Difficult Times, Difficult Decisions: Examining the Impact of Perceived Crisis Response Strategies During COVID-19

Abstract

Crises, such as the COVID-19 pandemic, require rapid action to be taken by leaders, despite minimal understanding of the impact of implemented crisis management policies and procedures in organizations. This study’s purpose was to establish a greater understanding of which perceived crisis response strategies were the most beneficial or detrimental to relevant perceptions and outcomes during the recent COVID-19 crisis. Using a time-lagged study design and a sample of 454 healthcare employees, latent profile analysis was used to identify strategy profiles used by organizations based on several policy/procedure categories (i.e., human-resource supportive, human-resource disadvantaging, behavioral/interactional human safety and protection-focused, and environmental and structural safety supports-focused policies and procedures). Results indicated that four perceived crisis response strategies were employed: (1) human resource-disadvantaging, (2) maximizing, (3) safety and human resource-supportive, and (4) inactive. Perceived crisis response strategy was linked to several employee well-being (e.g., work stress) and behavioral (e.g., safety behavior) outcomes via proximal perceptions (i.e., perceived organizational support, ethical leadership, and safety climate). Proximal perceptions were the most positive for employees within organizations that enacted safety and human resource-supportive policies and procedures or that utilized a maximizing approach by implementing a wide array of crisis response policies and procedures. This paper contributes to the literature by providing crucial information needed to reduce organizational decision-making time in the event of future crises.

The onset of the COVID-19 pandemic was a crisis situation characterized by significant loss of life, pressure to rapidly enact strategic responses, and a lack of available resources and information (WHO, 2020). As tremendous progress has been made since that time, it may be enticing to move past the devastation and focus on a more hopeful future. However, we argue that it is important to take a lessons-learned approach by evaluating the effects of implemented policies during the onset of the crisis in order to determine which may be most appropriate to deploy in the future.

Crises have widespread impacts for many entities (e.g., governments, communities), but can be particularly disruptive to organizations, rendering negative effects on employee attitudes and behaviors (Ayoko et al., 2017). However, due to an absence of resources on pandemic strategy approaches (Boin & Lodge, 2021), little is known about the impact that specific policies or strategic human resource responses have on employee perceptions and outcomes during crises. As such, identifying effective perceived crisis response strategies, particularly those not studied in previous research, is crucial in order to take a prescriptive approach for navigating future large-scale disasters which present similar crisis scenarios. Therefore, this study used a novel configural approach (i.e., latent profile analysis; LPA; a technique in which patterns across variables are examined to configure profiles; Spurk et al., 2020) to establish a greater understanding of which strategies enacted in response to the COVID-19 pandemic were the most impactful on relevant perceptions and outcomes in order to inform organizational actions in future crises.

Crisis Management Responses

A crisis is an ambiguous situation that poses an immediate threat to a group’s essential operations or values (Boin & Hart, 2007), arising from a disaster event, such as a pandemic (Shaluf et al., 2003). The crisis management process (i.e., an organized effort to avoid or respond to crises) has been argued to impact relevant outcomes, and scholars have emphasized the importance of observing lessons learned during a crisis to apply to future events (Pearson & Clair, 1998). As such, this study aims to further the crisis management literature by examining actions taken in response to a recent crisis (i.e., the COVID-19 pandemic) to better equip organizations for future health-related crises.

More specifically, by providing evidence of the beneficial or detrimental nature of strategic crisis responses on employee outcomes, the results of this effort can be used to significantly reduce organizational decision-making time in determining which strategies to enact. As crisis management consists of three phases (i.e., pre-crisis, crisis, and post-crisis; Coombs & Laufer, 2018), this study contributes to the post-crisis phase (i.e., a time in which crisis responses are evaluated for learning purposes) of the COVID-19 pandemic to serve as a valuable resource in the pre-crisis (i.e., preparation) phase for future crises and provide a roadmap for subsequent crisis phases. In particular, we present specific, actionable policies that have been empirically demonstrated to be beneficial to employee outcomes and can be immediately implemented by organizations, which is particularly vital as the crisis management literature lacks research investigating the post-crisis phase (Coombs & Laufer, 2018) and detailed, practical crisis management plans (Coombs, 2015) focused on internal stakeholders (Frandsen & Johansen, 2011).

Several conceptual models have been introduced that offer insight into existing conceptualizations of crisis management strategies. Crises impacting the healthcare system require concentrated efforts, as their severe impact on everyday life extends beyond the scope of traditional crisis management (Burkle, 2019). Work has been done to revise general crisis management frameworks (e.g., a three-stage model consisting of prevention and preparedness, response, and recovery and rehabilitation) to be tailored to health crises by considering what each phase would specifically entail in this context (Burkle, 2019). Similarly, Olu (2017) provided a framework highlighting how specific strategies (e.g., acquiring surplus equipment) align with crisis management stages and Heath et al. (2020) highlighted strategies likely to increase worker resilience during the pandemic (e.g., providing organizational justice). Further, reviews of crisis management conceptual models have highlighted the varying foci of crisis management strategies (Wang & Belardo, 2005) and tactics for determining which strategy is most appropriate for a given crisis situation (Harwati, 2013). In addition, research has identified specific competencies of effective leaders in healthcare organizations during crises, such as communication and influencing others (Sriharan et al., 2021). Taken together, the crisis management literature offers useful information regarding the selection of crisis management strategies.

However, the crisis management literature is lacking research examining how organizational internal crisis responses impact employee outcomes (Guzzo et al., 2021; Witala & Mistry, 2022), particularly in the healthcare sector. Therefore, the present effort addresses this gap by empirically examining which relevant outcomes are impacted by perceived enactment of specific strategies. More specifically, we recognize that leadership teams are forced to make rapid decisions in times of crisis and may inadvertently make crucial missteps in the crisis management process (Pearson & Clair, 1998). As such, we rely on theoretical tenets to investigate the relations between crisis response strategies and employee perceptions of their organization, the organization’s leaders, and the work environment, as well as their attitudinal and behavioral outcomes to determine which strategy approaches were associated with positive outcomes.

Perceived Crisis Responses During COVID-19

Although nearly all organizations were affected by the pandemic, healthcare organizations battled the crisis on multiple fronts (Grimm, 2020), including financial loss (The Associated Press, 2020) and increased employee safety risks (Cohen & Nigam, 2020). This is particularly concerning, as healthcare workers were on the frontlines in combating the COVID-19 crisis, and their efforts likely had wide-reaching impacts (Bapuji et al., 2020). Because of the prevalence of crises in this field, healthcare organizations have been deemed a “living laboratory” for the development of crisis management approaches (Nemeth et al., 2011). Thus, the identification of effective perceived crisis management strategies enacted within the healthcare sector is particularly imperative.

There has been widespread variability in enacted policies during COVID-19, with healthcare organization responses ranging from implementing employee furloughs (Frankel & Romm, 2020) to introducing telemedicine services to better serve patients (Centers for Disease Control and Prevention, 2020a). Due to this variability, there is a need to identify which policy implementation approaches were taken in order to determine the differential impacts. Notably, recent efforts have made strides in identifying COVID-19-related human resource management responses in various fields (i.e., hospitality, education), focusing on tactics such as communication and leader behaviors (e.g., Agarwal, 2021; Matthews et al., 2022; Witala & Mistry, 2022). Such efforts provide tremendous value to the literature by extending knowledge on policy-based crisis responses; however, the present effort is distinct in that it provides a specific list of policies that healthcare organizations can rapidly enact, using a latent profile approach to highlight which strategies were most effective to improve decision-making response time in future events. As better understanding the impact of such strategies on workplace outcomes can serve a prescriptive purpose, the first objective of this project was to identify unique profiles of perceived crisis response strategies.

As scholars have endorsed using a profile approach for examining organizational policies (Schulte et al., 2009), the method used to identify unique profiles is LPA, an exploratory analysis that groups participants with similar responses into categorical configural profiles (Spurk et al., 2020). As the profiles that emerge are dependent on which strategies respondents perceive to have been implemented, precise expectations of profiles and expected differences cannot be stated. However, as previous research has established that policymakers can delay or altogether fail to enact crisis response strategies (Bristow & Healy, 2015) and as crisis responses often involve implementing numerous policies in a limited time frame (Huang, 2008), we anticipate the emergence of both an inactive response profile in which no policies were enacted and a maximizing profile in which a multitude of actions were taken simultaneously. Further, as suggested by the diversity of policies compiled by subject matter experts (i.e., SMEs) for the purpose of this study (e.g., some organizations providing bonus pay and others decreasing pay; see Appendix 1), we expected a misalignment between profiles, such that some organizations primarily implemented strategies that were supportive to employees, whereas others implemented those that disadvantaged employees. Taken together, the following research question is proposed on an exploratory basis:

• Research Question 1 (RQ1): What perceived crisis response strategies were employed by healthcare organizations during COVID-19?

Perceived Crisis Response Strategy Impact

Employee perceptions of their organization’s crisis response strategy are likely to directly impact perceptions of the organization, the organization’s leaders, and the work environment. Although a number of theories have put forth explanatory mechanisms for this proposition, we focus specifically on social exchange and organizational support theories as guiding frameworks. Namely, social exchange theory (Blau, 1964; Homans, 1958) suggests that the quality of relationships between two parties is determined by their mutual exchanges (Cropanzano & Mitchell, 2005), which can be applied to employee relationships with both organizational leaders and the employing organization (Brown et al., 2005). Thus, when employees perceive that their organization has enacted favorable policies or procedures, employee attitudes and behaviors are also likely to be positive (Masterson et al., 2000).

Similarly, organizational support theory (Eisenberger et al., 1986) proposes that in the social exchange relationship between employees and an organization, employees develop attributions based on the favorability of the treatment they receive, resulting in reciprocating responses (Kurtessis et al., 2017). For instance, research has suggested that unsupportive organizational practices can result in feelings of exploitation, which can negatively impact employee attitudes and well-being (Caesens et al., 2017). Organizational support theory is rooted in the tenets of social exchange theory, as perceived support from an employee’s organization is the foundation for evaluating social exchange (Wayne et al., 1997).

Research has established that employees form perceptions about leadership through inferential processes, such that employees examine leader actions or processes in which they are involved, then make conclusions about leadership based on the outcome of those events (Jantzi & Leithwood, 1996). As such, perceived policy implementation may serve as a basis for employees to form perceptions about the organization and its leaders. Thus, our second objective was to examine how employee perceptions of the crisis responses employed by their organization were related to employee perceptions of the organization, leadership, and the work environment.

Proximal Employee Perceptions

We define proximal employee perceptions as perceptions that are formed as a direct result of perceived efforts undertaken by the organization in response to the crisis. Just as policies can be wide-ranging, so can the targets of employee attributions and perceptions. Namely, individual attributions of responsibility during organizational crises can vary (Zemba et al., 2006), as blame assignment during an organizational crisis is based on a complex sensemaking process (Gephart, 1993). Thus, whereas social exchange theory would suggest that perceived crisis response policies are likely to impact perceptions of employee treatment, an examination of the specific target(s) of those attributions (e.g., leaders) would provide greater insight into employee perceptions of responsibility.

Thus, in the current study, we focus on employee perceptions of three diverse workplace targets—leaders, the organization, and the work environment. Although a variety of perceptions would likely be impacted by organizational crisis management approaches, we selected three factors that align with the lenses set forth by the aforementioned theories and capture a wide range of phenomena while balancing survey length considerations. Namely, perceptions of ethical leadership, organizational support, and safety climate were included to assess employee perceptions of their leaders, organizations, and work environments, respectively.

Ethical leadership was selected to assess employee perceptions of leadership, as many of the policies enacted by healthcare organizations were innately ethical issues. Namely, healthcare organizations were required to make ethically complex decisions such as allocating scarce life-sustaining resources (e.g., ventilators; Springs, 2022) or evaluating the necessity of employee pay cuts. As severe ethical challenges in regard to healthcare policy and decision-making were raised by the pandemic (WHO, n.d.), examining perceptions of ethical leadership (i.e., a leader’s exhibition and promotion of normative workplace behavior via interpersonal communication, enforcement of consequences, and decision-making; Brown et al., 2005) can provide insight into how judgments of leader attributes are affected by perceived crisis response strategies, thereby providing organizations with a roadmap for effective crisis management.

Ethical leadership has often been identified as an important variable to consider within the context of social exchange theory (e.g., Brown et al., 2005; Mayer et al., 2009) as the theory suggests that leader actions toward employees are the basis of employee perceptions of leadership, such that interactions consisting of reciprocal equity and trust are likely to result in strong perceptions of ethical leadership (Mayer et al., 2009). Previous work investigating perceived ethical leadership in a healthcare context has established that organizational policies are often perceived to be ethical issues, and as leaders enact policies that set expectations for ethical standards (Mayer et al., 2009), organizational policies may impact employee perceptions of ethical leadership (Makaroff et al., 2014). Further, policies pertaining to ethical issues (e.g., pay) can impact employee perceptions of organizational ethics, which is reflective of the leaders implementing such policies (Weeks et al., 2004). Research has not only identified the overlap between policy-making and ethical leadership (Hudson, 1997) but also linked crisis scenarios to ethical leadership, as crises often result in leader decision-making errors (Pearson & Clair, 1998) and have been argued to impair leader abilities to make ethical decisions (Harwati, 2013).

In turn, POS was selected to assess perceptions of organizations, as human resource-focused responses more advantageous to employees (e.g., increased mental health benefits) are likely to impact employee perceptions of organizational support. Perceived organizational support (POS; i.e., an employee’s belief that the organization values their well-being and cares about their welfare) has frequently been examined in the context of social exchange theory due to the reciprocity of support perceptions that occur between an employee and their organization and the importance of those perceptions in determining relevant outcomes (Cropanzano & Mitchell, 2005).

Social exchange theory posits that the favorability of organizational actions are likely to strongly influence POS, particularly when the actions are perceived to be voluntarily implemented (Eisenberger et al., 1986), which is a proposition that has been supported by empirical evidence (Eisenberger et al., 1997). As organizations enacted a wide array of policies during the pandemic in a condensed period of time (Burrill et al., 2020), it is possible that employees perceived these policy implementations to be within the organization’s realm of control, therefore impacting POS as suggested by social exchange theory. Further, employees often view actions by agents of the organization (e.g., human resources) as indicators of the organization’s intentions (Eisenberger et al., 1986), and meta-analyses have demonstrated that favorable human resource practices are positively linked to POS (Kurtessis et al., 2017; Rhoades & Eisenberger, 2002).

Lastly, as a large number of safety-focused policies are frequently enacted during health-related crises, safety climate was selected to assess employee perceptions of their work environment. Safety climate is particularly relevant to employee health in health-related crisis situations (Sinclair et al., 2020) and has been identified to be a critical component in shaping healthcare work environments during active disease outbreaks (Wang et al., 2021). As both theoretical (see, e.g., Zohar, 2014) and meta-analytic work (He et al., 2019) have identified organizational policies as crucial to the development of safety climate perceptions, safety climate was identified as a particularly important factor to measure to assess work environment perceptions in response to crisis-enacted policies.

Notably, Zohar (2014) conceptualized safety climate differently based on the level of theory and analysis. At the individual level, safety climate is defined as an individuals’ perceptions of safety policies, procedures, and practices at work. At the unit level (e.g., organizations), safety climate refers to employees’ shared perceptions of safety policies, procedures, and practices within a unit (Zohar, 2014), calculated by aggregating employees’ perceptions of safety climate across meaningful clusters (Zohar, 2014). Collectively, safety climate emerges from individuals’ sense-making processes about organizational events. Specific to our study, the crisis response policies directly influence individuals’ sense-making processes regarding safety environments, which ultimately form the safety climate. We aimed to focus on this proximal, or direct, influence in the current study; therefore, the current study focuses on perceived safety climate at the individual level.

Safety climate perceptions have been investigated through the lens of social exchange theory. For example, it has been argued that organizational support of safety will result in greater employee safety climate perceptions (Cheung & Zhang, 2020). To this point, several efforts have reported a positive relationship between organizational support and safety climate perceptions (Cheung & Zhang, 2020; DeJoy et al., 2004). Similarly, scholars have theorized that organizational efforts to create a protected work environment will increase employee reciprocation of those values and safety climate perceptions (Cheung & Zhang, 2020). As such, an investigation of perceived ethical leadership, organizational support, and safety climate is warranted within the theoretical framework of social exchange theory and can provide insight into the diverse targets and nature of perceptions that can result from perceived organizational crisis responses.

Taken together, perceived crisis response strategies are likely to impact employee proximal perceptions such that strategies that are perceived by employees to be favorable may be positively related to ethical leadership, organizational support, and safety climate perceptions. As research has not yet identified the common strategic approaches employed by organizations during health-related crises, we broadly expect that the favorability of employee proximal perceptions will reflect the extent to which perceived crisis response strategies included supportive (e.g., increased pay) and maximizing (i.e., enacting a wide array of changes) human resource policies and procedures and avoided disadvantaging (e.g., decreased pay) policies or being inactive (i.e., minimal actions taken).

• Hypothesis 1 (H1): Perceived crisis response strategies that put greater emphasis on supportive and maximizing human resource policies and procedures will be linked to more favorable employee proximal perceptions (i.e., perceptions of ethical leadership, organizational support, and safety climate) than those that enacted primarily disadvantaging policies or were inactive.

Distal Employee Effects

Social exchange theory (Blau, 1964; Homans, 1958) suggests that subjective assessments play a key role in determining future attitudes and behaviors. Kurtessis et al. (2017) provided meta-analytic evidence for the impact of social exchange relationships, such that organizational policies and practices were linked to employee POS, which were related to employee well-being, work behavior, and orientation toward the organization. As this pattern of relationships has been demonstrated, it is likely that a similar trend will occur in the present effort via mediation effects. Although work orientation factors are certainly important outcomes that can have long-term implications, because of the toll the COVID-19 pandemic has taken on healthcare workers (Pappa et al., 2020), the current study focuses on the arguably more pressing concerns of employee well-being and behavior. More specifically, we examine two facets of employee well-being (i.e., job satisfaction and work stress) and two work behavior outcomes (i.e., turnover intentions and safety behavior).

Much support has been garnered for organizational support theory’s proposition that an employee’s perceptions about the extent to which they are valued by their organization are predictive of their subsequent attitudes and behaviors. For instance, in regard to employee well-being, perceptions of ethical leadership, organizational support, and safety climate have been positively linked to employee satisfaction (Church, 1995; Harris et al., 2007; Huang et al., 2016; Nixon et al., 2015) and inversely related to psychological stress or job strain (Chen et al., 2017; Elçi et al., 2012; Harris et al., 2007; Yu & Li, 2020). Further corroborating organizational support theory’s proposition regarding work behaviors, perceptions of ethical leadership, organizational support, and safety climate have been inversely linked to turnover intentions (Elçi et al., 2012; Harris et al., 2007; Nixon et al., 2015) and positively linked to safety behaviors (i.e., individual actions aimed at improving the health and safety; Christian et al., 2009; Chughtai, 2015; Hofmann & Morgeson, 1999). Thus, using both theoretical and empirical work as a foundation, we expect that proximal perceptions formed during COVID-19 will be an intermediary in the link between perceived crisis response strategies and distal employee outcomes (see Fig. 1).

• Hypothesis 2 (H2): Perceived crisis response strategy will impact distal employee outcomes via proximal employee perceptions such that individuals in organizations that are perceived to employ supportive or maximizing response strategies will perceive more favorable proximal perceptions (i.e., perceptions of ethical leadership, organizational support, and safety climate) than those that are perceived to employ disadvantaging or inactive response strategies, leading to more positive employee distal outcomes (i.e., job satisfaction, work stress, turnover intentions, and safety behavior).

Fig. 1.

Hypothesized model in which perceived crisis response profiles are expected to predict distal effects via proximal perceptions

Specific Perceived Crisis Response Policies and Procedures

Due to the unprecedented nature of the COVID-19 crisis, entities worldwide had to tackle numerous challenges with limited decision-making time (Burrill et al., 2020). This sense of urgency resulted in the rapid implementation of wide-ranging policies and procedures likely to have varying impacts (Burrill et al., 2020). In order to provide prescriptive information on which specific perceived crisis response policies and practices were the most impactful, we will conduct an exploratory examination of the links between specific perceived policies and practices implemented during COVID-19 and employee proximal perceptions and distal outcomes.

• Research Question 2 (RQ2): Which perceived crisis response policies and practices have the strongest links to employee proximal perceptions and distal outcomes?

Method

Participants

Individuals employed at healthcare facilities were recruited via the ResearchMatch platform, online healthcare forums and social media pages, researcher personal networks, and direct solicitations to hospitals. Participants who completed both surveys included 454 individuals who were at least 18 years of age and employed at a healthcare facility within the USA. In the sample, 66.5% worked in medical roles within their healthcare facilities, and 81.9% reported having direct interaction with patients in their roles.

Procedure

This study consisted of two data collection phases between May and August 2020, as time-lagged study designs are necessary to infer directionality in a mediation model (Cole & Maxwell, 2003). Phase one assessed policies enacted by organizations during COVID-19 and proximal employee perceptions, in addition to demographics, covariate measures, and distal outcomes to collect initial outcome scores. Approximately 30 days later, participants provided additional data on the distal effect variables, which were used in analyses as outcome variables. Participants were entered into a raffle for one of three $100 gift cards each time they completed the survey.

Measures

For each measure discussed below, five-point agreement scales were used, unless otherwise stated. Reliability coefficients are reported in Table 1.

Table 1.

Descriptive statistics and scale reliabilities for all study variables

Variable	M	SD	1	2	3	4	5	6	7	8	9
1. Age	41.29	13.51
2. Gender	0.85	0.36	− .07
3. Education	0.82	0.39	− .10*	− .07
4. Race	0.14	0.34	− .07	− .05	− .05
5. Career tenure	13.43	11.74	.74**	.04	− .03	− .07
6. Job tenure	7.10	8.05	.56**	.01	.03	− .12*	.66**
7. Patient interaction	0.82	0.39	− .10*	.00	− .06	.10*	− .12*	− .16**
8. Work location	0.45	0.50	− .01	− .06	− .13**	.05	− .06	− .11*	.05
9. Hours worked	40.71	19.24	− .03	− .05	.04	.01	.02	− .04	− .07	.04
10. # of children	0.40	0.80	− .10*	− .04	− .02	− .01	− .04	− .01	.01	.03	.05
11. Work intensity	3.32	1.01	− .05	.02	− .02	− .07	.03	.00	− .06	.04	.20**
12. Job title	0.33	0.47	.04	− .05	− .10*	− .01	− .06	.08	− .26**	.13**	.05
13. Human resource-supportive	7.86	3.84	.03	.06	.10*	− .07	.02	.15*	− .12**	− .30**	.01
14. Human resource-disadvantaging	2.57	1.55	.02	− .01	.01	− .06	.07	.13*	− .13**	− .10*	.00
15. Behavioral/interactional human safety and protection-focused	6.41	2.14	.00	.05	.06	− .01	− .02	.12*	− .04	− .24**	.05
16. Environmental and structural safety supports-focused	12.58	4.38	.00	.03	.12*	− .06	− .02	.17*	− .08	− .28**	.06
17. Perceived ethical leadership	3.62	0.97	.06	.00	.10*	− .01	.01	.06	− .12*	− .06	− .01
18. Perceived organizational support	3.43	1.04	.03	− .01	.09*	.05	− .02	.06	− .08	− .03	− .05
19. Perceived safety climate	3.85	0.93	− .06	.03	.09	− .01	− .10*	.05	− .03	− .09*	.02
20. Job satisfaction	3.86	0.91	.17**	− .02	.04	.04	.11*	.08	.03	− .06	− .01
21. Work stress	3.28	0.79	− .11**	.06	− .01	− .03	− .03	− .01	.02	.03	.20**
22. Turnover intentions	2.31	1.25	− .13**	.00	− .02	.03	− .08	− .12*	.01	.09*	.10*
23. Safety behavior	4.26	0.58	.24**	− .04	− .06	− .01	.15**	.13**	.03	.07	.09

Variable	10	11	12	13	14	15	16	17	18	19	20
11. Work intensity	− .06	(.86)
12. Job title	− .06	.03
13. Human resource-supportive	.04	− .07	.00	(.81)
14. Human resource-disadvantaging	.05	.14**	.05	.19**	(.47)
15. Behavioral/interactional human safety and protection-focused	.02	− .10*	.07	.62**	.24**	(.72)
16. Environmental and structural safety supports-focused	.09	− .07	.06	.67**	.27**	.83**	(.86)
17. Perceived ethical leadership	.10*	− .25**	.13**	.46**	.02	.44**	.47**	(.95)
18. Perceived organizational support	.10*	− .30**	.11*	.42**	− .04	.36**	.38**	.86**	(.95)
19. Perceived safety climate	.11*	− .19**	.13**	.42**	.06	.44**	.47**	.73**	.76**	(.93)
20. Job satisfaction	.11*	− .31**	− .04	.24**	− .01	.20**	.18**	.48**	.56**	.51**	(.90)
21. Work stress	− .03	.58**	.00	− .07	.13**	− .03	− .01	− .30**	− .38**	− .29**	− .53**
22. Turnover intentions	− .05	.26**	.03	− .17**	− .01	− .11*	− .14**	− .37**	− .44**	− .45**	− .70**
23. Safety behavior	.08	− .06	.06	.11*	− .01	.11*	.10*	.28**	.22**	.25**	.33**
Variable	21	22	23
21. Work stress	(.94)
22. Turnover intentions	.44**	(.89)
23. Safety behavior	− .15*	− .26**	(.85)

M and SD are used to represent mean and standard deviation, respectively. Items 12–15 represent the average number of policies enacted in each response category per participant. Gender was coded: 0 = male, 1 = female. Education was coded: 0 = less than a bachelor’s degree, 1 = bachelor’s degree and above. Race was coded: white = 0, non-white = 1. Work location was coded: 0 = hospital, 1 = non-hospital facility. Patient interaction was coded: 0 = no, 1 = yes. Job title was coded: 0 = medical staff, 1 = non-medical staff. Scale reliabilities are reported in the diagonal. *p < .05; **p < .01. The four organizational crisis response policy categories consisted of 17 items (human-resource supportive), 7 items (human-resource disadvantaging), 9 items (behavioral/interactional human safety and protection-focused), and 19 items (environmental and structural safety supports-focused)

Perceived Organizational Crisis Policies and Procedures

A list of 52 COVID-19-specific perceived organizational crisis response policies and procedures was compiled using publicly available information and data provided by SMEs. More specifically, we reviewed the literature on organizational crisis response policies and publicly available taxonomies (i.e., Centers for Disease Control and Prevention, 2020b; OHSU, 2020) and consulted 15 healthcare workers in the researchers’ personal networks with a variety of expertise across healthcare facilities and departments to qualitatively generate a comprehensive set of crisis response policies. In this process, the SMEs indicated in an open-response format which policies had been implemented or should have been enacted by healthcare organizations in response to COVID-19. Lists of policies were then independently generated by two members of the research team using the gathered information, who then calibrated their lists to reach full agreement on the initial set of items. Notably, the list of generated policies was directed toward primarily employees but also patients, as policies relevant to patients have a substantial impact on employees.

After identifying distinct policies and procedures, the research team used affinity mapping methodology (see, e.g., Murphy et al., 2018) to generate categories to represent the broad types of policies and procedures encompassed by the measure. More specifically, a qualitative sorting process was used to group perceived organizational policies and procedures into four categories: (a) human resource-supportive (i.e., focused on personnel compensation, benefits, job specifications, and work arrangements likely to be viewed as neutral or beneficial for employees; 17 items; e.g., introduced a temporary pay increase), (b) human resource-disadvantaging (i.e., focused on personnel compensation, benefits, job specifications, and work arrangements likely to be viewed as detrimental to employees; 7 items; e.g., introduced a temporary pay decrease), (c) behavioral/interactional human safety and protection-focused (i.e., focused on managing safety behaviors and/or providing protections in interactions among staff, patients, and/or visitors; 9 items; e.g., introduced visitor restrictions), and (d) environmental and structural safety supports-focused (i.e., focused on regulations related to the medical facility, safety equipment, and/or organization-level strategies for COVID-19 management; 19 items; e.g., imposed limits on PPE use). See Appendix 1 for the comprehensive list of policies comprising each category. The research team reached full agreement on the categorization of each item into the established categories. After these categories were identified, 12 student researchers with knowledge of organizational behavior independently sorted the 52 items into the categories to verify the placement of each item after undergoing rater training. The category that received the majority placement for each item was retained, with the study authors serving as tiebreakers.

We then solicited face and content validity evaluations of the generated policies (e.g., content relevance of the policies, the representativeness of the domain being assessed) from SMEs with various expertise (i.e., occupational health and safety researchers, healthcare practitioners in academia and medical facilities). We also asked for input on whether additional policies were overlooked in the initial version. Their feedback was independently incorporated by the authors through revising, deleting, and adding policy items to ensure the comprehensiveness, representativeness, and clarity of the crisis policy items. The authors then convened to discuss the discrepancies until reaching full agreement. The final list rendered a Cronbach’s alpha reliability coefficient of 0.91, consisting of 52 organizational crisis policy and procedure items which were then distributed to participants, asking them to indicate on a yes/no scale if each item was perceived to be implemented by their organization during the pandemic. This methodology follows that of Ramus and Steger (2000), who asked participants to agree or disagree that each policy had been enacted in their organization.

Proximal Perceptions

Perceptions of ethical leadership were assessed using Brown et al.’s (2005) 10-item scale, POS was measured with Eisenberger et al.’s (1986) nine-item scale, and perceived safety climate was evaluated with Huang et al.’s (2017) eight-item measure. Each scale was modified to include the stem, “during the COVID-19 crisis” and was phrased in past tense.

Distal Effects

Job satisfaction and turnover intentions were measured using three-item scales from the Michigan Organizational Assessment Questionnaire (Camman et al., 1979; Seashore et al., 1982). Stress was evaluated using Stanton et al.’s (2001) 15-item Stress in General scale, and safety behavior was measured with Griffin and Neal’s (2000) eight-item measure. Each of the distal effect scale items (with the exception of the following turnover intentions item: “I will probably look for a new job during the next year”) asked participants to respond in reference to the past 30 days and used past tense phrasing.

Demographic Variables and Covariates

Participants reported basic demographic information (i.e., age, gender, education level, and race), completed Boekhorst et al.’s (2017) five-item measure of work intensity, and reported average hours worked per week over the past 30 days, number of children in their household, job title, healthcare facility type, career and job tenure, and whether they directly interact with patients. Including a large number of variables in analyses is a common, recommended tactic when using an exploratory approach (Spector, 2017). As such, these measures were included as covariates either because employee work experiences were likely to be biased on the basis of the variable (i.e., job title) or because it had been shown to impact at least one outcome variable.

More specifically, job title was included as a covariate as non-medical staff are likely to have differential relationships with certain outcome variables (e.g., perceived safety climate), as their jobs do not entail interacting with infected patients unlike the medical staff. Further, research has demonstrated that all remaining explored covariates (i.e., age, gender, education level, race, work intensity, work hours, number of children, healthcare facility type, career and job tenure, direct interaction with patients) explain considerable variance in at least one outcome variable of interest (Besen et al., 2013; Carless & Arnup, 2011; Duffy et al., 1998; Grunfeld et al., 2005; Hersch & Xiao, 2016; Iranmanesh et al., 2017; Ozyurt et al., 2006). For instance, age has been shown to positively impact job satisfaction (Besen et al., 2013; Carless & Arnup, 2011), whereas job satisfaction has been shown to be negatively impacted by important considerations such as work intensity (Iranmanesh et al., 2017). As such, a failure to account for a wide array of influential factors may impact the ability to detect meaningful relationships between the variables of interest in this study.

Results

Confirmatory Factor Analysis

As strong correlations between proximal perceptions were demonstrated (see Table 1 for variable descriptive statistics and intercorrelations), we tested a series of confirmatory factor analyses to demonstrate the distinctiveness of our measured variables following previous recommendations (Brown, 2006). We used the lavaan package (Rosseel, 2012) in R (R Core Team, 2020) and implemented maximum-likelihood estimation with robust standard errors. Model fit was assessed with the χ² statistic, the comparative fit index (CFI), the Tucker-Lewis index (TLI), the standardized root mean square residual (SRMR), and the root mean square error of approximation (RMSEA). Hu and Bentler’s (1999) recommended cut-offs were used to evaluate model fit, with CFI and TLI exceeding 0.90, and SRMR and RMSEA below 0.08. The seven-factor model including the mediators and outcomes was compared with several alternative models by setting correlations between different combinations of latent factors equal to one; identified and alternative models were nested. We found that the seven-factor model demonstrated the best fit to the data compared to alternative models (Δχ2(1506) = 3222.35, p < 0.001, CFI = 0.92, TLI = 0.92, RMSEA = 0.05 [90% CI [0.05, 0.05]], SRMR = 0.06; see Appendix 2), therefore demonstrating that our constructs were distinct.

Latent Profile Analysis Results

A three-step approach (Asparouhov & Muthén, 2013) was used to examine the effects of perceived crisis response strategy on employee perceptions and distal outcomes. In the first step, we conducted LPA to identify unique strategy profiles. Next, profile assignments were made for each participant based on the posterior distribution of profile membership. In the third step, we used a regression-based approach to examine whether perceived strategy profiles predicted distal outcomes via proximal perceptions. Additionally, the impact of individual policies and procedures was assessed using relative weights analysis.

LPA was conducted using Mplus version 8.4 (Muthén & Muthén, 2019-2020). We considered the types of policies and procedures enacted, the number of individuals within each profile, and model fit to determine the number of meaningful perceived strategy profiles. Fit was assessed using the six indices recommended by Morin et al. (2017), and the Lo-Mendell-Rubin likelihood ratio test (LMR-LRT) and the bootstrapped likelihood ratio test (BLRT) statistics were used to compare adjacent models. In general, lower values of AIC, BIC, and SSA-BIC and higher Entropy values indicating greater profile separation are better (Lubke & Muthén, 2007). In comparing adjacent models, we used the LMR-LRT and BLRT statistics (Morin et al., 2017). Generally, a significant p value means that the examined model has a better fit than an adjacent model with one less profile. As SSA–BIC has been identified as the most accurate of these indices for LPA (Nylund et al., 2007), we prioritized SSA-BIC when interpreting model fit. Model fit indices were best for the five-profile model as it produced the lower SSA–BIC and BIC and as well as a significant BLRT and LMR-LRT (see Table 2). Although the five-profile model produced lower AIC values, one profile in this model was only a slight variation of one profile in the five-profile model and it had higher SSA–BIC and BIC. Thereby, based on those fit indices as well as parsimony and interpretability, we selected the four-profile model.

Table 2.

Fit statistics for profile structures

# of profiles	LL	AIC	BIC	SSA–BIC	Entropy	BLRT	LMR_LR
2	− 2242	4518	4588.01	4534.05	0.83	665.58*	621.26*
3	− 2127.19	4306.39	4413.46	4330.94	0.82	229.61*	214.32*
4	− 2077.27	4224.53	4368.66	4257.59	0.83	99.86*	93.21*
5	− 2065.17	4218.33	4399.53	4259.89	0.85	24.2	22.59*

LL log-likelihood, AIC Akaike information criteria, BIC Bayesian information criteria, SSA–BIC sample-size-adjusted BIC, BLRT bootstrapped log-likelihood ratio tests, LMR_LR Lo-Mendell-Rubin likelihood ratio test. *p < .05

Average values for each of the policy and procedure categories across the four strategy profiles in this model appear in Table 3 and Fig. 2, which reveal the pattern of results across categories, therefore aiding in the interpretation and naming of each profile. For example, we labeled the first profile (reflecting 12% of the data) as an inactive response, as average values for all four dimensions of policies and procedures were low. We labeled the second profile (reflecting 39% of the data) as a human resource-disadvantaging response due to its highest score being for human resource-disadvantaging policies and procedures and its low scores for the other three categories. Some specific policies that characterize the human-disadvantaging response likely include introducing temporary pay decreases and expanding required work duties.

Table 3.

Descriptive information per latent profile

Profiles	% of sample	Human resource-supportive		Human resource-disadvantaging		Behavioral/interactional human safety and protection-focused		Environmental and structural safety supports-focused
		M	SE	M	SE	M	SE	M	SE
Inactive response	12	− 1.66	0.06	− 1.03	0.09	− 1.39	0.22	− 1.46	0.23
Human resource-disadvantaging response	39	− 0.39	0.09	0.18	0.1	-0.38	0.1	-0.36	0.08
Safety and human resource-supportive response	15	0.43	0.16	− 0.54	0.14	0.24	0.1	0.62	0.11
Maximizing response	34	0.83	0.1	0.34	0.13	0.81	0.07	0.64	0.08

Mean scores on each of the four dimensions have been standardized to demonstrate the relative relationships among these dimensions

Fig. 2.

Perceived crisis response profile average latent profile values across policy/procedure categories. The four perceived crisis response policy categories consisted of 17 items (human-resource supportive), 7 items (human-resource disadvantaging), 9 items (behavioral/interactional human safety and protection-focused), and 19 items (environmental and structural safety supports-focused). The inactive response was implemented by 12% of the sample, the human resource-disadvantaging response was implemented by 39%, the safety and human resource-supportive response was implemented by 15%, and the maximizing response was implemented by 34%. The standardized values provided on the vertical axis represent the level of each policy category within each profile. Given that four types of policy responses have different ranges of values, we standardized them before performing LPA for ease of interpretation and presentation. The use of standardized versus unstandardized values results in no differences in profile identification

We labeled the third profile (reflecting 15% of the data) as a safety and human resource-supportive response, as this strategy focused on more positive policies and procedures with less reliance on human resource-disadvantaging actions, such as introducing additional mental health benefits for employees and providing bonus pay. The fourth profile (reflecting 34% of the data) was termed a maximizing response, as these organizations employed policies and procedures from all four categories. Taken together, in answering RQ1, the data indicated that there are four distinguishable strategy profiles that typify the perceived crisis response styles enacted by healthcare organizations during COVID-19, which aligned with the aforementioned profile expectations.

Regression Results on Proximal Perceptions

To investigate H1, which expected that perceived crisis response strategies would predict proximal perceptions, PROCESS model 4 (Hayes, 2018) was used with 5000 bias-corrected bootstrapping estimates and a custom seed to ensure estimate stability across models. More specifically, four models were run, each testing a different dependent variable but including the same three mediators, perceived strategy profile as a multicategorical independent variable, and covariates. Relationships between proximal factors were accounted for, as all predictors, including the three mediators, were simultaneously regressed on the distal outcomes. All continuous variables were standardized, and the inactive response strategy profile was set as the referent class in response strategy comparisons. Following Lum et al.’s (1998) methodology, only significant demographic variables and covariates were retained in the final models. Each of these retained factors was correlated with at least one mediator or dependent variable, following Becker’s (2005) recommendation, thereby removing impotent control variables (i.e., unrelated to any outcome variable), which diminish power; Becker, 2005).1

First examining relations between perceived crisis response strategy and proximal perceptions, as demonstrated in Table 4, employees within organizations that were perceived to employ human resource-disadvantaging, safety and human resource-supportive, or maximizing responses reported higher perceptions of ethical leadership and safety climate compared to organizations that were perceived to be inactive. Likewise, organizational support perceptions were higher within organizations that were perceived to employ safety and human resource-supportive or maximizing responses than within organizations with perceived inactive responses.

Table 4.

Regression results for proximal perceptions

Variable	Perceived ethical leadership		Perceived organizational support		Perceived safety climate
	B	SE	B	SE	B	SE
Covariates
Age	.04	.04	.00	.04	− .07	.04
Work hours	.03	.04	.00	.04	.04	.04
Work intensity	− .23**	.04	− .29**	.04	− .19**	.04
Job title	.28*	.09	.25**	.09	.30**	.09
Crisis response strategies
Human resource-disadvantaging response	.36**	.14	.26	.14	.68**	.14
Safety and human resource-supportive response	1.05**	.16	.85**	.16	1.13**	.17
Maximizing response	1.12**	.14	.93**	.14	1.21**	.14
	R2 = .26**		R2 = .24**		R2 = .21**
	F = 22.24		F = 19.46		F = 16.82

Crisis responses strategy variables were dummy coded with the inactive crisis response profile serving as the reference group. Job title was coded: 0 = medical staff, 1 = non-medical staff. Non-significant control variables such as career tenure, patient interaction, and work facility type were removed from analyses to follow Lum et al.’s (1998) methodology of only retaining significant covariates. *p < .05; **p < .01

To examine whether other effects differed across active strategy profiles, a series of ANOVAs with Bonferroni-adjusted pairwise comparisons were conducted. Results indicated that effects in each of the three models differed (p < 0.01) across the human resource-disadvantaging and safety and human resource-supportive responses, but not between the safety and human resource-supportive and maximizing responses. Notably, although the safety and human resource-supportive and maximizing response strategy profiles demonstrated similar patterns of results, differences between these profiles were deemed non-negligible based on the BLRT results that suggested meaningful differences were present, as well as the fact that human resource-disadvantaging policy and procedure reliance differed across strategies (i.e., one profile had above-average and one below-average reliance on such policies and procedures).

These findings provide full support for H1, as the utilization of a safety and human resource-supportive response strategy and enacting a wide array of crisis response policies and procedures was positively related to perceptions of ethical leadership, organizational support, and safety climate. The regression results for distal effects are provided in Appendix 3.

Mediation Effects

To test H2, which expected that perceived crisis response strategy would impact distal outcomes via proximal perceptions, all three mediators were included simultaneously in each of the four models run to test each dependent variable. Results indicated that perceived ethical leadership mediated all three effects of perceived crisis response strategy variables on safety behavior (see Table 5). Likewise, POS served as a mediator in relations between both the safety and human resource-supportive and maximizing perceived crisis response strategies on all dependent variables. Perceived safety climate mediated each of the effects between perceived crisis response strategies and the four distal outcomes. All other effects were non-significant, thereby providing partial support for H2. As limited direct effects between perceived response strategies and distal outcome variables emerged (i.e., the only direct effects were between perceived crisis response strategies and work stress), the indirect effects highlight the critical role that proximal perceptions played in linking perceived response strategies and distal outcomes.

Table 5.

Indirect effects of perceived crisis response strategy on distal employee outcomes

Outcome	Mediator	Strategy	Indirect effect	Boot SE	LL95%CI	UL95%CI
Job satisfaction	Perceived ethical leadership	HR-disadvantaging	− .03	.04	− .11	.03
		Safety and HR-supportive	− .10	.09	− .28	.07
		Maximizing	− .10	.09	− .29	.08
	Perceived organizational support	HR-disadvantaging	.10	.06	− .01	.23
		Safety and HR-supportive	.32*	.10	.15	.53
		Maximizing	.35*	.11	.17	.57
	Perceived safety climate	HR-disadvantaging	.21*	.07	.09	.36
		Safety and HR-supportive	.35*	.09	.18	.54
		Maximizing	.37*	.10	.20	.57
Work stress	Perceived ethical leadership	HR-disadvantaging	.03	.03	− .02	.10
		Safety and HR-supportive	.09	.08	− .07	.24
		Maximizing	.09	.09	− .08	.25
	Perceived organizational support	HR-disadvantaging	− .06	.04	− .16	.01
		Safety and HR-supportive	− .20*	.08	− .37	− .05
		Maximizing	− .21*	.09	− .40	− .06
	Perceived safety climate	HR-disadvantaging	− .09*	.05	− .20	− .01
		Safety and HR-supportive	− .15*	.07	− .31	− .02
		Maximizing	− .16*	.08	− .32	− .02
Turnover intentions	Perceived ethical leadership	HR-disadvantaging	.04	.04	− .02	.14
		Safety and HR-supportive	.11	.10	− .07	.31
		Maximizing	.12	.10	− .08	.33
	Perceived organizational support	HR-disadvantaging	− .07	.05	− .19	.01
		Safety and HR-supportive	− .22*	.10	− .42	− .05
		Maximizing	− .24*	.10	− .46	− .05
	Perceived safety climate	HR-disadvantaging	− .24*	.08	− .40	− .11
		Safety and HR-supportive	− .40*	.10	− .61	− .23
		Maximizing	-.43*	.10	− .64	− .25
Safety behavior	Perceived ethical leadership	HR-disadvantaging	.10*	.05	.01	.23
		Safety and HR-supportive	.30*	.10	.12	.51
		Maximizing	.32*	.11	.13	.55
	Perceived organizational support	HR-disadvantaging	− .05	.04	− .15	.01
		Safety and HR-supportive	− .17*	.09	− .36	− .01
		Maximizing	− .18*	.10	− .38	− .01
	Perceived safety climate	HR-disadvantaging	.16*	.07	.05	.30
		Safety and HR-supportive	.27*	.09	.10	.46
		Maximizing	.29*	.10	.10	.49

Crisis responses strategy variables were dummy coded with the inactive crisis response profile serving as the reference group. The three mediators were included simultaneously in each of the four models run, each testing a separate dependent variable. *p < .05

Relative Weights Analysis Results

To address RQ2, which questioned how specific perceived crisis response policies and procedures are linked to employee proximal perceptions and distal outcomes, multivariate relative weights analyses were conducted using the procedures recommended by Tonidandel and LeBreton (2011; see Appendix 1). Findings suggested that providing remote work options and implementing programs focused on employee morale may be particularly impactful human resource-supportive responses, whereas implementing employee travel restrictions and expanding required work duties were the most detrimental human resource-disadvantaging policies and procedures. Among behavioral/interactional human safety and protection-focused policies and procedures, implementing a COVID-19 prevention training program and utilizing remote communication between patients and staff were the most beneficial, and implementing a PPE and hygiene auditing process for personnel and frequently sanitizing equipment and clinical areas were the most beneficial environmental and structural safety supports-focused responses.

Discussion

This study investigated perceived strategic responses employed during the COVID-19 pandemic, as well as the impact of these responses on employee perceptions and work outcomes in order to provide information needed for rapid organizational action in future crises. Four distinct profiles of perceived crisis response strategies were identified, with the most prevalent profile being a human resource-disadvantaging response. These findings align with reports of actions taken by healthcare organizations during this study’s data collection period, such as widespread reductions of work hours and instituting furloughs (Paavola, 2020). As such, it is not surprising that employees perceived a disadvantaging response to be the most frequently implemented strategy, particularly when considering the severe financial pressure burdening organizations which required actions that were perceived as disadvantageous.

Notably, 35 of the 52 policies examined were enacted in over 50% of organizations, with those in the environmental and structural safety supports-focused category being implemented most frequently (see Appendix 1). However, some of the responses most likely to impact employee well-being and behavior were not as common as other less impactful policies and procedures. For instance, only 51.3% of respondents indicated that their employer implemented employee and patient COVID-19 prevention and control training, despite it being the most beneficial environmental and structural safety supports-focused procedure. Yet, 86.3% of respondents reported that their employer instituted visitor restrictions in patient areas, which was an important safety measure, but also one of the least impactful behavioral/interactional human safety and protection-focused responses. Such mismatches suggest that the prevalence of perceived policy implementation may not always translate directly into impacts on employee attitudes and behaviors.

In comparing the three active perceived crisis response strategies, results indicated that both the safety and human resource-supportive response and maximizing response profiles were linked to favorable proximal employee perceptions, with no meaningful differences between the two profiles. This suggests that perceived crisis response strategies are likely to have the greatest positive impact on employees when they include safety and positive human resource policies and procedures. These strategies resulted in more favorable outcomes than the human resource-disadvantaging approach, demonstrating that utilization of these positively perceived actions may offset the negative ramifications of enacting disadvantageous policies.

Notably, a few unexpected findings emerged, such as the positive link between the human resource-disadvantaging response and both ethical leadership and safety climate perceptions. As the data collection period was characterized by reports of hospital overcrowding and shortages (Bellware et al., 2020), participants were likely aware of their organization’s financial challenges, therefore perhaps perceived policies such as pay cuts and required overtime hours to be ethical and safe actions. Further, a negative relationship was found between POS and safety behavior, which may be similarly explained by the environment at the time in which organizations were perceived to provide support but were hindered by external factors such as PPE shortages, therefore POS may not have positively impacted safety behavior due to a lack of resources. Ethical leadership emerged as a mediator in the links between perceived crisis response strategies and safety behavior, but unexpectedly did not mediate relations with the other three distal outcomes, perhaps due to multicollinearity among the proximal variables.

Theoretical Implications

Several indirect effects demonstrated that greater perceived implementation of policies and procedures related to safety and human resource considerations was generally linked to more positive POS and perceived safety climate, which in turn had advantageous links to both employee well-being (i.e., job satisfaction and work stress) and behavioral outcomes (i.e., turnover intentions and safety behavior). These findings are consistent with the tenets of social exchange theory, as well as empirical work demonstrating that receiving support and beneficial treatment from organizations is likely to produce favorable organizational perceptions, improved well-being, and positive job behaviors (Caesens et al., 2017; Eisenberger et al., 1986, 1997; Kurtessis et al., 2017; Rhoades & Eisenberger, 2002).

Likewise, indirect effects in which perceived safety climate served as an intermediary are in line with extant workplace safety models (see, e.g., Beus et al., 2016) and empirical work (DeJoy et al., 2004), demonstrating the importance of safety policies and perceived safety climate in increasing safety behavior. Taken together, in line with theoretical expectations, the indirect effects demonstrated in this study extend these theories by highlighting how perceived crisis response strategies play a key role in the social exchanges between employees and organizations by providing cues that employees use to evaluate leader characteristics, organizational priorities, and work environment attributes.

Practical Implications

This study has several practical implications. Namely, developing a clearer understanding of approaches taken by organizations is beneficial, as policy perceptions were shown to be predictive of crucial employee judgments of their organization (e.g., POS), which, in turn, were found to predict worker attitudes and behaviors (e.g., turnover intent, safety behavior). This finding is paramount during this pivotal time in which healthcare worker safety is at risk and 22% of nurses have reported planning to leave their jobs in the next year (Berlin et al., 2021). As such, organizations could use such findings to implement policies that have been linked to key outcomes in an attempt to avoid negative worker attitudes and behaviors.

Further, as the perceptions employees form about the organization based on enacted policies play a crucial role in determining outcomes, organizations should clearly communicate new policies and the rationale behind them to employees in an attempt to encourage positive perceptions, thus potentially decreasing the likelihood of adverse outcomes. In addition, employees may be particularly averse to crisis management policies and procedures in which they perceive a loss, particularly within their personal lives (e.g., employee travel restrictions). Research demonstrating that negative phenomena can have a stronger impact than positive stimuli is robust (Baumeister et al., 2001). Yet, our findings indicate that negative salience effects in a crisis scenario may be softened by simultaneously implementing positively perceived policies and procedures.

Thus, in instances where policies that have been demonstrated to be detrimental to employee perceptions of social exchange with the organization are deemed necessary, implementation of a variety of resource-gain or safety-focused efforts may be particularly warranted to promote more balanced perceptions of the employee-organization social exchange relationships. Taken together, as organizational decision-making is often a complex process in which multiple considerations must be weighed (e.g., employee well-being, available resources), our findings can serve a prescriptive purpose by serving as a resource to help organizations make better decisions regarding the impact of perceived crisis response policies on employee well-being and behavior.

Limitations and Future Research Directions

Although this study has a number of strengths, we also acknowledge limitations. First, all data were collected from the same source, which increases the possibility of common method bias (Podsakoff et al., 2012). To mitigate this effect, we conducted time-lagged assessments of employee outcomes, which also allowed for a better understanding of effect directionality (Cole & Maxwell, 2003). However, given that policies and proximal perceptions were collected at the same time, temporal precedence is not established in the present effort; thus, causality cannot be inferred from our results. Future research could collect longitudinal data and examine the causal relationships among variables.

Further, given that our data consists of only two time points, future research could collect multi-wave longitudinal outcome data and examine how the profiles established in this effort may change or influence changes in various outcomes over an extended period of time. Relatedly, as our two waves of data collection were conducted only 30 days apart, more elapsed time may be necessary for enacted policies to impact employee outcomes. However, as there is evidence of healthcare organizations enacting policies as early as March 2020 (Grimm, 2020), it is possible that policies had been enacted for some time before we collected data on distal outcomes.

Additionally, it is possible that our policy measure may not have encapsulated all possible actions, especially uncommon approaches, thereby warranting more research on this topic. Further, to maintain participant confidentiality, respondents did not identify their employing organization, which precluded the examination of the collective perceptions and intraorganizational variability in reports of crisis response strategies at the organization level. Future work could collect multilevel data to investigate how shared perceptions (e.g., team safety climate perceptions) and perception variability of crisis response strategies influence organizational outcomes and exert trickle-down effects on individual outcomes.

This study focused on a single industry (i.e., healthcare), which could limit generalizability. However, as crises are especially common in healthcare organizations (Nemeth et al., 2011) and given the nature of the COVID-19 crisis, our focus on healthcare organizations seemed particularly necessary. As the policies examined in this effort were implemented in a variety of work sectors, such as introducing furloughs and pay cuts, we believe these results are likely to generalize to a number of work contexts outside of healthcare organizations.

As previous research has demonstrated that human resource practices also have the potential to impact customer satisfaction via their impact on the work climate (Rogg et al., 2001), future research should also examine the impact of perceived crisis response strategies on outcomes relevant to both patients/customers (e.g., satisfaction, trust) and potential employees (e.g., organizational attraction). In addition, examination of other relevant proximal perceptions (e.g., ethical climate perceptions) and distal outcomes (e.g., organizational commitment), as well as parallel effects (e.g., ethicality perceptions at both the leader and organizational level) would provide additional insight into who employees perceive as responsible for crisis responses.

Additional avenues include investigating potential moderators on the demonstrated relationships. For instance, culture may moderate the effects of crisis strategies on outcomes, with strategies employed in tight cultures (i.e., those with strict norms and a resistance to deviance) demonstrating stronger effects than loose cultures (i.e., those with flexible norms and receptive to deviance; Gelfand et al., 2011). In turn, team-member exchange or team trust could serve as moderators, with crisis strategies rendering stronger effects on outcomes when those factors are high. Thus, cultural context and team dynamics may influence crisis management strategy effectiveness.

In sum, this study sheds light on how organizations responded to the COVID-19 crisis, as well as how perceived response strategies during a large-scale crisis can impact employee outcomes. Thus, this paper not only contributes to the limited research on organizational crisis management impacts by highlighting the process by which employee attitudes and behaviors may be affected by perceived crisis response strategies, but it also provides prescriptive information that can aid organizational decision-makers.

Appendix 1

Table 6.

Multivariate relative weights analyses and prevalence of each organizational crisis policy/procedure

Category	Organizational crisis response	R 2	%R 2	Rank	Prevalence
Human resource-supportive	Promoted programs focused on employee morale (e.g., providing free meals)	0.010	11.6	2	53.1%
	Established a COVID-19 hotline for staff questions	0.007	8.7	3	59.3%
	Offered free COVID-19 testing for employees and/or employee family members	0.007	7.9	4	50.1%
	Provided PTO “loans”	0.006	7.5	5	26.7%
	Provided additional employee mental health benefits (e.g., counseling, stress prevention programs)	0.006	7.4	6	61.7%
	Implemented special policies for “high risk” employees (e.g., transfer to non-COVID unit, additional PTO)	0.005	6.3	7	47.2%
	Increased telemedicine benefits*	0.005	6.2	8	61.9%
	Implemented non-punitive sick leave policies that allow sick personnel to stay home	0.005	5.8	9	71.1%
	Provided additional childcare support/financial assistance	0.004	4.8	10	38.1%
	Implemented PTO for infected/symptomatic employees	0.004	4.2	11	65.4%
	Provided non-critical employees an option to perform roles/tasks outside of normal duties (e.g., job rotation)	0.003	4.1	12	63.1%
	Implemented full or partial PTO for unworked/missed/canceled hours	0.003	3.6	13	47.8%
	Reactivated retired or separated employees	0.003	3.3	14	20.4%
	Covered hotel costs for quarantining	0.002	2.6	15	21.7%
	Introduced a temporary pay increase	0.002	2.1	16	11.0%
	Provided bonus pay (excludes temporary wage increases)	0.002	2.0	17	17.6%
	Total R2	0.09
Human resource-disadvantaging	Implemented employee travel restrictions	0.015	40.1	1	62.1%
	Expanded required work duties	0.005	13.7	2	51.1%
	Implemented required furloughs/unpaid leave	0.005	13.1	3	39.2%
	Eliminated overtime hours	0.004	11.3	4	32.2%
	Introduced a temporary pay decrease	0.003	9.3	5	16.3%
	Implemented required overtime hours	0.002	6.6	6	12.1%
	Reduced work hours	0.002	6.0	7	44.6%
	Total R2	0.04
Behavioral/interactional human safety and protection-focused	Implemented a training program for personnel, patients, and their families concerning COVID-19 prevention and control measures*	0.019	25.8	1	51.3%
	Implemented a process to allow for remote communications between the patients and staff (e.g., video call applications on phone or tablets)*	0.013	18.2	2	74.8%
	Developed a written protocol for identifying, monitoring, and reporting COVID-19 among hospitalized patients, volunteers, and staff*	0.008	10.3	3	74.1%
	Implemented regular employee COVID symptom/temperature screening*	0.007	10.1	4	82.2%
	Designated a care team to take care of COVID-19 symptomatic patients	0.007	9.8	5	62.5%
	Provided regular COVID-19 related updates (e.g., potential employee COVID-19 exposure)	0.007	9.3	6	83.3%
	Posted signs instructing visitors to leave if they have a fever or are showing symptoms of respiratory infection*	0.005	6.6	7	84.3%
	Implemented policies restricting all visitors in patient areas*	0.004	5.6	8	86.3%
	Implemented policies allowing visitors to enter COVID-19 patient areas when safety precautions have been taken (e.g., wearing PPE, no touching, log of entry and exit)*	0.003	4.4	9	43.8%
	Total R2	0.08
Environmental and structural safety supports-focused	Implemented an auditing process for ensuring that personnel adhere to PPE and hygiene guidelines*	0.013	12.2	1	50.2%
	Sanitized/cleaned equipment and clinical areas more frequently*	0.011	10.3	2	89.8%
	Developed a process for adhering to recommended infection prevention and control practices (e.g., providing hand sanitizer in every patient room)*	0.009	7.9	3	86.1%
	Reduced patient ratios in COVID-19 units*	0.008	7.5	4	58.4%
	Created a multidisciplinary team or committee to address COVID-19 issues (e.g., preparedness, ethical issues)	0.007	6.7	5	81.3%
	Provided training on ethical issues concerning how decisions will be made in the event that services require prioritization and allocation (e.g., decisions based on probability of survival)	0.006	5.4	6	41.6%
	Provided training on donning and doffing PPE	0.006	5.2	7	78.0%
	Developed a process for triage during the COVID-19 outbreak (e.g., supply PPE to patients showing respiratory symptoms, instructions on prioritization of patients)	0.005	5.0	8	82.3%
	Introduced a requirement to change clothes/shoes at work arrival and/or departure	0.005	4.4	9	20.0%
	Increased the frequency of air exchange*	0.005	4.4	10	37.3%
	Imposed strict limits on PPE use	0.005	4.3	11	67.5%
	Introduced a strategy for continuing to provide required non-COVID-19 care (e.g., chronic illnesses, emergency services)*	0.005	4.2	12	82.6%
	Established criteria for employee COVID-19 testing	0.004	4.0	13	76.7%
	Developed plans for shifting healthcare services away from the hospital (e.g., home care)*	0.004	3.9	14	54.5%
	Established a process for tracking and allocating PPE and equipment	0.004	3.8	15	71.4%
	Quarantined symptomatic patients to specific locations in the facility*	0.003	3.0	16	81.7%
	Re-used/re-sterilized PPE	0.003	2.8	17	76.4%
	Used methods of communication (e.g., signs) to notify incoming people about the presence of COVID-19 in the facility*	0.003	2.5	18	63.0%
	Made determinations of critical (e.g., intensive care unit) vs. non-critical (e.g., dental unit) employees	0.003	2.4	19	65.6%
	Total R2	0.11

R², or the raw relative weight, refers to raw variance explained. %R², or the rescaled relative weight, refers to the estimates of relative weight using the metric of percentage of predicted variance accounted by each policy. Rank refers to the relative importance of each item in predicting all proximal and distal outcomes simultaneously. Prevalence refers to the percentage of respondents who indicated that the stated policy was enacted in their organization. *denotes policies oriented towards both employees and patients

Appendix 2

Table 7.

Confirmatory factor analyses

Model	χ 2	df	χ2 diff	CFI	TLI	RMSEA[90%CI]	SRMR
Model 1: Seven factors	3222.35***	1506	-	0.92	0.92	0.05[0.05, 0.05]	0.06
Model 2: Six factor (perceived ethical leadership + perceived safety climate)	3903.32***	1507	680.98***	0.89	0.88	0.06[0.06, 0.06]	0.07
Model 3: Six factor (perceived organizational support + perceived safety climate)	3792.30***	1507	569.96***	0.89	0.89	0.06[0.06, 0.06]	0.07
Model 4: Six factor (perceived ethical leadership + perceived organizational support)	3596.49***	1507	374.15***	0.90	0.90	0.06[0.05, 0.06]	0.07
Model 5: Six factor (job satisfaction + safety behavior)	3932.87***	1507	710.52***	0.89	0.88	0.06[0.06, 0.06]	0.10
Model 6: Six factor (turnover intentions + safety behavior)	3575.12***	1507	352.78***	0.90	0.90	0.06[0.05, 0.06]	0.08

N = 454. ^***p < 0.001; POS = perceived organizational support; χ2 = Chi-square statistic; df = degrees of freedom; CFI = the comparative fit index; TLI = the Tucker-Lewis index; AIC = Akaike information criterion; BIC = Bayesian information criterion; RMSEA = the root mean square error of approximation; SRMR = the standardized root mean square residual. The seven factor model includes ethical leadership, organizational support, safety climate, job satisfaction, turnover intentions, work stress, and safety behavior as seven distinct factors, with the subsequent models testing the fit of a reduced number of factors by combining any two out of those seven variables

Appendix 3

Table 8.

Regression results for distal outcomes

Variable	Job satisfaction		Work stress		Turnover intentions		Safety behavior
	B	SE	B	SE	B	SE	B	SE
Covariates
Age	0.17**	0.04	-0.08*	0.04	-0.15**	0.04	0.24**	0.04
Work hours	0.05	0.04	0.09*	0.04	0.06	0.04	0.09	0.05
Work intensity	-0.15**	0.04	0.47**	0.04	0.12**	0.04	-0.01	0.05
Job title	-0.23*	0.08	0.04	0.08	0.16	0.09	0.00	0.09
Crisis response strategies
Human resource-disadvantaging response	-0.21	0.13	0.30*	0.13	0.12	0.14	-0.36*	0.15
Safety and human resource supportive-response	-0.21	0.15	0.34*	0.15	0.12	0.17	-0.30	0.18
Maximizing response	-0.19	0.14	0.38**	0.14	0.22	0.15	-0.31	0.16
Proximal perceptions
Ethical leadership	-0.09	0.08	0.08	0.08	0.11	0.08	0.28**	0.09
POS	0.38**	0.08	-0.23**	0.08	-0.25**	0.09	-0.20*	0.09
Safety climate	0.31**	0.06	-0.14*	0.06	-0.35**	0.07	0.24**	0.07
	R2 = 0.40**		R2 = 0.41**		R2 = 0.29**		R2 = 0.17**
	F = 29.34		F = 30.25		F = 17.63		F = 8.91

Crisis responses strategy variables were dummy coded with the inactive crisis response profile serving as the reference group. Job title was coded: 0 = medical staff, 1 = non-medical staff. * p < 0.05. ** p < 0.01

Data Availability

The datasets generated during and/or analyzed during the current study are available from the corresponding author upon request.

Declarations

Ethics Approval

The questionnaire and methodology for this study was approved by the Human Research Ethics Committee of The University of Texas at Arlington (Protocol #2020–0644).

Consent to Participate and Publish

Informed consent was obtained from all individual participants included in the study, indicating that participants consented to both participation in the study and publication of results without naming them as participants.

Conflict of Interest

The authors declare no conflict of interest.