Corps de l’article

According to the World Health Organization[2], there is a global shortage of health personnel, particularly nurses and midwives that accounts for more than 50% of current unmet health workforce needs. Nine million nurses and midwives would need to be recruited by 2030. Nurses play a crucial role in health promotion, disease prevention and the delivery of primary and community care, but their physical and mental health has been put to the test during the COVID-19 pandemic[3] (Agalar & Ozturk Engin, 2020).

In France in 2020, 430,000 caregivers worked in nursing homes for elderly people, which are experiencing recruitment difficulties and high absenteeism (DRESS, 2016). A recent study in the French context of nursing homes shows how much absenteeism is a central concern in the health sector (Roussillon Soyer et al., 2021a). The authors explain how current absenteeism encourages future absenteeism. Another French study called for rethinking the management of caregivers to encourage positive behavior at work and reduce absenteeism (Roussillon Soyer et al., 2019). It is urgent to consider this problem because nearly 600,000 people are living in 7,200 nursing homes in France for elderly people, known as EHPADs.. Out of 30,000 COVID deaths occurring in France during the onset of the virus during the first year of the pandemic, 15,000 occurred in EHPADs (i.e. 48.35% June, 2020)[4].

In the Canadian province of Québec, there is currently a shortage of 4,000 nurses in the health network[5]. In order to restore its supply of nurses, the government of Quebec is considering various incentives[6] (e.g., for taking evening or night work, committing to a year’s work, recommending an employee, committing to working in a residence for seniors) that would keep nurses in place and bring back those who have left the field. These incentives were shown a year later to be ineffective[7], with some nurses feeling insulted by them and with more nurses quitting their careers rather than successfully attracting more to come back. In our study, we explored whether societal gratitude, a social form of incentive, might be more effective in retaining nurses in the profession, especially during difficult times such as the COVID-19 pandemic. To shed light on the likelihood of success of such measures, this study examined how social motives influenced the emotional engagement, turnover intentions, and absenteeism of French EHPAD nurses during the pandemic.

Our research is based on self-determination theory (Gagné & Deci, 2005; Rigby & Ryan, 2018) because it takes into account both the intensity and the quality of motivation (Deci & Ryan, 1985). Furthermore in terms of quality of motivation, this theory specifies a construct of socially controlled motivation that can help interpret the power of societal gratitude. In the context of COVID during which societies relied heavily on health care professionals, this form of motivation might be a powerful lever to sustain caregivers’ engagement and retention.

This study is positioned to be in line with Budhwar & Cumming (2020) to stimulate productive thought on the effect of COVID-19 on Management and Organization in the context of a pandemic. Furthermore, our research is an extension of the work of Shu et al. (2018) on job satisfaction and Gara Bach Ouerdian et al. (2018) on rebranding and emotional engagement in the context of turnover intentions. To this end, we propose a new link between emotional engagement and socially controlled motivation in its effect on turnover intentions and absenteeism, which complements the Pupion et al. (2017) approach to engagement based on the theory of planned behaviour (Ajzen, 1991; Ajzen & Fishbein, 1977).

In the next sections, we present hypotheses based on self-determination theory and engagement theory, followed by an empirical test of these hypotheses. We end with a discussion of the theoretical and practical implications of the results.

Self-determination theory: work motivation

Our study develops predictions derived from the logic of self-determination theory (SDT, Deci & Ryan, 1985; Olafsen et al., 2017; Trépanier et al., 2015), a theory of human motivation that has been successfully applied in various fields, including management (Deci & Ryan, 1985; Ryan & Deci, 2017; Van den Broeck et al., 2021). SDT provides a conceptual framework that differentiates between types of motivation according to sources of energization (“why” people engage in behaviour) that reflect the degree to which people’s actions are self-determined (i.e., volitional versus pressured). At its core, SDT distinguishes between intrinsic motivation (i.e., doing something out of interest and enjoyment; e.g., solving health dilemmas in patients) and extrinsic motivation (i.e., doing something for an instrumental reason). Extrinsic motivation is further divided to account for different types of instrumentalities that reflect different levels of internalization (Deci et al., 2017). Identified regulation reflects extrinsic motivation based on personal values (e.g., “I work hard because I value the health of my patients”), while introjected regulation reflects extrinsic motivation based on the enhancement or maintenance of self-worth (e.g., “I would feel ashamed if I did not do my job properly”). External regulation reflects extrinsic motivation based on externally controlled rewards and punishments, which can be divided into social and material types (Gagné et al., 2015). Socially controlled motivation is defined as doing something to gain the approval or avoid the criticisms of other people (e.g., “My colleagues put pressure on me to pull my weight”), while material external regulation is defined as doing something to gain a valued material reward (e.g., money) or avoid a material punishment (e.g., “If I make a mistake, I’ll lose my job”).

Research based on this conceptualization of work motivation has revealed that intrinsic and identified extrinsic motivation, together often labelled as autonomous forms of motivation, are most strongly and positively associated with optimal functioning which involves performance, attitudinal and well-being outcomes (Van den Broeck et al., 2021) and also strongly negatively associated with ill-being and turnover intentions; in contrast, introjected regulation was most positively associated with feelings of obligation towards an employer, and external regulation (irrespective of social or material focus) contributed scant variance in work outcomes (Van den Broeck et al., 2021). Given that most research has shown that autonomous forms of motivation are “better” than controlled forms (e.g. Deci & Ryan, 2000; Ryan et al., 2021; Van den Broeck et al., 2021), most research to date has concentrated on finding ways to enhance autonomous rather than controlled motivation. However, this research has not focused on specific contexts, such as crises such as the pandemic, where it might be possible for more external sources of motivation to be necessary to keep caregivers involved. A study of relations between leadership styles and work motivation demonstrated that when organizations are in crisis, more controlling forms of leadership can have some beneficial effects on work motivation (Gagne et al., 2020). As such, there appears to be some good reasons to explore whether socially controlled motivation is important to the engagement and retention of caregivers during the pandemic. Moreover, in a literature review of caregiver motivation, Dombestein et al. (2020: 9) concluded that “studies found that controlled (extrinsic) motivated reasons for providing care as feeling forced or obligated to take care, led to diminished well-being among caregivers.” However, none of the studies has differentiated between social and material external regulation.

Fang et al. (2021) recent work in the COVID-19 context highlights the importance of social support for the mental health of caregivers. The authors explain that since caregivers are in contact with COVID-19 cases, family and friends may fear becoming infected, resulting in their withdrawal of social support. In the absence of this social support, caregivers tend to choose negative coping styles, such as self-deprecation and withdrawal, which affect their mental health (Fang et al., 2021). Therefore, special attention should be given to improving the social support provided to caregivers. The pressures on healthcare workers (Muller et al., 2020) from society at large (America, 2021), from their employer, their stressed out colleagues, and even the financial pressures of retaining one’s job in a weakened economy during COVID might make societal gratitude something to hold onto in desperate times to keep going.

Given the lack of research focusing on socially controlled motivation as a type of motivation and the particular work circumstances of healthcare workers during the COVID-19 pandemic in French nursing homes, the present study sets out to explore how social external regulation factors were related to healthcare workers’ emotional engagement, absenteeism and turnover intentions.

Healthcare workers’ engagement

Caring for patients during the COVID-19 pandemic has been reported to be highly stressful in many countries[8] and has forced workers to drastically change the content of, and procedures in, their work. They have had to also alter their relationships within and outside of their work. These aspects include an increase in workloads (caseloads, schedules) combined with more complex work procedures to ensure safety (wearing uncomfortable protective gear, more frequent disinfection of work areas and equipment, social distancing, etc.). This is compounded with the workers’ fear of contaminating themselves, other patients, and their own family, and the potential social stigma of being seen as being infected with the virus. Yet the public (and their employer) expected healthcare workers to continue to do their job, despite exhaustion and fear. The weakened state of the economy also put financial pressures on them, obliging them to continue working in healthcare despite feeling burned out. Given that burnout is at the opposite end of the spectrum from work engagement (Maslach et al., 2001), we focused on how socially controlled rewards might help avoid the turnover and absenteeism of nurses by maintaining their work engagement.

The construct of work engagement is defined as “the harnessing of an employee’s full self in terms of physical, cognitive and emotional energies to work role performances” (Rich et al., 2010: 617). It seemed really important to differentiate between the three dimensions of engagement in the present study. Ulrich (1997: 125) wrote in his seminal book Human Resources Champions that “Employee contribution becomes a critical business issue because in trying to produce more output with less employee input, companies have no choice but to try to engage not only the body but the mind and soul of every employee.” Although the work of healthcare employees involves the three manifestations listed above (Shuck & Rose, 2013), we focused particularly on emotional engagement as an explanatory mechanism for turnover intentions. Because nursing work is typically described as involving a lot of “emotional labour” (Truc et al., 2009), an increased focus on external sources of motivation during a time of crisis might affect how emotionally engaged nurses are toward their patients and influence their desire to remain in their current job despite the extreme circumstances. The emotional labour of healthcare employees, particularly in nursing homes where residents’ entire lives take place, is particularly high (e.g., working with terminally ill patients, communicating to grieving family members, etc.), and has been associated with high levels of turnover (Jones & Gates, 2007; Shuck & Rose, 2013; Vahey et al., 2004). Moreover, societal gratitude is more likely to affect the emotional engagement of nurses rather than their cognitive or physical engagement, given that gratitude is usually expressed by invoking positive emotions.

Though autonomous forms of motivation are generally more strongly associated with work engagement (positive) and turnover intentions (negative) than controlled forms of motivation (Van den Broeck et al., 2021), research has rarely if ever examined how the relative importance of different types of motivation may shift in a situation of crisis. In the healthcare context in question, we were particularly interested in the role that external social regulation played in the immediate promotion of emotional engagement, turnover intentions, and absenteeism. In France, messages of support for caregivers in the context of COVID-19 occurred frequently: Examples of societal support have included songs by known[9] or amateur singers[10], children’s drawings, specially prepared dishes cooked by the locals, and audible “window” applause every evening at 8 p.m. to express gratitude to the caregivers. Though these social rewards might motivate healthcare workers to redouble their efforts, they may also be perceived as manipulative means of keeping them at work, and lead to feelings of obligation and guilt for wanting to quit because of fatigue, stress and anxiety. Nonetheless, given the unique social pressures on healthcare workers that have arisen during the COVID-19 pandemic context, praise and criticism might take on new meaning for healthcare workers. Praise may help restore depleted emotional resources, while criticism may take on greater significance than it deserves to someone working in a highly stressful environment. The question then becomes whether social pressure during a crisis can help convey the importance of the work of healthcare workers and thereby enhance their emotional engagement or decrease it by putting pressure on them. Given that meta-analytic relations between external regulation and work engagement were weak but in the positive direction, we expect that in a situation of crisis, societal expectations and gratitude will translate into a stronger positive link. At the same time, in line with previous research work (Jones & Gates, 2007; Shuck & Rose, 2013; Vahey et al., 2004), it would make sense if more emotional engagement led to lower turnover intentions.

As such, for the given healthcare context of the COVID-19 pandemic we hypothesize that:

Hypothesis 1: Emotional engagement mediates a relationship between socially controlled motivation and turnover intentions; the relationship between socially controlled motivation and emotional engagement is positive, and the relationship between emotional engagement and turnover intentions is negative.

The interest of studying turnover intentions is to put in place preventive interventions. It is difficult to implement retention strategies for staff when they have already left the organization (Williams, 1990). However, all perceptual measures are vulnerable to distortion. Using an objective measurement is scientifically justified. Like turnover, absenteeism is often considered an indicator of organizational withdrawal caused by negative attitudes (Johns, 2001). It is perceived as “a renunciation of the engagement to be present assiduously at work” (Bouville, 2009: 7).

Caregivers’ absenteeism has aroused the interest of researchers for a long time (e.g. Cohen & Golan, 2007; Hackett & Bycio, 1996; Shamian, 2003; Vandenberghe et al., 2009). The recent qualitative study of Roussillon Soyer et al. (2021b: 1) within the framework of self-determination theory showed that “The prevailing absenteeism has a harmful spiral impact on nurses’ and nursing assistants’ attitudes and behaviours, and, ultimately, on the quality of care received by the patients”. It is because absenteeism within a unit frustrates motivation that it will lead to even more absenteeism. Engaged employees do not perceive work as exhausting and stressful but rather as positively challenging (Bakker & Demerouti, 2009). Thus, they may be less sensitive to stress-related health problems (Bakker & Demerouti, 2009) and therefore less absent. Engaged employees tend to lose track of time while working and enjoy performing their tasks, resulting in lower rates of absenteeism (Shantz & Alfes, 2015). Engaged employees rate their overall health status and their ability to work more positively (Hakanen et al., 2008), so that we can anticipate that they will be less absent. Schaufeli et al. (2009) highlight through motivational processes (job demands and resources theory) the link between absenteeism and work engagement but did not differentiate the different dimensions of the engagement and whether emotional engagement might be particularly important. As such, and in line with hypothesis 1, for the given healthcare context of the COVID-19 pandemic, we hypothesize that:

Hypothesis 2: Emotional engagement mediates a relationship between socially controlled motivation and absenteeism; the relationship between socially controlled motivation and emotional engagement is positive, and the relationship between emotional engagement and absenteeism is negative.

Method

Sample and procedure

We conducted the study in nursing homes for dependent elderly people (Établissements d’Hébergement pour les Personnes Agées Dépendantes; EHPADs) in France between February 3 and 28, 2020 just before the EHPADs were closed to the public. The first official emails from the Regional Health Agency dated February 28, 2020, recommended that it would be necessary for the confinement of the residents within the nursing homes so that family visits would be discouraged. During February, there was talk of COVID-19[11], with fear and incomprehension setting in across the country. At the time, there was no screening for COVID-19 and therefore nobody knew who was infected and who was not.

The study participants (N = 303) were caregivers in ten private nursing homes located in France. These 10 private nursing homes volunteered to participate in research funded by the Occitania Region. The goal of the research was to reduce the absenteeism of caregivers in nursing homes. The “Caisse Primaire d’Assurance Maladie” checked that the rate of absenteeism in these nursing homes was representative of the national level of absenteeism in nursing homes.

All caregivers, without exception, of these nursing homes were invited to complete the questionnaire. Prior to the start of the study, all caregivers were briefed about the overall process including the study goals, measures, and potential publication targets. All participants were ensured of the strict confidentiality of their responses, and no one else, other than the researchers, had access to the data. Following their approvals, all employees agreed to participate in the survey. A paper questionnaire was administered to collect data from caregivers at their workplace during working hours.

A total of 303 invited participants completed the survey in February 2020, representing a response rate of 95%. These employees were spread over the 10 different nursing homes. Of the respondents, 94% were female, and represented different categories of care giving staff: nurses (n = 39), healthcare assistants (HCAs; n = 118), caregivers acting as HCAs (n = 126), and other caregivers (e.g., psychologist, occupational therapist; n = 20). HCAs have a state level professional qualification, whereas caregivers acting as HCAs have no such qualification. The average age was 41.15 years (SD = 11.25) and the average tenure was 7.56 years (SD = 7.11). Each employee was compensated with a fixed salary and with variable pay consisting of various bonuses according to the nursing home (e.g.: bonus for working on weekends, bonus for no absences, bonus for supervisors, participation in the organization’s profit sharing scheme).

Following the survey, we obtained archival data on absenteeism from the human resource departments of the EHPADs for the months of March and April 2020. Therefore, our predictor and mediator, and one of our outcomes, were measured at the same time and another outcome at a later time, which partially helps take care of some common method and directional influence concerns with the design of the study (Spector, 2019).

Measures

The items to assess the variables in the survey used a 5-point Likert scale ranging from 1 (strongly disagree) to 5 (strongly agree), except for work motivation, which used a 7-point Likert scale, as per the original scale validation (Dawes, 2008; Gagné et al., 2015).

Work motivation. Items were adapted from Gagné et al. (2015) Multidimensional Work Motivation Scale (MWMS) and Blais et al. (1993): Two items measured socially controlled motivation: “Every day, I make efforts in my work to be appreciated by my superiors, colleagues, residents, family…” and “Every day, I make efforts in my work to gain the respect of others my superiors, colleagues, residents, family…” (α = .93). The composite reliability (Raykov, 1997) was 0.928.

Emotional Engagement. Three of the 6-item emotional engagement subscale from Rich et al. (2010) were selected based on their semantic similarity. We translated all items into the French language consistent with the procedure used in Roussel (1996). The 3 items selected are as follow: “In my work, I am enthusiastic”; “In my work, I am proud of what I do”; “In my work, I am positive” (α = .78). The composite reliability (Raykov, 1997) was 0.770.

Turnover intentions. Four items from Wayne et al. (1997) were translated into French using Roussel (1996) procedure. Items are as follow: “I am seriously thinking of leaving my organization”; “As soon as I can find another organization, I will leave this one”; “I often think about leaving my work in this organization”; “I am actively looking for a job outside this organization” (α = .89). The composite reliability (Raykov, 1997) was 0.894.

Absenteeism. We obtained archival absenteeism data (in number of hours) during the COVID-19 lockdown period in March and April 2020 from all the nursing homes that participated in February.

Controls. We controlled for age (from 19 to 68 years old), gender (male or female), tenure (from zero years to 36 years), the name of the EHPAD (belonging to an EHPAD), and the occupational category of 303 employees (care profession: 118 nursing assistant, 126 acting as nursing assistant (does the job but does not have the diploma), 39 nurse, and 20 others such as psychologist, psychometrician).

These variables were measured by self-reported declarations by these different categories of caregivers: your age, etc. Age, gender, tenure, the name of the EHPAD (belonging to an EHPAD, and occupational category of employees were integrated into the study as potential control variables. Following the procedure proposed by Becker (2005), we examined the correlations between the control variables and the variables of our model to determine which control variables should be included in our analyses. Only the “age” and “gender” variables were significantly correlated with several variables in our model (see Table 1). Age was negatively associated with socially controlled motivation (r= -.16**), while gender was negatively associated with socially controlled motivation (r= -.14**). These variables were therefore integrated as predictors of absenteeism and turnover intentions in the final model. However, links between the variables in our model remained significant and identical to the model that did not include the control variables. Thus, as recommended by Becker (2005), the results are presented without the age and gender control variables.

Analyses and results

Confirmatory factor analyses

Prior to testing hypotheses, we conducted a series of confirmatory factor analyses (CFAs) to establish the convergent and discriminant validity of our variables. We assessed the fit of our data to a measurement model using Mplus 7.3 (Muthén & Muthén, 2017).

Since the data were collected with a single questionnaire, the problem of common method variance could represent a potential bias (Podsakoff et al., 2003). We performed Harman’s test (Podsakoff et al., 2003) to assess this potential bias. The measurement model (see Figure 1) included four factors (i.e., socially controlled motivation, emotional engagement, turnover intentions and absenteeism), yielded a good fit to the data (see Table 2). This model outperformed a three-factor model that merged socially controlled motivation and emotional engagement. It also outperformed a one-factor model that also tests for common method variance (Harman test; Podsakoff et al., 2003).

What’s more, according to this test, if a single factor emerges from the exploratory factor analysis or if a single factor accounts for more than 50% of the explained variance, the common method variance problem is present. The exploratory factor analysis carried out on all 10 items of the scales made it possible to extract four factors (socially controlled motivation, emotional engagement, turnover intentions and absenteeism), each having an eigenvalue greater than one. 71% of the variance is explained by the first three factors. The first factor alone accounts for 38% of the explained variance. Since several factors were restored by the exploratory factor analysis and none of these factors alone represents more than 50% of the explained variance, the absence of common variance is verified.

Therefore, these results support the distinctiveness of the four constructs in this study and assuages concerns around common method variance. Furthermore, all items loaded on their respective predicted factors, and the factor loadings were all acceptable and statistically significant.

Table 1

Confirmatory Factor Analysis of Measurement Models: Fit Indices

Confirmatory Factor Analysis of Measurement Models: Fit Indices

Note : N = 303; SRMR = standardized root-mean-square residual; CFI = comparative fit index; RMSEA = oot-mean-square error of approximation.

** p < .01, ***p ≤ .001

aThese models were estimated without the control variables because they do not change the significance and strength of the links (Becker, 2005; Bernerth & Aguinis, 2016)

-> Voir la liste des tableaux

Hypothesis testing

Table 2 summarizes the means, standard deviations, and correlations among the variables used to test our hypotheses. Of substantive interest, emotional engagement was correlated with hypothesized antecedents and outcomes in predicted directions. That is, respondents reported they were more emotionally engaged in their work when they also reported higher levels of socially controlled motivation. In addition, individuals reporting lower levels of turnover intentions or absenteeism reported higher levels of emotional engagement. We formally tested our hypotheses by specifying a series of structural equation models.

Table 2

Descriptive Statistics and Correlations for Key Study Variables

Descriptive Statistics and Correlations for Key Study Variables

Note : N = 303 * p < .05, **p < .01

-> Voir la liste des tableaux

To test the hypotheses, we specified a model in which socially controlled motivation predicted emotional engagement, which in turn predicted lower turnover intentions and lower absenteeism (see Figure 1). This model fit the data very well, χ 2 [32] = 48.07, p < .05, SRMR = .04, CFI = .98, RMSEA = .04.

Nonlinear bootstrapping was used with conditional process modeling to estimate the mechanism by which socially controlled motivation “influences” turnover intentions or absenteeism through emotional engagement, using of the PROCESS V3.0 macro “model 4” developed by Hayes (2017). The default data is represented by 95% confidence interval and the 5000 number of bootstrap samples. Results are presented in Table 3.

Figure 1

Measurement Model with Emotional Engagement as Mediator

Measurement Model with Emotional Engagement as Mediator

-> Voir la liste des figures

Table 3

Tests of indirect Relationships through Emotional Engagementa

Tests of indirect Relationships through Emotional Engagementa

Note : N = 303. All coefficients are standardized. 95% CI: bias-corrected confidence intervals.

aThese models run without the control variables because they do not change the significance and strength of the links (Becker, 2005; Bernerth & Aguinis, 2016)

-> Voir la liste des tableaux

In line with Hypothesis 1, respondents reported higher levels of emotional engagement when they reported higher levels of socially controlled motivation, β = .15, p < .05, R2 = 0.02, and reported lower levels of turnover intentions when they reported higher levels of emotional engagement, β = -.57, p < .01, R2 = 0.22. No significant direct relation was found between socially controlled motivation and turnover intentions when emotional engagement was included, β = .03, SE = .03, p = .19, 95% CI [-.02, .08], indicating complete mediation (see Figure 2). The indirect effect of socially controlled motivation was significantly related to turnover intentions through emotional engagement, β =—.063, SE = .02, p < .05, 95% CI [-.12, -.02].

In line with Hypothesis 2, the path estimates revealed that absenteeism was negatively related to emotional engagement, β = -17.95, SE = 5.51, p < .001, 95% CI [-28.79, -7.08], R2 = 0.04. Socially controlled motivation was related to emotional engagement, β = .05, SE = .02, p ≤ .01, 95% CI [.01, .09], and was not directly related to absenteeism, β = 3.49, SE = 1.97, p = .08, 95% CI [-.39, 7.36], indicating complete mediation (see figure 2). The indirect effect was significant β =—.025, SE = .01, p < .05, 95% CI [-.05, -.001] (see Table 3).

Figure 2

Structural Model, Hypothesis Testing

Structural Model, Hypothesis Testing

NS: non-significant

-> Voir la liste des figures

Discussion

A core question in motivation research concerns what drives employees to remain on the job and sustain their level of work engagement in a stressful work environment. This research focused particularly on social external forms of regulation that may take on more importance in such a context. The COVID-19 pandemic offered an opportunity to examine the external regulation dimension of motivation in a unique context, as pointed out recently by Budhwar & Cumming (2020). Socially controlled motivation was positively related to emotional engagement in healthcare work, which in turn was negatively related to turnover intentions and absenteeism. These findings draw attention to the significant influence of social forms of rewards in motivating healthcare workers during a crisis by exploring emerging phenomena that are not yet established (Farndale et al., 2020).

Because the data for this study were collected from nursing home establishments in France, the results must be interpreted within the French cultural environment. The three key identifying national cultural dimensions of Hofstede’s (2001) framework pertaining to the workplace in France are that it is high on the individualism, high on the power distance, and low on the masculinity dimensions (de Bruin, 2017; Soares et al., 2007). High individualism in the workplace indicates that French employees are more likely to give a higher priority to their own needs and those of their close friends and give lesser consideration for the needs of others such as the residents in a nursing home. High power distance in the workplace indicates that power in French establishments is more likely to be distributed unequally between managers and employees, with more hierarchical levels that separate them so that managers are less accessible to their subordinates. Low on masculinity in the workplace indicates that French employees value having a work-life balance with a good quality of life which means that they are less likely to work overtime hours because work should not interfere with their family life and personal activities. More research is needed to evaluate if our results would be replicated in cultures that are more collective, lower on power distance and higher in masculinity.

During the COVID-19 pandemic in this French work context, caregivers were more likely to have a propensity to behave individualistically and engage in absenteeism and voluntary turnover to protect their own health and that of family members from being infected with the virus compared to how caregivers would behave by giving more consideration to the needs of nursing home residents in a country where individualism is lower and collectivism is higher such as in Japan or Sweden (de Bruin, 2017; Soares et al., 2007). Furthermore, the caregivers were likely to experience higher levels of power distance with managers so that managers might have been less aware of the need to improve the working conditions of caregivers during a public health crisis in comparison to workplaces in lower power distance countries such as the United States where upward communication is less hindered by hierarchical barriers. Lower perspective taking from managers might have led to decreased engagement, and increased turnover and absenteeism. With high absenteeism in nursing homes during the pandemic French caregivers with relatively low levels on the masculinity cultural dimension might be less likely to work overtime to fill in for absent coworkers than would be expected from caregivers in countries with cultures displaying higher masculinity such as Japan and Germany.

Theoretical contributions

Results extend support for the relative contribution of the different forms of motivation proposed by SDT to highly stressful situations of crisis. Indeed, results highlight the unsuspected role of socially controlled motivation in times of crisis. This corroborates recent research demonstrating that different forms of leadership behaviors have differential effects on different types of motivation in crisis versus non-crisis organizational contexts (Gagné et al., 2019). More precisely, within organizations facing a crisis, individual perceptions of transactional leadership were related to decreased individual controlled motivation, while collective perceptions of transactional leadership were related to increased collective autonomous motivation, and decreased collective amotivation (Gagné et al., 2019). These results, along with ours, may indicate that in crisis situations, social rewards may be interpreted differently by employees, perhaps as more legitimate and even as serving a “guiding” function when the work context is more chaotic.

Our study extends the work of Pupion et al. (2017) on the motivational approach to engagement. Pupion et al. (2017) model a motivation-engagement process with regard to public service, in the light of the theory of planned behaviour (Ajzen, 1991). We extend this model by studying the consequences (turnover and absenteeism) of the link between motivation and engagement in the light of the self-determination theory.

We also adds breadth to the findings of Williams et al. (2014) within the framework of self-determination theory and it sheds additional light on ways to limit absenteeism and turnover intentions beyond those provided by Shu et al. (2018) and Gara Bach Ouerdian et al. (2018) thanks to our study of the socially controlled motivation facet. In a context of a public health crisis where every health worker counts, it is even more important not to neglect any potential lever to reduce the absenteeism of the healthcare staff in nursing homes. Until now, studies have focused only on autonomous motivation because it is an important lever. For example, Trépanier et al. (2015), Gagné et al. (2015), and Howard et al. (2017) revealed that autonomous forms of motivation, were most strongly associated with optimal functioning (i.e., performance, attitudinal and well-being outcomes), whereas external and introjected forms of extrinsic motivation, together referred to as controlled types of motivation, were less associated with positive outcomes. In the global shortage of highly skilled workers that results during a pandemic, we cannot afford to neglect the contribution of controlled types of motivation when it comes to retaining caregivers, as every factor that contributes to their retention counts in saving lives. Our study showed that societal gratitude, a form of socially controlled motivation, helped keep healthcare workers emotionally engaged in their work despite all the challenges they encountered, forestalling their turnover and absenteeism.

Practical contribution

The goal of the public in praising healthcare workers during the pandemic was to encourage and motivate them to work hard and not give up in almost unbearable working conditions. That social pressure could also be coming from nursing home administrators who, during the pandemic, have been struggling to staff their facilities appropriately to respond to the needs of residents. As long as this pressure comes with a good dose of gratitude and a focus on the benefits healthcare workers bring to residents, and as long as their health and safety is taken into consideration, at least in the short-term, that social pressure seems to enhance the emotional engagement of healthcare workers, and their desire to stay in their job and show up at work.

The results can help raise public awareness of the importance of encouraging the public and administrators to express gratitude to the caregivers for their contribution to society. For example, the administrators of healthcare organizations can ensure health and safety are promoted through good work design, voice, and adequate resources, which have all been shown to foster emotional engagement (Reina et al., 2018). In addition, caregivers’ families can encourage healthcare workers on the merits of their work, the community could express its support by singing optimistic songs out of open windows at a given time each day or bring them coffees (or even champagne!)[12], and journalists can enhance caregivers’ professional identity by reporting on stories for the media of their courageous work in nursing homes. The advantage of gratitude is that it is not costly to deliver.

Limitations and future directions

Our research is subject to a few limitations that suggest avenues for further investigation. Our survey respondents consisted of a homogenous population of French caregivers in a COVID-19 context in nursing homes, which limits our knowledge of how generalizable the results are to other countries or professions during the pandemic. Thus, replications of our study in other countries and professions would be welcome.

What would be interesting to examine is how long socially controlled motivation can sustain emotional engagement when the pandemic subsides. SDT predicts, consistent with behaviorist logic, that rewards and sanctions will operate only as long as they are present (Deci, 1971). SDT further explains that they would no longer work when removed because they do not promote the internalization of the value of the rewarded behaviour. Our findings suggest that emotional engagement, which might indicate in the present context the degree to which caregivers have internalized the value of their work for the well-being of the residents, is important to mitigate the effects of a stressful crisis on turnover intentions and absenteeism. If the social pressure on healthcare workers subsides post-pandemic, the question is whether it would decrease emotional engagement or even backlash into a dip lower than baseline levels, similar to the effect tangible expected rewards have been shown to have on intrinsic motivation (Deci et al., 1999). However, it is worth noting that meta-analytic evidence on the effects of tangible rewards on intrinsic motivation (Deci et al., 1999) did not generalize to praise (positive feedback, verbal rewards) (Hewett & Conway, 2016), so it could be that socially controlled motivation might not be as harmful to ongoing engagement.

The study was mostly cross-sectional, which could raise issues around direction of causality. However, the hypothesis we proposed rests on years of evidence on the role of motivation in predicting outcomes similar to the ones used herein, and the use of absenteeism data collected in the supplemental analysis can reassure us partially about the logic behind our hypotheses. Furthermore, a majority of studies in the literature that analyze a mediation process utilize cross-sectional designs (Goodman et al., 2017; Maxwell & Cole, 2007; Spector, 2019). To increase the temporal stability of the interpretation of our findings we conducted analysis with absenteeism (Spector, 2019), representing an alternative source of data collected in a later time period with findings that agreed with those of the cross-sectional analysis. The analysis with absenteeism using archival data reinforces our results that used turnover intentions as the dependent variable.

A suggestion for future research would be to examine the interaction between prosocial (Grant, 2008) and controlled social reward motivations. Prosocial motivation is close to identified regulation (one of the types of autonomous motivation) that represents doing something because it benefits others (i.e., a form of meaningfulness). It would be interesting to test whether the addition of social rewards and pressures onto existing prosocial motivation would increase or decrease emotional engagement towards residents (i.e., beneficiaries).

Conclusions

Our findings show how social motives through the external social regulation dimension of controlled motivation, is positively linked to emotional engagement, and the outcomes of reduced caregiver turnover intentions and absenteeism during the COVID-19 pandemic. Societal gratitude contributed to keep healthcare workers emotionally engaged in their work despite all the challenges they encountered, forestalling their turnover and absenteeism. Our study suggests that gratitude incites positive energy to both the giver and the receiver and consider it is important to encourage that individuals and organizations express gratitude to caregivers for their contribution to society.