Self-efficacy, social distancing, and essential worker status dynamics among SGM people

Abstract

As of October 2020, the COVID-19 pandemic has accounted for over 210,000 deaths in the U.S. Sexual and gender minority populations are more likely to work in essential industries while bearing a disproportionate burden of the virus. Constructs consistent with Protection Motivation Theory (perceived severity, vulnerability, self-efficacy, and response efficacy) were measured using an abridged version of Kleczkowski et al.’s 4-factor Protection Motivation Theory Psychological Measures to examine social distancing behaviors of these populations. 32.6% of the sample were essential workers. Greater self-efficacy predicted stricter social distancing behaviors. Non-essential and unemployed worker statuses were associated with increased odds of stricter social distancing behaviors relative to essential worker status. Essential worker status predicted lower self-efficacy. The indirect effect of essential worker status on social distancing through self-efficacy was significant. Findings suggest that interventions that encourage social distancing through enhanced self-efficacy may optimize health for sexual and gender minority essential workers.

Introduction

Background

Since December 2019, the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), also known as Coronavirus Disease 19 (COVID-19), pandemic has affected nearly every corner of U.S. society.¹ While pharmaceutical trials for interventions to treat and prevent infection are underway, social distancing—a practice of maintaining six feet of distance between persons—remains one of the few options to slow COVID-19 infection rates and reduce the risk of overburdening local healthcare systems.^2–8 In spite of social distancing measures, emerging research on the effect of the COVID-19 pandemic on sexual and gender minority (SGM) individuals broadly has found that such populations bear a higher burden of the COVID-19 pandemic than others due to a number of factors (e.g., a higher likelihood of living below the poverty line, inadequate health coverage, and higher rates of smoking tobacco products).^9–13 Further, SGM populations have a greater propensity to occupy jobs in essential industries than their non-LGBTQ+ counterparts, increasing their likelihood of COVID-19 exposure.¹¹

Study Aims

Considering SGM populations’ elevated COVID-19 risk, research identifying the psychological mechanisms that inform the social distancing practices of these populations is greatly needed. One such avenue for investigating these factors is via the application of a Protection motivation theory (PMT) model.¹⁴ PMT is a psychosocial framework that has been widely used in research on health outcomes for a myriad of populations.^14–17 The objective of the current study was to test a hypothesized path model implied by PMT in a sample of SGM respondents. PMT constructs serve as proximal determinants of social distancing behaviors. Subsequently, employment in an essential industry would serve as a distal predictor of social distancing behaviors with the potential of operating through PMT constructs.

Employment Considerations for SGM Populations

Notwithstanding the devastating economic effects of the COVID-19 pandemic, SGM populations face unique employment challenges due to their marginalized status.^18,19 Although recent Supreme Court precedents have expanded workplace discrimination protections to include SGM populations in the U.S., such minority groups have long faced job insecurity as well as harassment at the workplace due to their gender or sexual identity expression.^20–24 In fact, legal proceedings that would extend protections to SGM populations are indicative of the pervasiveness of minority stress in the daily lives of SGM workers nationwide.²⁵

Furthermore, burgeoning research on the financial toll of the COVID-19 pandemic on SGM populations also indicates that such minority groups are positioned to bear the brunt of the current economic recession.²⁶ In one study, sexual minority men reported disproportionate rates of unemployment due to the COVID-19 pandemic as a function of their employment as hourly workers in service industries.²⁶ Such findings may be emblematic of SGM populations’ experience in the workforce broadly, one of being devalued due to either their sexual or gender identity, or both.²⁷ Yet in spite of job loss due to the pandemic and the systemic depreciation of SGM workers in the U.S., employment in the service industry (amongst others) may also be characterized as “essential” in context of the COVID-19 pandemic.²⁸

The definition of essential work in the U.S. differs state-by-state and as a result, the array of industries considered essential consists of healthcare, food services, manufacturing, and a diverse selection of others.²⁸ Individuals who occupy these positions are also categorically at higher risk for COVID-19 infection due to increased rates of disease exposure and variability in the mandated use of personal protective equipment (PPE) while at the workplace.²⁸ In respect to LGBTQ+ Americans specifically, 5 million SGM individuals occupy such positions at present.¹¹

Additionally, recent research has emphasized the uniquely at-risk position of Black and Latinx SGM populations.²⁹ Studies have identified that structural racism and discrimination are associated with inadequate healthcare coverage, health care access, and increased risk of COVID-19 exposure for Black and Latinx populations in the U.S.³⁰ The convergence of these multiple marginalized identities places LGBTQ+ communities of color at unique risk of enduring adverse health-related and psychosocial effects due to the virus.

COVID-19 and Social Distancing in the United States

As of October 2020, the pandemic has accounted for over 35 million infections worldwide with more than 210,000 deaths in the United States alone.³¹ Based upon evidence from previous pandemics and outbreaks (e.g., H1N1 influenza, SARS-CoV-1) the Centers for Disease Control and Prevention (CDC) has indicated that social distancing, in conjunction with other behavioral measures, is effective at preventing infection.^4,7,8 As a result, this collection of practices (e.g., maintenance of six feet of physical distance to others, face covering, isolation of those with exposures, symptoms, etc.) has served as a common fixture in COVID-19 prevention strategies.^4,7

Yet prior research on the implementation of social distancing in the U.S. has found multiple barriers to optimal implementation (e.g., concerns over deleterious effects on local and federal economies, as well as the possible infringement upon personal liberties).³² Emerging research has also indicated that essential workers, who are at high risk of COVID-19 infection, have reported suboptimal social distancing practices when compared to those with non-essential employment.³³ Further, essential workers in the aforementioned study also anticipated being able to self-isolate for a shorter period of time than others, which may indicate the financial strain imposed by such social distancing practices.³³ However, research on social distancing practices among SGM populations in the U.S., notwithstanding SGM essential workers, remains limited. As SGM essential workers are at unique risk of COVID-19 infection and especially vulnerable due to their marginalized socioeconomic status, research is needed to inform policies that would serve to protect such populations who might otherwise be omitted from aggregate research.³⁴

Protection Motivation Theory and Social Distancing

PMT is framework composed of constructs that seek to capture the mechanisms behind the motivation to adopt certain protective behaviors—in this case, those related to one’s health.¹⁷ According to PMT, protective behaviors are motivated by one’s perception of a given threat’s severity (health-related or otherwise), how likely an individual believes they would be adversely affected by the threat (i.e., vulnerability), how effective one believes a certain protective behavior is in response to a threat (i.e., response efficacy), and how effective one believes their execution of this protective behavior will be (i.e., self-efficacy).¹⁷

Kleczkowski et al.’s study is one of the few to examine how PMT variables predict the intention and implementation of social distancing behaviors during a simulated epidemic.^35,36 Their findings suggest that while the endorsement of PMT variables could predict the intent of adopting social distancing practices, no aspect of PMT could predict enactment of social distancing during the simulation itself.³⁶ The COVID-19 pandemic has provided the context to expand upon this research and in so doing provide a real-world application of the hypothesis put forth by Kleczkowski et al. (i.e., that PMT constructs would predict social distancing behaviors) though in respect to SGM populations in the U.S.³⁶

As previously stated, the psychosocial mechanisms which inform the social distancing practices of SGM essential workers warrant particular attention due to the increased COVID-19 infection risk and socioeconomic vulnerability of these populations during the ongoing crisis.¹⁹ The application of the PMT framework is particularly relevant to the investigation of SGM populations in this context for a number of reasons. Firstly, research has identified numerous social factors that contribute to SGM populations’ vulnerability broadly (e.g., social stigma, discrimination, etc.).^37,38 Secondly, studies have also called attention to how these factors also undermine the self-efficacy of SGM individuals to remediate such vulnerability.^39,40 Thus, PMT provides a lens to examine SGM populations in the U.S. that is consistent with the extant literature on the lived experience of SGM individuals in the U.S. today.

As previously stated, the purpose of the current study is to test a hypothesized path model indicated by the PMT framework in a sample of SGM respondents. We hypothesized that PMT constructs would serve as proximal determinants of social distancing—a practice that may be considered a novel protective health behavior in the U.S. As such, we also hypothesized that employment in an essential industry would assume the role of a distal predictor of social distancing behaviors with the possibility of operating through PMT constructs.

Methods

Recruitment

Data was collected from a larger study on examining social and behavioral factors associated with COVID-19 among SGM populations in the U.S. Between May 6th and May 15th 2020.¹⁰ Participants were recruited using targeted advertisements on geo-location-based dating apps that target SGM populations. As described in Starks et al.¹⁰, advertisements included an image of one or more adolescent or young adult men. The advertisements included the study name “PRIDE Endures” and one tagline (i.e., “Sex in the time of social distancing” or “COVID-19 and the LGBTQ + community”). All materials were available in English. To be eligible for the current analysis, participants needed to be older than 18 and identify as a sexual and/or gender minority and have not or believed to have not contracted COVID-19. Participants were shown a study advertisement and then clicked on the advertisement to get directed to a landing page with a summary of the study. Interested participants would be consented and were routed to the Qualtrics-based survey. Upon completion, participants had the option of providing their email address for an opportunity to win 1 of 5 Amazon gift cards worth $50. All procedures were approved by the Institutional Review Board at Hunter College, City University of New York.

Measurement

Demographics.

Participants reported their age, education, race and ethnicity, and income.

Sexual and gender identity.

Regarding sexual gender identity, participants were asked what they consider themselves to be. They were able to select from a set of options which included gay, lesbian, bisexual, queer, straight/heterosexual, or other. Gender identity was assessed using “Which best describes your gender identity?” The response format included a diverse set of options: cis man, cis women, transwoman, transman, nonbinary, and other. Categories were collapse due to the lack of responses in a particular category.

Employment.

Participants were also asked about their employment status and classification. Specifically, participants were asked if they are employed or unemployed. For those who indicted being employed, they were then asked if their employment was considered to be essential or non-essential with the definition: “you have been told you must continue to go to work because your job is part of essential service.” A three-category employment variable was created i.e., Essential Worker, Non-Essential Worker, and Unemployed.

Subsequently, essential worker participants were asked to identify the industry in which they are carrying out their essential service. Possible responses included; health and social services, food services, and transportation among others.

Constructs consistent with PMT were measured using an abridged version of a 4-factor self-report measure which asked them to state the extent to which they agreed or disagreed with several statements designed to measure their beliefs on social distancing in the context of the COVID-19 pandemic.^35,36

Perceived severity.

Perceived severity was measured using the 2 items: “If I were to develop COVID-19, I would suffer a lot of unpleasant symptoms” and “Developing an infectious disease would be unlikely to cause me to die prematurely.” The response format was a 5-point likert-type scale (1 = Strongly Disagree and 5 = Strongly Agree). One of the items was reverse-coded so that greater scores indicate greater perceived severity. The two items showed good reliability (α= 0.79).

Perceived vulnerability.

Perceived vulnerability was measured using the 2 items:” My chances of developing COVID-19 in the future are likely” and “I am unlikely to develop COVID-19 in the future.” The response format was a 5-point likert-type scale (1 = Strongly Disagree and 5 = Strongly Agree). One of the items was reverse-coded so that greater scores indicate greater perceived vulnerability. The two items showed good reliability (α= 0.89).

Response efficacy.

Response efficacy was measured using a single item: “If I were to engage in social distancing (e.g. by avoiding public transport and social events) I would lessen my chance of developing COVID-19.” The response format was a 5-point likert-type scale (1 = Strongly Disagree and 5 = Strongly Agree).

Social Distancing.

Social distancing was measured using the ordinal item: “To what extent are you social distancing?” Possible responses included: (1) All of the time. I am staying at home nearly all the time,” (2) “Most of the time. I only leave my home to buy food and other essentials, (3) Some of the time. I have reduced the amount of times I am in public spaces, social gatherings, or at work, (4) Some of the time. I am limiting social interaction to family members who live in my community, and (5) None of the time. I am doing everything I normally do. The item was reverse-coded so that greater scores indicate stricter social distancing.

Analytic Plan

First, bivariate analysis tested associations among demographic variables, PMT constructs and the three-category employment status. Second, to further describe the sample, we conducted a frequency analysis on the variable that asked participants to indicate the industry which they carry out essential functions. Both of these analytic steps were conducted in SPSS (v. 25). Finally, the hypothesized path model was tested. The primary outcome, social distancing, was predicted by PMT constructs and employment status. Subsequently, PMT constructs were predicted by employment status. The model adjusted for demographic covariates (age, race/ethnicity, sexual identity, gender, education, income, and region of residence) in the prediction of all endogenous variables. The significance of indirect effects associated with employment status to social distancing behaviors through PMT constructs were evaluated using a model constraint approach in which the indirect pathways coefficients were constrained to be “0” and a model test (Wald χ²) examined if the constrained model was statistically significant than the model in which coefficients are freely estimated. To account for missing data, the ordinal regression model was conducted in Mplus (v8) using full information maximum likelihood estimation.

Results

Table 1 presents the results of the bivariate tests of association. More than half of the sample was a person of color (62%) and identified as gay or lesbian (71.3%). The vast majority of the sample indicated being cis men (93.8%). The sample was geographically diverse with participants from the South comprising of 41.5% of the sample. The Northeast comprised of 21.5% and the West comprised 19.0% of the sample. About 17% of the participants lived in the Midwest. Participants who were unemployed were less likely to have a 4-year degree and less likely to earn more than $30,000 annually. Essential workers scored significantly lower on self-efficacy scores compared to both non-essential employees and participants who were unemployed.

Table 1.

Demographic characteristics

	Overall		Essential Worker	Non-Essential Worker	Unemployed	Test Statistic	p-value
Subgroup n (%)	917 (100.0)**		299 (32.6)**	184 (20.0)**	375 (40.9)**
	n(%)		n(%)	n(%)	n(%)
Race/Ethnicity						χ2(6) = 6.18	.402
African American	368 (40.1)		129 (43.1)	75 (40.7)	132 (32.5)
White	349 (38.0)		112 (37.4)	71 (38.5)	151 (40.2)
Latino	118 (12.9)		31 (10.3)	23 (12.5)	56 (14.9)
Multiracial or other	82 (8.9)		27 (9.0)	15 (8.1)	36 (9.6)
Sexual Identity						χ2(6) = 7.54	.274
Gay and Lesbian	654 (71.3)		221 (73.9)	135 (73.3)	257 (93.7)
Bisexual	192 (20.9)		8 (2.6)	8 (4.3)	18 (6.3)
Queer	36 (3.9)		8 (2.6)	4 (2.1)	20 (5.3)
Other	35 (3.8)		8 (2.6)	4 (2.1)	20 (5.3)
Gender Identity						χ2(4) = 5.30	.257
Cis male	860 (93.8)		283 (94.6)	178 (96.7)	345 (92.0)
TG/GNC	65 (5.2)		13 (4.3)	5 (2.7)	25 (6.6)
Other	9 (1.0)		3 (1.0)	1 (0.5)	5 (1.3)
Education						χ2(2) = 32.91	< .001*
Less than a 4 year degree	506 (55.2)		159b (53.1)	68c (36.9)	235a (62.6)
4 year degree or more	411 (44.8)		140b (46.8)	116c (63.0)	140a (37.3)
Income						χ2(2) = 110.18	< .001*
$30,000 or more	484 (52.8)		203b (67.8)	143c (77.7)	137a (36.5)
<$30,000	377 (41.1)		96b (32.1)	41c (22.2)	238a (63.4)
Region of Residence						χ2(6) = 7.38	.286
South	381 (41.5)		135 (45.1)	78 (42.8)	143 (38.5)
Northeast	197 (21.5)		54 (18.0)	45 (24.7)	85 (22.9)
West	174 (19.0)		53 (17.7)	34 (18.6)	77 (20.7)
Midwest	157 (17.1)		57 (19.0)	25 (13.7)	66 (17.7)
	M(SD)		M(SD)	M(SD)	M(SD)
Age	37.96 (13.19)		38.30 (14.68)	36.45 (12.10)	39.06 (14.68)	F(2, 855) = 2.423	.089
PMT Severity	3.75 (1.34)		3.69 (1.30)	3.72 (1.31)	3.81 (1.39)	F(2, 631) = 0.564	.569
PMT Vulnerability	3.24 (1.25)		3.32 (1.29)	3.30 (1.19)	3.15 (1.26)	F(2, 631) = 1.265	.283
PMT Response Efficacy	5.35 (1.66)		5.26 (1.61)	5.60 (1.53)	5.29 (1.75)	F(2, 631) =2.152	.117
PMT Self-Efficacy	4.25 (1.34)		4.03 (1.39)a,b	4.41 (1.29)a	4.34 (1.32)b	F(2, 631) = 1.798	.009*

^**Sample sizes do not equal 917 due to missing data

Table 2 presents descriptive data on essential workers’ industries of employment. Essential worker participants working in health and social services (29.1%) and food service workers (17.8%) comprised the largest segment of the essential worker sample. Additionally, pharmacy staff (1.6%) and those working in security (1.6%) industries comprised the smallest segment of the essential worker sample.

Table 2.

Essential Worker Industries (n = 245)

	Frequency*	Percent (%)
Health & Social Services	75	29.1
Food Services	46	17.8
U.S. Local & Federal Government	33	12.8
Logistics & Transportation	28	10.9
Banking & Finance	15	5.8
Retail, Sales & Hospitality	13	5.0
Energy & Utilities	8	3.1
Manufacturing	8	3.1
Construction & Maintenance	6	2.3
Education	5	1.9
Technology	5	1.9
Pharmacy	4	1.6
Security	4	1.6
Other	8	3.1

^*Subsample size does not equal 245 as participants were able to select more than one category

Table 3 presents the results of the final hypothesized path model. Among PMT constructs, only self-efficacy was significantly associated with social distancing. Higher self-efficacy was associated with increased odds of indicating higher ordinal levels of social distancing (AOR = 1.28; 95% CI 1.11, 1.44; p<.001). Regarding demographic factors directly associated with social distancing, the model revealed racial disparities with Black (AOR = 2.17; 95% CI 1.50, 3.12; p <0.001) and those who indicated a multiracial or different race/ethnicity (AOR = 2.31; 95% CI 1.63, 3.22; p <0.001) had greater odds of engaging in stricter social distancing behaviors relative to White participants.

Table 3.

Adjusted Multivariable PMT Model

	Self-Efficacy					Social Distancing
	B	CI	β	p		B	CI	AOR	CI	p
Age	0.00	−0.02, 0.01	0.04	0.294		0.01	0.00, 0.02	1.01	0.99, 1.02	0.112
Education	0.13	−0.09, 0.36	0.05	0.262		0.11	−0.19, 0.41	1.11	0.82, 1.51	0.493
Income	0.25	0.01, 0.49	0.09	0.042		0.04	−0.27, 0.36	1.05	0.76, 1.45	0.773
Race/Ethnicity (Ref:White)
Black	0.21	−0.05, 0.46	0.07	0.119		0.37	0.02, 0.71	1.45	1.02, 2.04	0.034
Latino	0.13	−0.23, 0.49	0.03	0.488		0.36	−0.05, 0.78	1.44	0.95, 2.18	0.085
Multiracial or other	0.25	−0.09, 0.59	0.05	0.160		0.53	0.03, 1.03	1.71	1.03, 2.80	0.037
Gender (Ref: Cismale)
TGNC	0.10	−0.37, 0.58	0.01	0.670		0.60	−0.18, 1.39	1.82	0.83, 4.01	0.136
Other	0.19	−0.80, 1.20	0.01	0.697		0.19	−0.85, 1.24	1.21	0.43, 3.45	0.713
Sexual Identity (Ref: Gay)
Bisexual	−0.01	−0.25, 0.24	−0.01	0.986		−0.34	−0.72, 0.03	0.70	0.48, 1.03	0.072
Queer	0.15	−0.35, 0.65	0.02	0.566		−0.29	−0.92, 0.33	0.74	0.39, 1.39	0.356
Other	−0.15	−0.77, 0.47	−0.02	0.641		0.05	−0.87, 0.97	1.05	0.41, 2.63	0.913
Region
Midwest	−0.18	−0.49, 0.13	−0.05	0.251		−0.19	−0.63, 0.23	0.83	0.53, 1.25	0.369
South	−0.18	−0.47, 0.10	−0.06	0.215		−0.01	−0.36, 0.33	0.98	0.69, 1.39	0.940
West	0.05	−0.24, 0.34	0.02	0.718		−0.18	−0.57, 0.19	0.83	0.56, 1.21	0.334
Employment (Essential)
Non-Essential	0.33	0.04, 0.61	0.10	0.026		0.77	0.41, 1.14	2.17	1.50, 3.12	<0.001
Unemployed	0.39	0.14, 0.64	0.14	0.002		0.84	0.49, 1.17	2.31	1.63, 3.22	<0.001
PMT
Response Efficacy	-	-		-		0.10	−0.02, 0.22	1.11	0.98, 1.24	0.117
Self-Efficacy	-	-		-		0.24	0.11, 0.37	1.28	1.11, 1.44	<0.001
Severity of Illness	-	-		-		0.07	−0.06, 0.20	1.07	0.94, 1.22	0.325
Vulnerability	-	-		-		0.08	−0.06, 0.23	1.08	0.94, 1.25	0.259
Thresholds
Social Distance = 1	-	-		-		−2.18	−2.71, −1.66	0.11	0.06, 0.16	<0.001
Social Distance = 2	-	-		-		−1.55	−2.04, −1.05	0.21	0.13, 0.34	<0.001
Social Distance = 3	-	-		-		0.07	−0.39, 0.54	1.07	0.67, 1.71	0.761
Social Distance = 4	-	-		-		2.27	1.76, 2.77	9.67	5.81, 15.95	<0.001

The pattern of significant direct effects observed in path model results implied that only indirect pathways involving PMT self-efficacy might be significant (see Figure 1). Results of model constraint tests indicated that employment group explains a significant amount of the variance in social distancing through associations with PMT self-efficacy. Both non-essential workers (Wald χ²(2) = 17.56, p <.001) and unemployed participants (Wald χ²(2) = 21.48, p <.001) reported greater self-efficacy and indirectly greater social distancing compared to essential workers.

Figure 1:

Model of the Indirect Effects Through Self-Efficacy

Discussion

These findings highlight the potential for the PMT model to provide a framework for the study of COVID-19 prevention among essential workers in particular. Though Kleczkowski et al.³⁶ previously studied the effect of PMT constructs on social distancing behaviors in simulated form, the current study is the first to apply this model in examining social distancing behaviors in an actual pandemic. Findings specifically indicated the salience of self-efficacy as a predictor of social distancing. In this case, self-efficacy for social distancing also had the potential to explain shared variance between essential worker status and social distancing. Essential workers were less likely to practice strict social distancing and their self-efficacy to engage in this behavior was likewise lower than those of other employment groups. This finding exposes an intra-psychological point of vulnerability that, in turn, renders essential workers at greater risk of COVID-19 infection.

As previously discussed, engagement in a certain protective health behavior, according to PMT, hinges upon the belief that (1) forgoing the behavior increases one’s vulnerability to a given threat, (2) that this behavior is effective in mitigating one’s risk, and (3) that one is capable of implementing this behavior effectively.⁴¹ PMT proposes that high scores in all constructs captured by the abovementioned statements (i.e., severity, vulnerability, response efficacy, and self-efficacy) inform high engagement in protective behaviors to mitigate a threat to one’s health.⁴¹ Although our findings regarding social distancing behaviors are not consistent with this model as a whole, research has shown that self-efficacy is the most consistent and robust predictor of behavioral intentions within the PMT model.^42–45 In this sense, the present finding that only self-efficacy was positively associated with social distancing is consistent with trends in prior PMT research.⁴²

One possible explanation for study findings is that more rigorous social distancing behaviors tied to workplace settings may undermine essential workers’ perceived self-efficacy to practice optimal social distancing. Rotter’s theory⁴⁶ on the internal and external locus of control has been applied to research on the role of self-efficacy in the implementation of a variety of health behaviors.^47–49 The internal locus of control (i.e., the belief that one controls their own health-related actions) is typically associated with positive health outcomes.^46,48,49 If essential workers attribute their social distancing to workplace leadership or local governments (an external source of control), it may therefore undermine their perceived self-efficacy to engage in this novel health behavior, particularly in circumstances where those mandates are relaxed by external authorities or in settings where they are not applicable.^46,48

With respect to race and ethnicity, findings suggest that those in the sample who identified as Black or other unspecified race or ethnicity were more likely to report stricter social distancing behaviors than those who identified as White. Given evidence of substantial disparities in COVID-19 mortality and hospitalization rates of Black compared to White populations in the U.S., this finding is especially notable.^50–52 It suggests that disparities related to COVID-19 may not be a result of the mere failure of racial and ethnic minority groups to engage in personal protective behaviors. Instead, these individuals may face additional structural barriers that confer COVID-19 vulnerability which diminish the benefits of social distancing for them (i.e., residential segregation, inadequate health insurance coverage, etc.).^51,53,54

Furthermore, research has found that Black and Latinx populations are more likely to work in an essential industry.⁵¹ As this is the case, rates of exposure to COVID-19 and variability in PPE while at the workplace for essential workers may undermine the stricter social distancing efforts made evident by ethnic minority participants in the study sample.²⁸ Such a finding may also be evidence of health phenomena such as weathering. Weathering is a well-documented process which grounds the increased risk of adverse health outcomes for racial and ethnic minority populations in the experience of chronic stress and disenfranchisement across the lifespan.^55,56 Thus, our findings may indicate the power of such social factors for the Black respondents in the current study who practiced social distancing with higher rigor.

These findings have implications for public policy and COVID-19 prevention guidelines. The current study has shown that self-efficacy to implement social distancing behaviors is an entry point for the improvement of social distancing practices among those who carry a high burden of the COVID-19 pandemic. While our findings suggest that mandated social distancing practices in the workplace may undermine the self-efficacy to implement these behaviors for essential workers in particular, social distancing guidelines and subsequent mandates from local governments and workplace leadership are arguably unavoidable during the ongoing crisis. However, turning to past research on interventions to improve self-efficacy to engage in novel, positive health behaviors may guide the continued enforcement of social distancing directives at the workplace (i.e., modeling of effective social distancing practices, providing positive feedback for optimal social distancing, etc.).^57,58 Efforts to bolster the self-efficacy to practice social distancing may prove vital considering the COVID-19 pandemic’s expected duration in the absence of pharmaceutical interventions.⁵⁹

With respect to SGM essential workers specifically, data collection for the current study ended before three landmark Supreme Court rulings that extended workplace discrimination protections to SGM individuals.^20–22 As such, a number of factors at the policy-level may explain our findings of low self-efficacy to engage in social distancing among this subsample: (1) Prior to these legal cases, SGM individuals in the workforce were already at risk of job loss due solely to their sexual or gender identity;⁶⁰ (2) limited research has shown that social distancing at the workplace continues to produce challenges for essential workers;⁶¹ and (3) research prior to the aforementioned Supreme Court decisions found that state-level non-discrimination edicts are associated with increased productivity in the workplace.⁶⁰ Therefore, our findings may indicate that job loss is already top of mind for SGM essential workers and, by extension, they may report lower self-efficacy to practice social distancing at the workplace in order to maintain productivity and mitigate said job loss risk. Provided the job security such legal rulings may provide, data collected today may indicate higher rates of self-efficacy to engage in social distancing for SGM essential workers.

Study findings regarding perceived self-efficacy to enact social distancing practices for the sample broadly also have implications for future research and clinical practice. Minority stress theory suggests that under duress, those with marginalized sexual and gender identities turn to their communities and interpersonal relationships more readily than others in order to cope.²⁵ Limited research on COVID-19-related stress and anxiety has found that SGM populations experience higher rates of stress and lower rates of optimal adherence to CDC COVID-19 prevention guidelines than the general public.⁶² The implications of our findings may contextualize the desire for social support among SGM populations during the current crisis, a claim which necessitates further research on the role of social support for those at unique risk of infection during the pandemic.^25,62 Future research should further disaggregate SGM identities as it related to worker experiences given that they face unique challenges based on societal norms.

There are several limitations of note to the current study. Self-report measures on social distancing may not accurately capture actual social distancing practices of participants, but rather their perceived social distancing practices. The study could also not detect the granular, though meaningful, differences within certain items (e.g., commuting to work via public transportation or by car, social gatherings while implementing other prevention efforts like face covering, etc.). As a result, capturing CDC defined social distancing in practice presents challenges.⁴ Also, considering the diverse group of essential industries found in the current study, generalizing the efficacy of social distancing across categorically different industries may disregard the difficulties unique to certain workplaces which may make the ability to practice social distancing more challenging than in others. Finally, social distancing guidelines have varied and evolved across state lines since the COVID-19 pandemic took root in the U.S.⁶³ Therefore, social distancing practices interpreted as suboptimal may be a reflection of the maximum required by order of certain states’ officials and not necessarily as a function of low social distancing rigor.

Conclusions

Overall, the study findings support the conclusion that perceived self-efficacy to implement social distancing behaviors predicts social distancing practices among SGM populations in the U.S. Further research on workplace settings and PPE requirements for those settings may add to burgeoning research on the plight of essential workers in the U.S. and their self-efficacy (as informed by workplace conditions) to practice social distancing in the workplace. Our study also has implications on the policy-level for Black SGM populations during the pandemic. Namely, study findings indicate that Black respondents practiced greater social distancing rigor, yet concurrent research has found that ethnic minority groups have higher morbidity and mortality rates during the pandemic than their White counterparts.⁵² Such findings indicate a need for institutional changes in health care and workplace settings along the lines of (1) linking such at-risk populations to adequate care and (2) combatting the structural racism and discrimination that contribute to health and socioeconomic vulnerability.

Finally, our study highlights the health disparities between essential and non-essential workers in the U.S. As such, policy changes would be warranted to protect SGM essential workers considering their representation in the workforce and high risk for infection.¹¹ Considering the growing research on the projected long-term effects of a second wave of the COVID-19 pandemic in the latter half of 2020, such research is needed in order to better tailor prevention practices and preclude the deepening of already existing health disparities in the U.S.⁶⁴

Public Health Significance:

During the U.S. COVID-19 pandemic, essential workers are at greatest risk of COVID-19 infection. Essential workers are simultaneously asked to work and protect themselves from contracting COVID-19. The current study reveals that essential workers are in need of policy mechanisms to facilitate adequate measures to socially distance and continue working.