of the relation 287 between comfort from companion animals and each dependent variable (see Figure 1B). 288 Covariates were included in each model: age (continuous), race/ethnicity (White/non-Latinx = 1, 289 minority race/ethnicity or multiple racial/ethnic identities = 0), gender modality (gender minority 290 = 1, cisgender = 0), extent to which current needs are met (continuous), whether the participant 291 reported being the primary caretaker of a dog and/or cat in the past 12 months (=1) or not (=0), 292 social support (continuous), and whether participation occurred prior to (=0) or after (=1) 293 interviews began being conducted online due to COVID-19. Both race/ethnicity and gender 294 identity were dichotomized due to insufficient power to analyze the differences between each 295 identity category. 296 We tested for the multivariate assumptions of normality, linearity, multicollinearity, 297 singularity, and homoscedasticity, which were all met. Mahallanobis distance scores were 298 computed and indicated that there were no outliers. We also standardized all continuous 299 variables included in the analyses (Baron & Kenny, 1986). We conducted a post-hoc power 300 analysis using G*Power software (Faul et al., 2009); the results indicated that our sample size (n 301 = 155) was sufficient (> .80) to detect a hypothesized incremental medium (f 2 = .15) or large (f 2 302 = .35) effect size (Cohen, 1977) at an alpha level of .05 and a critical F value of 3.91. However, 303 we had less than adequate power to detect a small effect size (f 2 = .02). 304

Results 305
Means, standard deviations, and correlations among constructs are reported in Table 2. 306 The correlation between personal hardiness and self-esteem was statistically significant (r = .56, 307 p < .001). Personal hardiness was positively and significantly associated with comfort from 308 companion animals (r = .29, p < .001), while self-esteem was not significantly associated with 309 comfort from companion animals. Comfort from companion animals was not significantly 310 associated with victimization but was positively and significantly associated with interpersonal 311 microaggressions (r = .33, p < .001). However, this effect was not strong enough to violate the 312 assumption of multicollinearity, as VIF and Tolerance were all acceptable (Hair et al., 2010). 313 The only covariates significantly associated (p < .05) with our dependent variables in our 314 moderation models were gender modality (i.e., identifying as a gender minority) and social 315 support. Identifying as a gender minority was negatively associated with personal hardiness and 316 self-esteem in all moderation models; social support scores were positively associated with 317 personal hardiness and self-esteem across all models. All covariates examined were included in 318 the final models, despite non-significance, as they did not affect power. 319

Simple Moderation Analyses 320
In our simple moderation model that examined victimization as a moderator of the 321 relation between comfort from companion animals and personal hardiness, comfort from we hypothesized that comfort from companion animals would be positively related to self-393 esteem and personal hardiness. Further, we hypothesized that the positive association between 394 comfort from companion animals and these indicators of psychological adjustment would be 395 strongest for those who report greater exposure to microaggressions and victimization. 396 Our first hypothesis was partially supported. We found a significant association between 397 comfort from companion animals and personal hardiness across the simple and multiple 398 moderation models, in which higher levels of comfort from companion animals were related to 399 greater personal hardiness. These results coincide with existing literature that finds HAI is 400 attachment (e.g., love) were frequently linked to the promotion of self-esteem. In contrast, 412 activities associated with emotional comfort (e.g., tactile interactions) were more likely to be 413 discussed in relation to arousal regulation (i.e., emotion regulation, stress reduction), which may 414 help to explain the differences in significant associations between comfort from companion 415 animals and self-esteem and personal hardiness. Our study highlights the importance of 416 investigating different characteristics of HAI in relation to multiple aspects of human health and 417 wellbeing, in order to adequately delineate the mechanism through which HAI may confer 418 benefits and risks to those who live with pets. 419 In support of our second hypothesis, we found that exposure to interpersonal 420 microaggressions was a significant moderator of the relation between comfort from companion 421 animals and personal hardiness. The results of the conditional effects suggest that the magnitude 422 of the effect of comfort from companion animals on personal hardiness may be strongest for 423 those who report moderate and high levels of interpersonal microaggressions. However, at low 424 levels of interpersonal microaggressions, the relationship between CCAS and personal hardiness 425 was no longer significant. Our results support the hypothesis that individuals who experience 426 adverse social contexts characterized by high levels of interpersonal microaggressions may 427 receive more benefits from HAI. This may be due to the link between LGBTQ-related 428 interpersonal stressors and inadequate community-level support and/or lower levels of perceived 429 social support from friends, family, and significant others (Dakin et al., 2020;Ehlke et al., 2020). 430 In the absence of human support, LGBTQ+ individuals exposed to interpersonal 431 microaggressions may rely on their pets more as a source of emotional comfort, which may 432 strengthen the positive relationship between comfort from companion animals and individual-433 level coping skills and resilience. Indeed, our results also indicated that comfort from companion 434 animals was positively correlated with interpersonal microaggressions. This could also explain 435 why there was not a significant relation between comfort from companion animals and personal 436 hardiness for participants who reported low levels of interpersonal microaggressions. 437 there is limited empirical evidence that supports why victimization may interact with HAI in 469 relation to self-esteem in this way. Results from qualitative studies suggest that caring for a pet 470 may lead to enhanced feelings of worth and responsibility in LGBTQ+ youth (Bryant, 1990; receiving comfort from one's pet) may be an important facilitator of self-esteem for participants 473 exposed to greater levels of victimization, as it is evidence of their abilities to care for another 474 being. However, these results should be interpreted with caution, and replication with a larger 475 sample size adequate to detect small effects is needed. 476 Limitations 477 There are a few notable limitations of the current study. As a result of our community-478 engaged research approach, our data were collected using convenience sampling methods. 479 Additionally, our results are based on cross-sectional data. Although our sample size provided 480 adequate power to detect moderate and large effect sizes, we lacked sufficient power to detect 481 small effect sizes. Due to these power limitations, we relied on dichotomized race/ethnicity and 482 gender modality variables that likely did not capture the full extent of participants' diverse 483 experiences. Similarly, due to the limitations of our sample size, we were unable to test whether 484 there were differences in the relation between HAI and our outcomes based on species type 485 (dogs vs. cats). A limitation related to our measurement approach is that the GMSR assesses 486 lifetime victimization experiences due to gender identity and expression and not sexual 487 orientation, whereas the LGBQ Microaggressions on Campus Scale asks respondents about 488 experiences of interpersonal microaggressions related to their sexual orientation. Although 489 experiences of sexual orientation and gender-based discrimination frequently co-occur across 490 LGBTQ+ identities, there may be differences in how participants responded to items based on 491 whether the participant identified as a sexual and/or gender minority. Another limitation is our 492 use of measures that assessed experiences based on different timeframes (i.e., past year, 493 lifetime). For example, we assessed lifetime experiences of victimization; in contrast, interpersonal microaggressions were measured by assessing experiences over the past year. 495 Further, we did not collect data on the source of microaggressions or victimization (e.g., friend, 496 family, peer). Due to the hypothesis that HAI may be especially beneficial for those who lack 497 human social support, capturing whether adverse interpersonal experiences are caused by family, 498 peers, and/or close others versus strangers is important to further delineate under what conditions 499 and for whom HAI may provide the most benefits. 500

Implications 501
Given our findings suggesting that the magnitude of the effect of comfort from 502 companion animals on positive psychological adjustment (i.e., self-esteem, personal hardiness) is 503 greatest for those who experience high levels of LGBTQ-related interpersonal stressors, we 504 continue here with implications for policy and practice. Considering that LGBTQ+ individuals, 505 and individuals with other marginalized identities, are disproportionately at risk for issues related 506 to economic vulnerabilities, such as (lack of) access to pet-friendly rental housing and veterinary 507 care, we emphasize a need for communities to encourage partnerships between social service 508 providers and animal welfare organizations. These community partnerships should focus on 509 supporting people and pets through collaborative measures that reduce barriers to health and 510 wellbeing for LGBTQ+ emerging adults and their pets. For example, providing free or reduced-511 cost basic veterinary care or pet supplies in the same location where individuals receive mental 512 health services could reduce some of the burden of issues with access to transportation or the 513 need to take time away from paid work for multiple appointments. Additionally, social service 514 and mental health providers should be cognizant of the animal services that may be available in 515 their communities in order to assist individuals with pet-related needs. These types of efforts 516 within communities, and more generally framed within public policy, could make great strides in 517 supporting the health and wellbeing of individuals who take comfort in their pets while 518 simultaneously supporting the welfare of their pets. 519

Future Directions 520
Our findings emphasize the benefits of companion animals for LGBTQ+ communities 521 and the importance of considering the nature and severity of stressors and how adversity 522 exposure may impact HAI. Future research should replicate and expand on our cross-sectional 523 study. For example, the hypothesis that the benefits of HAI will be most pronounced among 524 those who experience adversity and lack social support should be explored with additional types 525 of LGBTQ+ minority stress, such as discrimination and rejection. Future research should also 526 capture multiple forms of minority stress (e.g., racism, ableism) that can intersect and co-occur 527 with LGBTQ+ minority stress and complicate experiences of adversity and related social and