Employee work status, mental health, substance use, and career turnover intentions: An examination of restaurant employees during COVID-19

Abstract

The COVID-19 pandemic has negatively impacted the restaurant industry and employees in the worst possible way. This empirical study aims to examine the relationships between employees’ work status (working, furloughed, or laid-off), mental health (psychological well-being and psychological distress), substance use (drug and alcohol use), and career turnover intentions during the pandemic. Analyzing the responses of 585 restaurant employees using structural equation modelling (SEM), findings revealed that working employees experienced higher levels of psychological distress, drug and alcohol use than furloughed employees. Moreover, psychological distress increased drug and alcohol use, as well as career turnover intentions. Lastly, all employees, regardless of their mental health, increased their substance use and indicated a desire to seek future employment in alternate industries during the pandemic. Practical and theoretical implications are discussed in detail.

1. Introduction

The COVID-19 pandemic, which first confirmed case of infection was reported in the United States on January 20, 2020 (Holshue et al., 2020), has resulted in a multitude of lost lives, untold economic devastation, and caused the global economy to shut down almost overnight (UNWTO, 2020). Restrictions to try to stem the spread of the virus have included orders to stay at home, travel limitations, social distancing, and the temporary closure of many businesses, especially in the hospitality sector (Bartik et al., 2020). There has also been a decrease in demand for many businesses that have been permitted to continue to operate (Bartik et al., 2020). During the month of April 2020 alone, job losses in the United States increased by 15.9 million, resulting in the national unemployment rate increasing by a record 10.3 percentage points to 14.7% (U.S. Bureau of Labor Statistics, 2020). The hospitality industry has been particularly hard hit, with most restaurants having to curtail their on-premise dining and restrictions on travel, leading to a precipitous drop in tourism and hotel occupancy rates (Gursoy and Chi, 2020). Two out of three restaurant workers have lost their jobs, equating to eight million furloughed or laid-off restaurant workers (National Restaurant Association, 2020). By late April 2020, all 50 states and the District of Columbia had ordered the closing or restricting of foodservice establishments in response to COVID-19, with most restricting restaurants to take-out/curbside pickup and delivery only (Restaurant Law Center, 2020). While by June 2020 most states had permitted restaurants to start reopening, with restrictions ranging from limited seating capacity, limited party size, and physical distancing to no restrictions at all (Sontag, 2020), many consumers reported that they were likely to leave home and eat at restaurants less often due to concerns about the virus (Maze, 2020). In fact, research has shown that even with an easing of restrictions, more than 50% of individuals would be unwilling to dine at restaurants right away (Gursoy et al., 2020). In July 2020 it was reported that over 26,000 restaurants listed on the Yelp website had closed, and that nearly 16,000 of these would not be reopening (Pesce, 2020). Given this dramatic scenario, it is timely to examine how restaurant workers are faring under such stressful circumstances, especially with regard to their mental health and, more specifically, its antecedents and outcomes.

Recent studies have shown that about one in five Americans lives with a mental illness (National Institute of Mental Health, 2019), and a large proportion of diseases are attributed to neuropsychiatric disorders, mostly due to the chronically disabling nature of depression and other mental disorders, as well as alcohol- and substance-use disorders (Prince et al., 2007). Along these lines, a study performed in the hospitality industry revealed that 80% of employees believe that mental health issues, such as feeling depressed, anxious, or manic, represent a major challenge facing the industry (Randall, 2018). In the restaurant industry, because of its fast paced and stressful work environment, the numbers are even more staggering. For instance, a two-year study, which surveyed 17,000 employees in 19 industries, not only concluded that the work environment in the food and beverage industry correlated with a high level of mental health issues, but that it was one of the three worst industries to work in, along with retail and manufacturing (Mental Health America, 2017). Further, the restaurant industry is ranked the highest of any industry for illicit drug use and third highest for heavy alcohol consumption (Bush and Lipari, 2015). Such drug and alcohol use have been posited to lead to career dissatisfaction (Deery and Jago, 2015) and, in turn, career turnover (McGinley et al., 2014).

While previous studies have associated unemployment with psychological distress (Murphy and Athanasou, 1999; Winefield et al., 2012), drug use (Compton et al., 2014; French et al., 2001), and alcohol use (Khan et al., 2002; Popovici and French, 2013), one in three restaurant workers still working during the pandemic are doing so under quite different and possibly more stressful circumstances, especially in light of the heightened possibility of exposure to the virus. Given the concerning statistics regarding the number of furloughed and laid-off restaurant workers in the wake of the COVID-19 pandemic, and especially coupled with the equally concerning statistics about restaurant industry employees’ mental health, this study sets out to examine the relationships between restaurant employees’ work status (working, furloughed, or laid-off), mental health (psychological well-being and psychological distress), substance use (drug and alcohol use), and career turnover intentions during the pandemic.

As previously stated, work status has been divided into three groups for the purposes of this study. While “working” implies that the person is still employed in their restaurant job during the pandemic, those “furloughed” and “laid-off” are not working and may, thus, be considered to be “unemployed.” An explanation of the main differences between these two groups is therefore warranted. According to the Society for Human Resource Management (2020), a furlough is a temporary leave of absence mandated by an employer, often when there is insufficient work for all employees, whereas a layoff is a termination for which the employee holds no blame. Furloughed employees may be able to continue to receive benefits coverage and collect unemployment insurance for the reduction in time worked. A layoff may be permanent or may be called temporary, although the latter may become permanent. Some employers may continue benefits coverage to laid-off employees for a specific period of time, and laid-off workers are typically eligible to receive unemployment benefits.

The SARS epidemic that originated in Asia in 2002 and rapidly spread to five continents spawned a number of academic studies. These studies included restaurant diners’ self-protective behavior and restaurant management’s attempts to modify workers’ self-protective behavior through various incentives (Chuo, 2014). However, no previous study has been found to have investigated the relationships between employee work status, mental health, substance use and career turnover intentions in the restaurant industry during a major health and economic crisis.

Results of this study revealed that work status affects employees’ mental health differently during a pandemic than previously reported in studies conducted prior to the pandemic. More specifically, the current study revealed that employees who were still working during the COVID-19 pandemic experienced higher levels of psychological distress, drug use, and alcohol use than those that had been furloughed. Furthermore, the current study also revealed that psychological well-being increases drug and alcohol use, thereby contradicting previous studies. Given the importance of understanding how the investigated relationships may differ because of the added stresses of the pandemic, this study begins to fill a gap in the literature and, additionally, provides important theoretical and practical implications for restauranteurs.

2. Literature review

2.1. Relationship between employee work Status and mental health

The World Health Organization (2004) defines mental health as “a state of well-being in which the individual realizes his or her own abilities, can cope with the normal stresses of life, can work productively and fruitfully, and is able to make a contribution to his or her community.” Over the years, researchers have developed a variety of mental-health survey instruments that include both positive and negative measurement items, in order to more accurately portray psychological or mental health (Winefield et al., 2012). Although not exactly opposite theoretical ends, psychological well-being and psychological distress are often used to assess the overall mental health of the general population (Jiang, 2020). Psychological well-being is usually conceptualized as a combination of positive affective states (the hedonic perspective) and functioning with optimal effectiveness in individual and social life (the eudaimonic perspective) (Deci and Ryan, 2008); while psychological distress is largely defined as a state of emotional suffering characterized by symptoms of depression (e.g., lost interest; sadness; hopelessness) and anxiety (e.g., restlessness; feeling tense) (Mirowsky and Ross, 2002). Prior studies have addressed the relationships between employees’ work status (working, furloughed, laid-off) and psychological well-being and distress, although such studies were not performed during global economic and sanitary crises such as the COVID-19 pandemic, and few studies have focused on these relationships in the hospitality industry, the sector that has suffered the most negative economic impact since the start of the pandemic.

For instance, a study by Winefield et al. (2012) demonstrated that low psychological well-being was associated with being unable to work and receiving a pension or benefit. Similarly, a meta-analytical study, which examined the impact of unemployment on worker well-being across 104 empirical studies with 437 effect sizes, found that unemployed individuals had lower psychological well-being than did their employed counterparts (McKee-Ryan et al., 2005). Based on these findings, it is reasonable to suggest that employees that are still working in the restaurant industry during the pandemic will tend to experience higher psychological well-being than those that have been furloughed or laid-off; and employees that have been furloughed will tend to experience higher psychological well-being than those that have been laid-off, since furloughed employees have a higher probability of still be receiving benefits (e.g., health insurance) and may be able to return to work, while laid-off employees tend not to receive benefits, and their chances of getting back to work are slimmer in comparison to furloughed employees. Consequently, the following hypothesis is proposed:

H1a

The work status of employees will affect their psychological well-being.

In that same vein, a study by Murphy and Athanasou (1999), which conducted a meta-analysis on the effect of gaining or losing employment on mental health, showed that job loss was associated with an increase in psychological distress. Likewise, in a study by Winefield et al. (2012), high psychological distress was associated with being unable to work and receiving a pension or benefit. As a result, it is probable that employees who are working will experience lower psychological distress than those who have been furloughed or laid-off, while employees who have been furloughed will probably suffer less psychological distress than those who have been laid-off for the same reasons as described earlier. Hence, the following hypothesis is suggested:

H1b

The work status of employees will affect their psychological distress.

2.2. Relationship between employee work Status and substance use

A study by the Substance Abuse and Mental Health Service Administration (Bush and Lipari, 2015) performed in the U.S. ranked the restaurant industry highest among 19 industries for illicit drug use and third highest for heavy alcohol consumption. Likewise, in a secondary analysis of a large nationally representative survey used to identify the prevalence of alcohol and drug use and intoxication at work, which included a total of 9,828 workers aged at least 14, differential patterns were identified by drug type, worker characteristics and occupational setting (Pidd et al., 2011). The results revealed that nearly nine percent of workers surveyed usually drank alcohol and 0.9% usually used drugs at work, and that hospitality industry workers were 3.5 times more likely than other workers to drink alcohol and two to three times more likely to use drugs at work or attend work under the influence of alcohol or drugs. Despite these worrisome statistics, no research has been conducted in the hospitality field to assess the antecedents of drug and alcohol consumption, including the effect of employee work status, especially relevant during the COVID-19 pandemic.

With regard to drug use, most hospitality studies have focused on pre-employment drug testing (e.g., Kitterlin and Moll, 2013; Thomas et al., 2014), mainly due to employers’ concerns about the effects of drug use on outcomes such workplace productivity, absenteeism, and safety (French et al., 2001). Although studies about the relationship between employee work status and drug use in the hospitality field are almost non-existent, this relationship began to be examined more closely in the early 2000s in the fields of psychology, sociology, and economics. For instance, a study that estimated the probability of employment and labor force participation for different types of drug users indicated that chronic drug use was negatively related to employment for both males and females (French et al., 2001). However, non-chronic drug use was not significantly related to employment. In another study, using data from 405,000 non-institutionalized adult participants, results demonstrated that higher rates of past month illicit drug use and past year drug abuse/dependence were found among the unemployed (Compton et al., 2014). Given the aforementioned findings, it is suggested that restaurant employees who are still working during the COVID-19 pandemic will have lower drug usage than those who have been furloughed or laid-off, due to those working still having an income and occupation, despite the financial burdens imposed on restaurants by lockdowns. Moreover, employees who have been furloughed will have lower drug usage than those who have been laid-off, due to most furloughed employees still receiving benefits and the likelihood of reemployment once their restaurants return to full operation. Hence, the following hypothesis is proposed:

H2a

The work status of employees will affect their drug use.

The same rationale applies regarding the relationship between employees’ work status and alcohol use. Indeed, prior studies have confirmed that unemployment is highly correlated with increased alcohol consumption. For instance, in a study by Compton et al. (2014), higher rates of past month heavy alcohol use and past-year alcohol abuse/dependence were found among the unemployed. These findings are similar to those from a study by Khan et al. (2002), which demonstrated that longer unemployment increases alcohol use, while shorter unemployment decreases it. Lastly, Popovici and French (2013) also showed that unemployment has significantly positive effects on overall alcohol consumption, binge-drinking episodes, and a diagnosis of alcohol abuse and/or dependence. As a result, the following hypothesis is suggested:

H2b

The work status of employees will affect their alcohol use.

2.3. Relationship between mental health and substance use

Both mental health and substance use (alcohol and drugs) are considered integral components of behavioral health (Levin and Hanson, 2020; Robles, 2019) and established topics in psychology and hospitality literature (Kitterlin and Moll, 2013; Zheng et al., 2020). Zhang et al. (2020) found, in their exploratory study, mental health problems to be prevalent in the hospitality workplace, and that employers should, therefore, provide support and resources to their staff. Likewise, several scholars concluded that hospitality employees had a higher consumption of alcohol and drugs than people in other industries (Hight and Park, 2018; Pidd et al., 2011; Pizam, 2010, 2012). As mental health and substance use are among the main drivers of mortality in the United States (The U.S. Burden of Disease Collaborators, 2018) and the leading cause of disability worldwide (Heslin, 2020), further research on the relationship between employees’ mental health and substance use is needed.

While several studies in the medical (Battams and Roche, 2011; LaMontagne et al., 2014; Roberts et al., 2011), sociological (Schwartz, 2002) and management fields (Danna and Griffin, 1999) consider substance use as an outcome of poor mental health, a gap in the hospitality literature regarding the verification of this relationship remains. To the authors’ knowledge, only a single hospitality study so far has confirmed that work stress intensifies employees’ substance use in the restaurant industry (Hight and Park, 2019). In that regard, Kitterlin and Moll (2013) suggested that the restaurant industry needs further investigations on the causes of substance use. With the current COVID-19 pandemic, employees’ psychological well-being represents an important psychological factor that could have a negative effect on restaurant employees’ drug and alcohol usage, and therefore, such factor needs to be taken into consideration and empirically examined during a global sanitary crisis like the COVID-19 pandemic. For instance, studies conducted with children and young adolescents (e.g., Battams and Roche, 2011; Roberts et al., 2011) found that their psychological well-being, known as the positive side of mental health, decreases their substance use. Danna and Griffin (1999) showed a similar result in the workplace environment. Further, a study by Baumeister et al. (2007) demonstrated that self-control, one of the two factors of psychological well-being, represents a mental resource that can prevent negative behavioral outcomes, including substance use. Hence, it is possible to argue that employees who exhibit higher psychological well-being, through increased behavioral control and positive affect, will adopt lower drug and alcohol usage, even during the COVID-19 pandemic. Thus, the following hypotheses are suggested:

H3a

Employees’ psychological well-being will decrease their drug use.

H3b

Employees’ psychological well-being will decrease their alcohol use.

Investigations on the negative side of mental health, also known as psychological distress, have confirmed substance use as an outcome (e.g., LaMontagne et al., 2014). In that same vein, some studies argued that people respond to frustration and distress by increasing the use of alcohol (Horwitz, 2002) and drugs (Schwartz, 2002), in order to forget about their daily problems (Hull, 1981). Adolescents with mental health issues also increase the incidence of substance use (Roberts et al., 2011). In the hospitality field, Pizam (2010, 2012) left an open question to scholars regarding the main reasons why hospitality workers have higher substance use levels than employees from other industries, and assumed that stress could be an explanation. Hence, this question is still open, since few investigations were conducted on this topic in the hospitality industry (e.g., Hight and Park, 2018, 2019; Kitterlin and Moll, 2013). Furthermore, because of the severity of events resulting from the COVID-19 pandemic (e.g., restaurant closures, social distancing, high unemployment), more empirical research is needed in order to understand how restaurant employees’ psychological distress can affect their drug and alcohol use. Based on the aforementioned literature, the present study hypothesizes that:

H4a

Employees’ psychological distress will increase their drug use.

H4b

Employees’ psychological distress will increase their alcohol use.

2.4. Relationship between substance use and career turnover intentions

Career turnover refers to an employee’s “movement to a new occupation that is not part of a typical career progression" (p. 631), as defined in the highly regarded career change model created by Rhodes and Doering (1983). During the conceptualization of their model, Rhodes and Doering (1983) drew upon various job turnover models that were voluntarily initiated by employees. Subsequent studies based on this model have found that career turnover is caused by a variety of factors, such as low levels of person-work environment fit (Donohue, 2006), lack of upward career mobility (McGinley and Martinez, 2018), job insecurity (Shropshire and Kadlec, 2012) and, notably, substance use (Hoffmann et al., 2007).

Staffing and employee turnover has consistently remained a challenge for many firms, including restaurants. Within the hospitality literature, several factors have been found to impact voluntary job turnover, including workplace demands, job stress, and coping strategies such as substance use (Deery and Jago, 2015). Further, prior conceptual models on career change for hospitality employees have argued that illicit substance use can lead to individuals’ desire to leave their current profession and seek employment in an alternative field. For example, Deery and Jago (2015), in their exploration of employee retention strategies, posited that employee alcohol abuse can lead to career dissatisfaction, which in turn, can lead to career turnover. In addition, McGinley et al. (2014) interviewed several hospitality career changers in the hotel industry, and found that the combination of workplace stress plus alcohol use led to the occupational turnover of some respondents.

Currently, empirical evidence analyzing the direct path between substance use and career turnover remains sparse in the hospitality literature. It has been noted, however, that substance use was a prevalent factor among employees in hospitality firms (Giousmpasoglou et al., 2018; Hight and Park, 2018), and that hospitality workers’ substance use had a significant impact on employee job satisfaction and turnover intention (Cain et al., 2018; Hight and Park, 2018, 2019). Based on this rationale, the current study argues that restaurant employees’ drug and alcohol use could lead them to consider leaving their current career, in order to pursue opportunities in other industries. Thus, the following hypotheses are proposed:

H5a

Employees’ drug use will increase their career turnover intentions.

H5b

Employees’ alcohol use will increase their career turnover intentions.

2.5. Relationship between mental health and career turnover intentions

The hospitality industry is a labor-intensive and stressful work environment with high employee turnover (Chang and Busser, 2020; Zhang et al., 2020). Several studies have already explored the influence of stress, burnout, and mental exhaustion on turnover intentions (e.g., Haldorai et al., 2019; Jung and Yoon, 2014). These scholars concluded that a stressful workplace makes employees more likely to leave their jobs. Zhang et al. (2020) found that the hospitality industry shows a prevalence of mental health issues, with stress considered to be the main source of concern for employees. On the other hand, some investigations found negative relationships between psychological well-being and turnover intentions (e.g., Deery and Jago, 2015; Gordon et al., 2019; Wang et al., 2020). Gordon et al. (2019) concluded that measuring well-being is critical for the hospitality industry, since employees' lack of psychological well-being results in higher turnover intentions.

Chang and Busser (2020) argued that turnover investigations focused too significantly on employees’ intention to leave their current job, instead of industry career retention. Likewise, Deery and Jago (2015) concluded that staff turnover is higher among employees who have other career opportunities. McGinley et al. (2014) explored why employees change their careers and leave the hospitality industry, and found that stress was one of the main reasons. Furthermore, Haldorai et al. (2019) found that employees who suffered non-work-related stress showed little motivation to dedicate their physical and mental efforts in their current hospitality career. Therefore, it is possible to conclude that employees who experience psychological distress are more likely to leave their careers in the restaurant industry, while employees who experience psychological well-being are more likely to remain. Hence, the current investigation hypothesizes that:

H6a

Employees’ psychological well-being will decrease career turnover intentions.

H6b

Employees’ psychological distress will increase career turnover intentions.

The conceptual framework of this study is represented in Fig. 1 .

Fig. 1.

Conceptual framework.

*: Three groups were included (working, furloughed, and laid-off), and furloughed employees were used as the reference group.

3. Methodology

3.1. Data collection

This study was designed to examine the relationships between restaurant employees’ work status, mental health (psychological well-being and psychological distress), substance use (drug and alcohol use), and career turnover intentions. The population of the study was defined as non-managerial employees who currently work or have worked in the restaurant industry immediately prior to the COVID-19 pandemic. As such, this study’s sample was defined as restaurant employees: (a) currently working, furloughed, or laid-off during the COVID-19 pandemic; and (b) who were at least 18 years of age.

To address this study’s research objectives, a questionnaire was developed using the Qualtrics online survey platform. An online marketing company was then utilized to identify and distribute the survey to qualified participants in the month of June 2020. To ensure the qualifying criteria of age and work tenure, each respondent was required to answer several screening questions designed to verify eligibility. Overall, 300 responses were collected from each of the three groups (working, furloughed, laid-off), a total of 900 responses. After removing respondents who failed the screening questions and those with missing data, a total of 585 samples were used for the study, a usable response rate of 65%.

3.2. Research instrument

This study used a self-reported questionnaire to survey non-managerial employees (and former employees) in the restaurant industry. After satisfying all screening questions, the respondents confirmed their current employee work status – working, furloughed, or laid-off. Mental health was measured using the 18-item Mental Health Inventory originally developed by Veit and Ware (1983). Mental health is comprised of two dimensions, psychological well-being and psychological distress. All mental health items were assessed using a 7-point Likert-type scale ranging from never (1) to always (7).

Items used to measure drug use were adapted from the Obsessive Compulsive Drug Use Scale (Franken et al., 2002), while the Short Michigan Alcoholism Screening Test (Selzer et al., 1975) was taken to measure employee alcohol use. To measure career turnover intentions, items were adapted from Farkas and Tetrick’s (1989) study. All drug use, alcohol use, and career turnover intention items were measured using a 7-point Likert-type scale ranging from strongly disagree (1) to strongly agree (7). In addition, various sociodemographic questions were included in the questionnaire.

3.3. Analysis method

This study adopted a two-step approach suggested by Anderson and Gerbing (1992). In the first step, the validity of each construct was tested by confirmatory factor analysis (CFA). Once the validity of each construct had been confirmed, structural equation modelling (SEM) was used to test the proposed hypotheses.

4. Results

4.1. Descriptive statistics

Among 585 samples, 59.49% of participants were male and 40.17% were female. In terms of age, 47.35% of participants were between 18 and 30, and 33.50% were between 31 and 40. Furthermore, 51.62% indicated that they had a 4-year college degree and 18.46% held a master’s degree. In terms of employee work status, 39.83% of participants were still working in the restaurant industry, 35.56% were laid-off, and 24.62% were furloughed. Descriptive statistics are presented in Table 1 .

Table 1.

Descriptive statistics.

Variable		Frequency	Percent
Gender	Male	348	59.49
	Female	235	40.17
	Other	2	0.34
Age	18 to 30	277	47.35
	31 to 40	196	33.50
	41 to 50	73	12.48
	51 to 60	31	5.30
	60 or higher	8	1.37
Annual	Less than $10,000	13	2.22
Household	$10,000 to $19,999	37	6.32
Income	$20,000 to $29,999	62	10.60
	$30,000 to $39,999	75	12.82
	$40,000 to $49,999	90	15.38
	$50,000 to $59,999	137	23.42
	$60,000 to $69,999	50	8.55
	$70,000 to $79,999	38	6.50
	$80,000 to $89,999	23	3.93
	$90,000 to $99,999	24	4.10
	$100,000 to $149,999	31	5.30
	$150,000 or more	5	0.85
Education	Lower than High School	1	0.17
	High School or equivalent	69	11.79
	2-year college	95	16.24
	4-year college or university	302	51.62
	Master's Degree (MS)	108	18.46
	Doctoral Degree (PhD)	2	0.34
	Professional Degree (MD, JD, etc.)	8	1.37
Employment	Laid-off	208	35.56
Status	Furloughed	144	24.62
	Working	233	39.83
Total		585	100.00

4.2. Measurement model

Prior to proceeding to the structural model, a confirmatory factor analysis (CFA) was conducted to test the validity of constructs (Table 2 ). χ² was significant at 1% significance level. However, due to the fact that χ² value is sensitive to large sample sizes (Hair et al., 2010), it was necessary to check other goodness-of-fit indices. All indices were higher than the minimum threshold (NFI = .935, TLI = .948, CFI = .953, RMSEA = .064), and thus, the model appropriately fitted the data.

Table 2.

Measurement model.

		Loading	S.E.	t-value	AVE	CR	Cronbach's α
Psychological	depression_2	0.881			0.742	0.953	0.955
Distress	depression_3	0.882	0.035	30.497***
(PD)	depression_4	0.894	0.028	36.921***
	anxiety_1	0.809	0.037	25.601***
	anxiety_2	0.868	0.033	29.458***
	anxiety_4	0.850	0.034	28.251***
	anxiety_5	0.843	0.034	27.817***
Psychological	emotcont_1	0.677			0.618	0.906	0.911
Well-Being	emotcont_2	0.663	0.051	20.297***
(WB)	paffect_1	0.803	0.07	17.365***
	paffect_2	0.838	0.073	17.999***
	paffect_3	0.868	0.075	18.518***
	paffect_4	0.843	0.073	18.088***
Drug Use	drug_1	0.920			0.867	0.983	0.983
(DU)	drug_2	0.934	0.025	41.841***
	drug_3	0.935	0.025	42.003***
	drug_4	0.918	0.026	39.489***
	drug_6	0.934	0.025	41.885***
	drug_7	0.944	0.025	43.510***
	drug_8	0.933	0.025	41.707***
	drug_9	0.942	0.024	43.206***
	drug_10	0.919	0.026	39.637***
Alcohol Use	alcohol_2	0.924			0.856	0.973	0.973
(AU)	alcohol_3	0.935	0.024	42.286***
	alcohol_4	0.921	0.025	40.183***
	alcohol_6	0.915	0.026	39.280***
	alcohol_7	0.932	0.025	41.776***
	alcohol_1	0.924	0.024	40.507***
Career Turnover	turnover_1	0.830			0.884	0.884	0.896
Intention	turnover_2	0.793	0.039	23.910***
(CTRI)	turnover_3	0.797	0.088	11.096***
	turnover_4	0.819	0.089	11.225***

Model fit: χ² = 1530.260, df = 448, Normed χ² = 3.416, NFI = .935, TLI = .948, CFI = .953, RMSEA = .064.

Note:

^***p < .01.

Second, convergent validity was tested. After eliminating measurement items with low factor loadings, standardized factor loadings ranged from .663 to .944. The average variance extracted (AVE) values were higher than the .50 cutoff value (Hair et al., 2010). The values for construct reliability (CR) were calculated, and all values were greater than the .70 cutoff value (Hair et al., 2010). Cronbach’s α values were calculated for each construct. The results confirmed that all α values were higher than the .70 cutoff (Nunnally, 1978). In summary, it can be concluded that constructs’ convergent validity was confirmed. To test for discriminant validity, AVE values were compared to squared correlations between constructs (Fornell and Larcker, 1981). Results suggested that all AVE values were greater than the squared correlation between constructs. Thus, discriminant validity was also confirmed (Table 3 ).

Table 3.

Discriminant validity.

	1	2	3	4	5
Well-being	(0.618)
Distress	0.000	(0.742)
Drug	0.031	0.462	(0.867)
Alcohol	0.027	0.491	0.850	(0.856)
Turnover	0.004	0.138	0.030	0.025	(0.884)

Note: The diagonal numbers in parentheses indicate the AVE. The remaining numbers are squared correlations.

4.3. SEM results

A structural equation model (SEM) was used to test the proposed hypotheses. The goodness-of-fit indices of the structural model suggest that the model fits the data well (χ² = 1616.962, df = 504, NFI = .933, TLI = .947, CFI = .953, RMSEA = .061). The SEM results are presented in Table 4 .

Table 4.

SEM results.

Hypotheses				Estimate	S.E.	t-value	Results
H1a	Working	→	WB	0.026	0.094	0.473	Not supported
	Layoff	→	WB	0.078	0.096	1.437
H1b	Working	→	PD	0.138	0.158	2.599***	Partially supported
	Layoff	→	PD	0.033	0.161	0.631
H2a	Working	→	DU	0.125	0.150	3.270***	Partially supported
	Layoff	→	DU	−0.005	0.153	−0.136
H2b	Working	→	AU	0.077	0.153	2.018**	Partially supported
	Layoff	→	AU	−0.022	0.156	−0.578
H3a	WB	→	DU	0.176	0.074	5.368***	Not supported
H3b	WB	→	AU	0.166	0.076	5.089***	Not supported
H4a	PD	→	DU	0.665	0.048	18.376***	Supported
H4b	PD	→	AU	0.691	0.049	19.078***	Supported
H5a	DU	→	CTRI	0.052	0.082	0.431	Not supported
H5b	AU	→	CTRI	−0.279	0.084	−2.240**	Not supported
H6a	WB	→	CTRI	0.095	0.069	2.139**	Not supported
H6b	PD	→	CTRI	0.525	0.059	8.117***	Supported

Model fit: χ² = 1616.962, df = 504, normed χ² = 3.208, NFI = .933, TLI = .947, CFI = .953, RMSEA = .061.

Note:

^***p < .01.

^**p < .05.

First, it was hypothesized that employee work status would have a significant influence on participants’ mental health (H_1a and H_1b). Due to this study adopting employee work status as exogenous variables, it was necessary to create two dummy variables (working and laid-off), making people who were furloughed as the reference group. No significant influence of employee work status on psychological well-being was found, indicating that H_1a was not supported. However, when it comes to the psychological distress, a significant influence was found. More specifically, compared to the reference group (furloughed participants), employees who were still working experienced significantly higher psychological distress (β_{Working → PD} = .125, p < .01). However, participants who were laid-off did not experience any difference in psychological distress compared to those who were furloughed. Thus, H_1b was partially supported.

Second, employee work status was hypothesized to have a significant influence on participants’ substance use (H_2a and H_2b). Results suggested that participants who were still working showed a higher tendency to use both drugs and alcohol compared to those who had been furloughed (β_{Working → DU} = .125, p < .01; β_{Working → AU} = .077, p < .05). No significant results were found for those who had been laid-off compared to the furloughed employees. Thus, H_2a and H_2b were partially supported.

Third, the impact of participants’ mental health (psychological well-being and psychological distress) on substance use (drug and alcohol use) was hypothesized (H_3a, H_3b, H_4a, and H_4b). Particularly, a negative influence of psychological well-being on both types of substance use was hypothesized. However, results showed the opposite (β_{WB → DU} = .176, p < .01; β_{WB → AU} = .166, p < .01), thus not supporting H_3a and H_3b. Moreover, participants’ psychological distress was hypothesized to increase both drug and alcohol use (H_4a and H_4b). Indeed, results showed that psychological distress significantly increased drug and alcohol use (β_{PD → DU} = .665, p < .01; β_{WB → AU} = .691, p < .01). Thus, both H_4a and H_4b were supported.

Next, the positive influences of participants’ substance use on career turnover intentions (H_5a and H_5b) were examined. No significant influence of drug use on career intentions was found (β_{DU → CTRI} = .052, p > .05). However, alcohol use significantly decreased career turnover intentions (β_{AU → CTRI} = -.279, p < .05). Thus, H_5a and H_5b were rejected. Lastly, the influences of participants’ mental health on career turnover intentions were examined (H_6a and H_6b). Psychological distress was found to have a significant positive effect on career turnover intentions (β_{PD → CTRI} = .525, p < .01). Thus, H_6b was supported. Comparably, participants’ psychological well-being was found to increase career turnover intentions (β_{WB → CTRI} = .095, p < .05), thus H_6a was not supported.

5. Discussion, implications and future research

5.1. Discussion and theoretical implications

The purpose of this study was to examine the relationships between restaurant employees’ work status, mental health (psychological well-being and psychological distress), substance use (drug and alcohol use) and career turnover intentions during the COVID-19 pandemic. Results revealed that work status affects employees’ mental health differently than previously reported in the extant literature in respect of studies conducted prior to the pandemic.

More specifically, the current study revealed that employees who were still working during the COVID-19 pandemic, versus those who had been furloughed, experienced higher levels of psychological distress, drug use, and alcohol use. Such findings contradict prior studies that associated unemployment with high psychological distress (Murphy and Athanasou, 1999; Winefield et al., 2012), drug use (Compton et al., 2014; French et al., 2001), and alcohol use (Khan et al., 2002; Popovici and French, 2013). Such contradictory findings could be explained by the fact that restaurant workers who are still working during the pandemic are doing so under more stressful circumstances, especially given the heightened possibility of exposure to the virus and a significant reduction in gratuities (Center for Disease Control and Prevention, 2020; Selevitch, 2020).

The effect of employee work status on psychological well-being was not confirmed, thus not supporting the prior findings of McKee-Ryan et al. (2005), which showed that unemployed individuals had lower psychological well-being than their employed counterparts. This finding could be partly explained by a study by Winefield et al. (2012), which revealed that low psychological well-being was associated with both “being unable to work” and “receiving a pension or benefit.” Moreover, with the Coronavirus Aid, Relief and Economic Security (CARES) Act signed into law on March 27, 2020, unemployed individuals received compensation and tax credits from the federal government – a $600 per week increase in benefits to each unemployed recipient for up to four months, and up to $1,200 in refundable tax credit for individuals and $500 for each child. Such financial compensations and tax credits may have contributed to an increase in the overall well-being of furloughed and dismissed employees (Watson et al., 2020).

Additionally, the current study also revealed that psychological well-being increases drug and alcohol use, hence contradicting prior studies which showed the exact opposite (Baumeister et al., 2007; Battams and Roche, 2011; Roberts et al., 2011). Accordingly, several reports have confirmed significant increases in alcohol sales and consumption during the COVID-19 pandemic, possibly because individuals were either coping with the stress provoked by the ongoing health crisis, or by staying in the comfort of their homes having more free and/or flexible time to consume alcoholic beverages (American Heart Association, 2020). Likewise, a recent study by Czeisler et al. (2020) confirmed that a variety of sociodemographic groups – such as younger adults, racial/ethnic minorities, and essential workers – representative of the overall restaurant employee population, reported increased substance use, as well as disproportionately worse mental health outcomes and elevated suicidal ideation during the COVID-19 pandemic.

Nonetheless, the findings of the current study support prior literature in respect of the positive influences of psychological distress on drug use (Roberts et al., 2011; Schwartz, 2002), alcohol use (Horwitz, 2002; LaMontagne et al., 2014), and career turnover intentions (Haldorai et al., 2019; McGinley et al., 2014). Lastly, the proposed effects of psychological well-being (Gordon et al., 2019; Wang et al., 2020) and drug and alcohol use (Deery and Jago, 2015; Hoffmann et al., 2007) on career turnover intentions were not supported. These findings could be explained by previously mentioned phenomena, including that employees’ work status did not have an impact on their psychological well-being, and that working restaurant employees – who tended to experience more psychological distress than their unemployed counterparts – were fortunate enough to keep their jobs, despite the psychological distress generated by their professional activities during such dramatic health crisis.

5.2. Practical implications

From a practical perspective, it is apparent that individuals still working experienced greater instances of negative psychological and behavioral outcomes, such as increased levels of psychological distress, drug use, and alcohol use, versus their furloughed counterpart. As mentioned earlier, these results can be attributed to various reasons.

First, the United States’ government provided a financial stimulus package that increased unemployment benefits for all furloughed and laid-off workers (Watson et al., 2020). Thus, it seems that furloughed employees’ access to a steady income helped to alleviate insecurity related to the uncertainty of their work return (Jones, 2020). Second, employees working during the pandemic were highly impacted due to various and unusual workplace circumstances. For instance, social distancing guidelines and the forced closure of in-house dining meant most working employees were forced to quickly adapt to a brand-new service delivery system geared toward takeout and delivery modalities. Further, working employees also had to contend with the increased risk of COVID-19 exposure, while furloughed and laid-off employees did not. Thus, this study makes it apparent that firms that remain open or re-open while the pandemic is still ongoing need to consider the strains and stressors put onto working employees. For example, in addition to a modified service delivery system and an increased exposure to the virus, working employees were not eligible for unemployment benefits. As such, due to drastic decreases in restaurant sales resulting in a corresponding decrease in gratuities, working restaurant employees were not able to generate income similar to what they were previously accustomed.

As a result, restaurant firms should take the time to prioritize the well-being of their employees by adopting strategies that can mitigate their risk exposure, both physically and financially. For example, restaurant owners could partially furlough their employees and then stagger shifts in a manner that allows all employees to have some working shifts and furloughed days. This strategy could reduce employee stressors by providing a shortened work week, which would decrease exposure to the virus, provide access to some unemployment benefits due to the partial furlough status, and offer ample time for recovery from work shifts. Moreover, employers could inform their restaurant employees about available resources and hotlines – which offer free and unrestricted support in cases of employees experiencing exacerbated mental health and/or substance use issues – such as the National Suicide Prevention Lifeline and the Substance Abuse and Mental Health Services Administration (SAMHSA) National Helpline, among others. Lastly, employers could coach their staff on healthy ways to cope with stress, such as taking care of one’s physical and mental health (e.g., through healthy eating, regular physical activities, quality sleep, and the avoidance of alcohol or drug use), connecting with others (e.g., by frequently socializing and talking with people either online, through social media, or by phone or mail), and taking time to unwind (e.g., by taking breaks and performing enjoyable activities).

5.3. Limitations and future research

This study was designed to examine the relationships between restaurant workers’ employee work status (working, furloughed, laid-off), mental health (psychological well-being and psychological distress), substance use (drug and alcohol use), and career turnover intentions during the COVID-19 pandemic. Although the study provides significant results and implications, it is naturally not free of limitations. First, this study was conducted at a single point in time, when most restaurants were closed for dine-in business or altogether. While this was considered to be optimum timing for conducting this particular research, future studies should investigate how the aforementioned relationships may differ once restaurants return to dine-in service with limitations, and again once the situation normalizes. Second, this study surveyed only restaurant workers in the United States. It would be interesting for future studies to investigate these relationships in other countries to determine whether the results are generalizable across cultures. Third, this study concentrated solely on workers within the restaurant industry. Future studies should investigate the aforesaid relationships in other sectors of the hospitality industry and in other high-touch industries to determine whether the results are applicable across a variety of sectors and industries. Finally, a qualitative study to address the lived experiences of restaurant employees during a pandemic using a phenomenological approach may add greater depth to the current quantitative research.