We conducted an analysis of nonprobability cross-sectional data collected from a sample of SGM students (N = 565) to explore the effect of the early stages of the pandemic (spring and summer 2020) on SGM college students. Eligibility criteria included being at least 18 years of age, identifying as an SGM person, and being a college student. Data were collected between May 27^th and August 14^th in 2020 via online questionnaire. All measures were self-reported. Participants were incentivized with a raffle for a $50 Amazon gift card. The University of Maryland Institutional Review Board granted approval for the study.

Participants were recruited using an electronic recruitment flyer with a link to an online Qualtrics survey. The recruitment flyer was distributed via multiple social media platforms (i.e., Facebook, LinkedIn, and Twitter). We also recruited through email campaigns within our internal and external professional networks, and at historically Black colleges and universities, Hispanic-serving institutions, and lesbian, gay, bisexual, transgender, and queer or questioning (LGBTQ) student centers across the U.S. The research team distributed over 2,000 emails during study recruitment (not including forwarded emails). Available data from the survey URL link management platform reveals that the survey link was opened (clicked) over 1,400 times; 71% of clicks were from emails, and 23% were from social media platforms (with 60% from Twitter and 29% from Facebook).

Upon clicking the link to the survey, participants immediately underwent a self-administered electronic informed consent process. Those who opted to participate provided implied informed consent by clicking an arrow at the end of the informed consent process to proceed and begin the survey. To take the full survey, participants had to meet the study eligibility criteria, which was assessed at the beginning of the survey with four questions to confirm that participants were full-time graduate or undergraduate students in the U.S., identified as an SGM person, and were at least 18 years old.

Survey duration was approximately 20–25 minutes, with a total of 87 questions (including single and multi-item questions). Study participants had the right to skip or not answer any question in the survey. Upon completion of the survey, participants were provided with a listing of mental health and crisis management resources and contact information for the study principal investigator in case they had any questions.

Descriptive statistics were used to examine the distribution of demographic and educational outcomes. Frequencies and percentages were used for categorical variables. For continuous variables, distributions were examined using histograms and values for skewness and kurtosis. Significantly skewed variables were dichotomized. For normally distributed variables, means and standard deviations were reported.

To examine the role of living arrangement during the early stages of the COVID-19 pandemic, a series of Chi-square analyses were used. Specifically, differences based on living arrangement were examined for all academic performance outcomes, LGBTQ-related stressors, access to care, and alcohol and other substance use. A One-Way ANOVA was used to determine the extent to which psychological distress differed based on living arrangement. Finally, to determine the extent to which findings related to living arrangements and outcomes differed between undergraduate and graduate students, a stratified analysis was completed for these two populations separately using the same statistical tests described above.

Responses were obtained from 565 college students enrolled in schools across 37 different states. Table 1 shows a summary of the demographic characteristics for the total sample and for the three living arrangements during the early stages of the COVID-19 pandemic. Most of the students were enrolled in undergraduate programs (70%), with 2.5% earning their Associates Degree and 64% earning a Bachelor’s degree. While most identified as White (75%), 14% of the sample were Hispanic or Latina/o/x, 7% another race or ethnicity, 14% Asian, and 11% Black or African American. Additionally, 54% of respondents identified as cisgender women, 15% as cisgender men, 10% as transgender, 12% as non-binary, and 6% as genderqueer. With regards to sexual orientation, 32% of the sample identified as bisexual, 14% as gay, 18% as lesbian, 16% as queer, 7% as asexual, and 9% as pansexual. When further examining the group of students who were living on campus and had to move, it should be noted that the overwhelming majority (<90%) of these students returned to their parents’ homes.

When asked about the impact of the pandemic on their grades, about a quarter indicated that their grades suffered (24.7%), with 5% of students failing classes. Regarding academic progression, some students reported having to drop a class (10%) or drop out of school (1.4%).

A series of Chi-square tests of association were used to determine whether living arrangements during the early stages of the COVID-19 pandemic were associated with academic performance outcomes and LGBTQ-related stressors.

When examining academic performance outcomes, no significant differences based on living arrangements were found related to grades suffering (χ² (N = 479) = 2.97, p = .227, Cramer’s V = .079), failing classes (χ² (N = 485) = 1.72, p = .423, Cramer’s V = .060), having to drop a class (χ² (N = 490) = 1.04, p = .595, Cramer’s V = .046), or having to drop out of school (χ² (N = 496) = 1.29, p = .524, Cramer’s V = .051).

On LGBTQ-related stressors, significant differences based on living arrangements were found related to LGBTQ-related family rejection (χ² (N = 509) = 10.49, p = .005, Cramer’s V = .144), internalized transphobia (χ² (N = 153) = 8.26, p = .016, Cramer’s V = .232), and LGBTQ-related identity concealment (χ² (N = 501) = 7.19, p = .027, Cramer’s V = .120), with students that remained living on campus being less likely to experience these stressors compared to their peers with other living arrangements. No significant differences were found based on living arrangements and LGBTQ victimization (χ² (N = 509) = 0.54, p = .765, Cramer’s V = .032), and internalized homophobia (χ² (N = 509) = 0.59, p = .748, Cramer’s V = .034).

When examining access to care, students who never lived on campus were more likely to report difficulty accessing mental health care (χ² (N = 509) = 6.73, p = .035, Cramer’s V = .115).

No significant differences were found based on living arrangements and psychological distress (F(2, 506) = 0.37, p = .691, η² = .001). When examining differences on increased alcohol and substance use since the start of the COVID-19 pandemic, students who never lived on campus were more likely to report increased alcohol use (χ² (N = 509) = 12.63, p = .002, Cramer’s V = .158) and increased tobacco use (χ² (N = 509) = 7.78, p = .020, Cramer’s V = .124). Additionally, students that never lived on campus or who stayed on campus were more likely to report an increase in recreational cannabis use (χ² (N = 509) = 12.48, p = .002, Cramer’s V = .157).

Table 2 provides a summary of results disaggregated by type of program. There were significant differences in the proportion of students who never lived on campus, lived on campus but had to move back home, or remained living on campus after the start of the pandemic between undergrad and graduate students (χ² (N = 509) = 88.79, p < .001, Cramer’s V = .418). To examine the extent to which results for undergraduate and graduate students yielded consistent findings, the analyses were stratified by educational program.

For undergraduate students, no academic performance outcomes were found to differ by living arrangements. For graduate students, academic performance outcomes that were found to differ by living arrangements included grades suffering (χ² (N = 131) = 6.65, p = .036, Cramer’s V = .225) and failing classes (χ² (N = 135) = 9.39, p = .009, Cramer’s V = .264). Graduate students who were living on campus and had to move were more likely to report that their grades suffered compared to those who never lived on campus. While only graduate students reported failing classes during the Spring 2020 semester, these students either lived on campus and had to move or remained on campus.

Among undergraduate students, significant differences based on living arrangements were found related to racism (χ² (N = 154) = 6.74, p = .034, Cramer’s V = .209). Among graduate students, no significant differences based on living arrangements were found related to any LGBTQ stressor.

There were no differences between undergraduate and graduate students in-terms of their reported access to mental health care.

Lastly, when examining differences on increased alcohol and substance use among undergraduate students, significant differences based on living arrangements were found related to recreational marijuana (χ² (N = 358) = 9.45, p = .009, Cramer’s V = .162), alcohol (χ² (N = 358) = 6.52, p = .038, Cramer’s V = .135), and tobacco (χ² (N = 358) = 6.59, p = .037, Cramer’s V = .136), with undergraduate students who remained on campus and those who never lived on campus reporting higher rates of increased recreational marijuana, alcohol, and tobacco compared to undergraduate students who lived on campus and had to move. Among graduate students, significant differences based on living arrangements were found related to increased alcohol use (χ² (N = 151) = 6.81, p = .033, Cramer’s V = .212), with graduate students who never lived on campus reporting the highest rates of increased alcohol use.

The findings of this study should be interpreted in light of several limitations, including the electronic nature of the study. The research team used data collection methods that yielded an anonymous survey from which data could not be validated, and the possibility of duplicate responses could not be excluded (though protections against duplicate survey entries were implemented), particularly given that the survey was incentivized. Selection bias is a consideration based on the survey distribution methods. The fact that survey completion required internet access presents another possible selection bias, but not a probable one given that the college students targeted by this study are likely to have regular internet access. Another limitation is that data were self-reported and retrospective, possibly resulting in recall bias or social desirability bias. We did not collect institution name or city-level location (to protect participant anonymity) or the level of available resources at each institution, which limited our ability to analyze regional distribution of the sample and the available resources at each institution based on whether students remained on campus or not. Sexual orientation and gender identity questions allowed for single answers only, possibly impacting young peoples’ ability to reveal their full identities if they identified with multiple sexual orientations and/or gender identities. We implemented the gold standard 3-step question process (Cahill et al., 2014; Cahill et al., 2016; Deutsch et al., 2013), which minimizes the probability of this occurrence. Also related to sexual orientation and gender identity, we did not disaggregate the analyses by sexual orientation and gender identity because we were statistically underpowered to do so. In addition, some analyses of subgroups may have been underpowered resulting in Type II error and, therefore, an inability to detect significant effects.

Future studies should examine the role of sexual orientation and gender identity. We acknowledge that the experiences of these subgroups are not identical and, as a result, the outcomes are not all the same. Lastly, as a result of the cross-sectional nature of the study, temporality and therefore causality cannot be established. For example, SGM students’ academic performance, SGM stress experiences, and mental health now likely differ (and will in the future) compared to summer 2020. Despite this limitation, the study still provides important findings and implications relevant for student affairs practitioners and the academic performance and mental wellbeing of SGM students. Longitudinal studies will be able to address this study limitation to examine the impact of the COVID-19 pandemic over time.

Our findings have important implications for higher education institutions and public health in the context of SGM college students in the time of COVID-19 and beyond. It is imperative for higher education institutions and public health systems to recognize and address SGM young people and college students as vulnerable health disparity populations with unique concerns and needs. Given the ongoing nature of the COVID-19 pandemic (Johns Hopkins University, 2021), higher education institutions in the U.S. must address the changing needs of SGM students since the start of COVID-19. Thus, higher education institutions should take steps to prevent and address highly relevant SGM-stressors that could be impacting the mental health, well-being, and academic success of SGM college students, especially given the endemic and long-lasting effects of COVID-19. Recommendations include implementation and expansion of access to LGBTQ affirming mental health care (Williams & Fish, 2020) on and off-campus, including provision of remote or online services and supports (Fish et al., 2020b), as well as dissemination of online or remote resources for SGM college students and their parents and families that are useful for the prevention of SGM stressors (e.g., family rejection, internalized transphobia, identity concealment) (Cohen, Mannarino, Wilson, & Zinny, 2018; Diamond & Shpigel, 2014; Ryan, 2009; SAMHSA, 2014).

Potential interventions include offering more remote and in-person support groups specifically for SGM students (including specifically for SGM students with multiple marginalized identities [Latinx, Black, immigrant, etc.]) that address mental health, identity-related stress, and academic-related concerns (Duran, 2020; Algarin, Salerno, Shrader, Lee, & Fish, 2022). Higher education institutions may also offer (online or remote) trainings that provide education and skills building for students, professors, and administration toward supporting students who are struggling with their mental health and academic performance in the context of COVID-19 (Gato et al., 2021). Student Affairs practitioners can become strong advocates for SGM student success and wellbeing by being leaders in offering on-line office hours and creating various programs and initiatives. Another recommendation would be to establish a mentorship program for students to meet with positive role models that can be facilitated either in person or virtual (Csaposs, B., 2022; Hogan el al., 2020; Sarna et al., 2021). Student Affairs practitioners can also work collaboratively with their campuses academic support programs and health centers to establish much needed resources for LGBTQ students (Conron, 2021). Lastly it would be recommended that campuses counseling centers be sought out for additional support and to identify possible liaisons to work with to best meet the mental health needs of the SGM student populations. University stakeholders are advised to work with campus LGBTQ student centers, offices of diversity and inclusion, health and mental health stakeholders, and expert faculty on such initiatives.

This study has shared important insights into the academic, minority stress, and mental health experiences of SGM college students during the early months of the COVID-19 pandemic. Our findings have highlighted the importance of providing LGBTQ affirming mental health care to SGM students on college and university campuses. The study reinforces the ways in which college and university campuses continue to serve as safe spaces for marginalized SGM populations. Future scholarship in this area is needed to provide further insights into the impact of the COVID pandemic on SGM college students in relation their academic, mental health, minority stress, and behavioral outcomes. This study can serve as an important starting point for mental health care providers and college practitioners as they find ways to meet the needs of SGM college students.