Emotional labour in service work: Psychological flexibility and emotion regulation

Abstract

Drawing from job demands–resources theory (Demerouti et al., 2001), this article investigates the effect of psychological flexibility, relative to surface and deep acting, in the relationship between day-level emotional demands and exhaustion. A total of 170 not-for-profit service workers first filled in a questionnaire and then completed a diary survey over three consecutive workdays. The results of multilevel analyses suggest that person-level psychological flexibility was associated with lower levels of daily emotional exhaustion (measured at bedtime). Moreover, person-level psychological flexibility was found to attenuate, whereas person-level surface acting was found to strengthen, the association between day-level emotional demands and day-level exhaustion. Person-level deep acting had no significant effect on daily exhaustion. These findings extend previous research by demonstrating the role of psychological flexibility in encouraging employees to handle their emotions primarily by accepting them rather than actively regulating (i.e. suppressing or changing) them.

Keywords

deep acting emotional demands emotional exhaustion job demands–resources model psychological flexibility surface acting

Introduction

The service industry has become the largest employment sector in most developed countries (Lee and Wolpin, 2006). Because service roles involve frequent and intense interpersonal contacts with customers, service providers are often required to engage in emotional labour (EL). Defined as ‘the management of feeling to create a publicly observable facial and bodily display’ (Hochschild, 1983: 7), EL involves psychological processes necessary to manage emotions in accordance with organizational or professional display rules that serve as guidelines for the appropriate expression of emotions at work (Ashforth and Humphrey, 1993; Diefendorff and Gregarus, 2009; Hochschild, 1979). Put differently, emotional labour is the individual’s response to emotional job demands, defined as work characteristics that put the employee in emotionally stressful situations (Peeters et al., 2005), such as handling difficult customers.

It has been proposed that the relationship reported by others (e.g. Chebat and Kollias, 2000; Diefendorff and Richard, 2003) between emotional job demands and employee strain might vary as a function of different types of EL (e.g. Grandey, 2000). More specifically, the current literature tends to focus on two types of EL, namely surface acting and deep acting (e.g. Ashforth and Humphrey, 1993; Hochschild, 1983; Zapf, 2002). In surface acting, the employee suppresses his or her felt emotions and simulates emotional displays according to organizational directives. Deep acting, in contrast, is an effort to transform one’s felt emotions so as to actually experience the emotions one is required to display. Common to both types of EL is the need to regulate the emotions arising; that is, to suppress, transform or otherwise control thoughts and feelings (e.g. Bond and Flaxman, 2006). As a result, these types of EL are likely to be costly in terms of psychological effort, and so may make it more difficult to handle emotionally demanding job tasks (Muraven and Baumeister, 2000).

At the same time, there is scant empirical work on modes of EL that do not entail active emotion regulation and the energy and attention involved therein (Diefendorff et al., 2005; Larsen and Prizmic, 2004). To address this gap, we draw from job demands–resources theory (JD-R: Demerouti et al., 2001) to examine a cognitive-behavioural pattern, psychological flexibility, which involves the acceptance of emotions – even negative ones – while maintaining a focus on goal-oriented behaviours (e.g. Hayes et al., 1999). For example, imagine a bank teller facing an irate customer. Using surface acting, the teller might fake a smile and a calm demeanour while internally seething in frustration. In deep acting, the teller might employ trusted techniques (deep breathing, thinking positive thoughts) in an effort to release his frustration and actually feel calm (Grandey, 2000; Hochschild, 1979). A teller who is psychologically flexible, in contrast to both, accepts the emotions arising from the situation. He is mindful of his frustration, regarding it as an essentially automatic, idiosyncratic reaction to the circumstances. Not needing to divert his attention to suppressing or transforming his emotions, he can focus his energy on the concrete issue to be dealt with, rather than the emotional content of the interaction.

Between- and within-person variation

Testing for between-person variation, as is usually done in cross-sectional survey research, is important because it highlights how employees differ in their perceptions and behaviour. However, it is also important to take into account that working conditions – and employees’ reactions to them – may vary from day to day (Butler et al., 2005; Simbula, 2010). Employees who are usually energetic may have ‘off-days’, while employees who are exhausted can find themselves reinvigorated. Or, in the words of Sonnentag et al. (2010), ‘not all days are created equal’. Most previous studies have treated emotional job demands and exhaustion as relatively stable across time (Simbula, 2010), and so have measured only between-person differences in these variables. Examining these two variables at the within-person level (i.e. the day level) contributes to our understanding of work experiences as they unfold in the course of a working week (Ilies et al., 2007).

While EL may also vary within the same person and may be sensitive to training, individuals appear to use certain types of EL habitually (e.g. Brotheridge and Lee, 2002; Gross and John, 2003), making EL a more trait-like characteristic. We therefore use a daily (within-person) measure of emotional demands to predict daily emotional exhaustion, and trait-like (between-person) measures of EL as moderators in the demands–exhaustion association.

Emotional job demands and employee exhaustion

Burnout, a key concept in research on the health and well-being of service employees (e.g. Chebat and Kollias, 2000), can be defined in general terms as a syndrome of exhaustion, cynicism and reduced professional efficacy (Maslach et al., 2001). The importance of burnout is suggested by its relationship with such outcomes as absenteeism, decreased performance and physical/mental disorders (Taris, 2006). Emotional exhaustion, the dimension of burnout investigated here, involves feelings of fatigue, irritability and frustration, and a general sense of being worn out, as employees’ emotional resources become depleted (Maslach and Jackson, 1981). We selected emotional exhaustion as the outcome variable for this study because: (a) there appears to be general consensus in the literature that emotional exhaustion is the central dimension of burnout (e.g. Bakker et al., 2003); and (b) emotional exhaustion is regarded as the first stage of burnout, and therefore provides a critical point for intervention (e.g. Maslach, 2003).

The service industry offers the classic setting for the study of emotional exhaustion. Service employees work in the front line and act as boundary spanners, linking the organization and the public it serves. As a result, the nature of service work is demanding not only in terms of mental and physical effort (for instance, through imposed quotas and time pressure), but also in terms of emotional effort (e.g. Chebat and Kollias, 2000). That is, while all individuals experience emotions in reaction to various conditions of the organizational environment, for service workers, dealing with emotional issues and stimuli (e.g. customer complaints, ambiguous expectations, verbal aggression) is a requirement of the job (Grandey et al., 2005; Zapf, 2002). With such demands often resulting in emotional exhaustion (e.g. Deery et al., 2002; Grandey et al., 2007), it is not surprising that service jobs are often graded among the most stressful jobs in modern countries (e.g. D’Ausilio, 1997). Nevertheless, while most prior research has convincingly shown the relationship between demands and exhaustion from a between-person perspective, knowledge on how day-to-day fluctuations in these variables are related remains limited. In their recent study, Judge et al. (2009) call for more research on this question (p. 82). In that our model considers effects of day-level demands on affective outcomes, we seek to address this call, and propose:

Hypothesis 1: Day-level emotional job demands are positively associated with day-level emotional exhaustion.

Emotional labour and employee exhaustion

Emotional labour is intended to facilitate the attainment of organizational goals by contributing to a positive service experience for the customer (e.g. Diefendorff and Richard, 2003), and by allowing service providers to detach themselves cognitively from negative emotions so that they may maintain their objectivity and emotional equilibrium. Referring to the functionality of EL, Thayer et al. (1994) suggest that EL allows for the flexible adaptation of the organism to changing environmental demands. Indeed, previous studies have found EL to be positively associated with well-being and task effectiveness (e.g. Ashforth and Humphrey, 1993).

Other studies (e.g. Morris and Feldman, 1996; Wharton, 1993), however, view EL as a double-edged sword. First, the transformation – almost unconsciously – of what is typically a private act into a public act, and one conforming to rules set by others, causes both a loss of personal control and self-alienation (Hochschild, 1983). Second, as a result of obeying display rules, employees may experience emotional dissonance, or a conflict between expressed and experienced emotions (Morris and Feldman, 1996). Such emotional dissonance has been found to be negatively associated with well-being and performance (e.g. Cote and Morgan, 2002).

The inconclusive evidence regarding the effect of EL on exhaustion may reflect our incomplete understanding of the different types of EL. From a JD-R perspective (Demerouti et al., 2001), EL often requires effort and leads to the depletion of resources used for emotional control (Muraven and Baumeister, 2000). Yet, as noted above, this may not always be the case. Consistent with recent research in the JD-R tradition (e.g. Xanthopoulou et al., 2007), which suggests that the possession of personal resources may be crucial in buffering the impact of job demands on emotional exhaustion, we examine whether a cognitive-behavioural pattern – namely, psychological flexibility – may enable employees to better handle their emotions at work, over and above surface and deep acting. Notably, while most studies on the JD-R model have been restricted to external/organizational job resources, in the current study we focus on internal, personal recourses – aspects of the self that are generally linked to resiliency, or to individuals’ sense of being able to influence their environment successfully (Hobfoll, 2002). Such personal resources may serve the same three functions as job resources: aid in achieving work goals; reduce job demands and their associated physiological and psychological costs; and/or stimulate personal growth, learning and development (Bakker and Demerouti, 2007).

Surface and deep acting

Surface and deep acting involve deliberate efforts to align felt emotions with expected emotional displays (both are also known as controlled affect regulation: Parkinson and Totterdell, 1999). Research suggests that surface acting is likely to drain energy, as it involves prolonged internal tension between displayed and true feelings. In other words, the continuous experience and management of emotional dissonance entails high levels of psychological effort, and thus contributes to resource loss (Diefendorff and Gosserand, 2003; Grandey et al., 2005).

It is not clear whether deep acting has similar effects (Grandey, 2003). On the one hand, dealing with ongoing emotions, as occurs during surface acting, requires more effort than preventing the development of emotions that have not fully emerged, as occurs during deep acting (Kanfer and Kantrowitz, 2002). Deep acting involves the modification of feelings, so that continuous efforts to monitor expressions are only necessary up to the point where the employee has successfully modified his/her feelings. Thus, deep acting holds the promise of restoring energy and attention by reducing the conflict between feelings and expressions (Brotheridge and Lee, 2002; Grandey, 2003; Hochschild, 1983). On the other hand, deep acting involves conscious processes at the intellectual level of regulation, during the phase where the employee actively tries to influence his or her feelings to bring them into line with the emotional displays required by the firm. Deep acting is therefore considered cognitively effortful (Ashforth and Humphrey, 1993; Hochschild, 1983), but, again, only to a certain point. In line with these views, surface acting has consistently been shown to be linked to job stress, while deep acting has often displayed a weaker or no relationship with stress (e.g. Brotheridge and Lee, 2002; Grandey, 2003; Gross and John, 2003; Totterdell and Holman, 2003; Zapf, 2002).

Psychological flexibility

When facing a specific service situation, employees may choose ways other than surface or deep acting to handle their emotions. In particular, they may engage in acceptance and commitment – that is, accepting the present moment without needing to regulate associated thoughts and emotions, and adapting their behaviour to the situation as needed for the pursuit of goals (Hayes et al., 2006; Martin et al., 1998). More specifically, acceptance refers to a willingness to experience thoughts and emotions, including negative ones like anger or frustration, without letting them determine one’s actions (Bond and Bunce, 2003). Commitment refers to adapting one’s behaviour as needed to meet goals irrespective of negative thoughts and emotions (Hayes et al., 2006). Acceptance and commitment are inherently tied together as the two main processes characterizing the cognitive-behavioural pattern known as psychological flexibility (Hayes et al., 2006). This integration becomes apparent when we speak more technically about what psychological flexibility is: the degree to which cognition interacts with contingencies of reinforcement to help or hinder the ability to pursue goals. Thus, we believe that psychological flexibility satisfies the three definitional conditions of personal and job resources (helpful in achieving work goals; reducing job demands and their physiological and psychological costs; and stimulating personal growth, learning and development – Bakker and Demerouti, 2007).

The construct of psychological flexibility was developed in the context of acceptance and commitment therapy (ACT), a modern cognitive-behavioural therapy. ACT helps people to learn useful behaviours to deal with problematic psychological events, such as unpleasant thoughts and feelings (Bond et al., 2006). Inside and outside the realm of emotional job demands discussed here, many people become profoundly constrained by such negative internal events, as they divert energy and attention away from the present moment. Inflexible responses aimed at changing, minimizing or banishing unwanted thoughts and feelings are detrimental to goal-based actions and psychological well-being (Bond and Bunce, 2003). Indeed, such responses may paradoxically increase the severity and frequency of the unwanted thoughts, and amplify psychological symptoms (e.g. Wenzlaff and Wegner, 2000).

The premise of ACT is that people can learn to accept their unwanted feelings and respond to them with committed action that serves valued goals, thereby undermining the ability of negative stimuli to determine subsequent behaviour and emotional states (Bond et al., 2010). ACT argues that a negative internal event will not necessarily lead to poor well-being and performance. Rather, it is only when people ‘buy into’ the negative content of the event that it will have harmful emotional and behavioural effects. A key aim of ACT is to train people to treat their thoughts and feelings as automatic chatter, or more technically, as the ongoing classically and operantly conditioned responses that they are, such that they will remain aware of their thoughts and feelings, but base their actions upon their values and goals, not upon the vagaries of their internal events (Bond and Hayes, 2002; Bond et al., 2006).

In terms of the JD-R model, because psychologically flexible employees are willing to experience even unwanted internal events rather than engaging with or mulling over these events, and because they have greater goal-related context sensitivity, they are likely to observe the opportunities inherent in the current situation, and focus on the concrete issue to be dealt with, so as to maximize goal attainment (Bond and Bunce, 2003; Hayes et al., 1999). Consistent with this notion, Hayes et al. (1999) suggest that psychological flexibility is likely to promote mental health by lessening the impact of negative internal events, and by helping individuals to clearly define and accomplish goals that are meaningful to them.

While interventions based on acceptance and commitment have produced positive outcomes in people with a wide range of psychological problems (for a review see Hayes and Strosahl, 2004), the potential contribution of psychological flexibility to work and organizational settings has only rarely been investigated. However, the few studies that have examined this personal resource in work settings offer promising results. For example, psychological flexibility has been found to enhance employees’ ability to handle work strain (Bond and Bunce, 2003) and to reduce employee burnout (Hayes et al., 2006). In line with these findings and the aforementioned arguments, we propose:

Hypothesis 2: Person-level psychological flexibility is negatively associated with day-level emotional exhaustion, even after controlling for person-level surface and deep acting.

Emotional job demands, emotional labour and exhaustion: interaction effects

As described above, previous research (e.g. Bakker et al., 2004; Karasek, 1979; Karasek and Theorell, 1990) suggests that high job demands combined with low or inappropriate job resources are associated with an elevated risk of negative outcomes. More recent work by Bakker et al. (2005) also confirms that job resources protect people from the unhealthy effects (e.g. exhaustion) of a demanding work environment. While these and other researchers tend to focus on external (i.e. organizational) job resources as potential buffers against stress (e.g. social support, performance feedback), in the current study we respond to the call for more research on personal resources (e.g. Bakker and Demerouti, 2008). Accordingly, we propose that psychological flexibility, which concerns accepting ones’ emotions, is an individual characteristic that determines the degree to which emotional demands relate to exhaustion. In particular, we propose that psychological flexibility, in contrast to surface and deep acting, is likely to attenuate the demands–exhaustion relationship.

There are two reasons why psychological flexibility should have this effect. First, from a JD-R perspective (Demerouti et al., 2001), psychological flexibility may act as a personal resource that buffers negative aspects of emotionally demanding work conditions. More specifically, psychologically flexible employees need not invest effort in actively regulating their emotional reactions, and so will have more energy and attention available to deal effectively with job demands. In contrast, with surface acting (and, to some extent, deep acting), energy and attention are depleted by the continuous psychological effort needed to regulate genuinely felt emotions (Martínez-Iñigo et al., 2007). Under such conditions, it is likely that the effect of emotional job demands on emotional exhaustion is only exacerbated.

Second, to the degree that employees use different types of EL (i.e. active regulation vs. acceptance), they may interpret or frame job demands differently. Under conditions of surface acting (and again, to some extent, deep acting), employees may frame emotional job demands as problematic work situations likely to necessitate emotional and/or behavioural modification. In such cases, anticipation of the difficult situation may lead to increased emotional exhaustion. In contrast, psychologically flexible employees may frame the same job demands as simply part of the normal workday, in an open, non-elaborative and non-judgmental manner (Hayes et al., 2006; Lazarus, 1966). Taking these considerations into account, we propose:

Hypothesis 3a: Person-level surface acting, and to a lesser extent person-level deep acting, strengthen the positive association between day-level emotional job demands and day-level emotional exhaustion.

Hypothesis 3b: Person-level psychological flexibility attenuates the positive association between day-level emotional job demands and day-level emotional exhaustion.

Method

Sample

The sample comprised 254 participants identified through a network of family and friends of seven research assistants. The research assistants contacted potential respondents both directly (first-hand acquaintances) and indirectly (based on referrals by their friends/colleagues/family members, some of whom had already been invited to participate in the study). Such a convenience sample is often applied in diary studies, as completing the diary can require a considerable investment of time and dedication on the part of respondents, which may be more easily expected from relatives and friends (Conway and Briner, 2002; Higgins et al., 1985). In order to participate in the study, respondents had to: (a) be employed in a not-for-profit service organization; (b) have direct contact with customers for at least 50 percent of their working time; (c) have a non-managerial position; (d) have a job relating to the core business of the organization; and (e) have a full-time job, with a contract of at least 30 hours per week. Moreover, we sought to create a balanced sample in terms of gender and age. Accordingly, the research assistants were instructed to approach equal numbers of male/female as well as younger/older employees. In that participants were recruited by seven different research assistants, on different dates and in different locations, the likelihood that the sample method might have affected the results is small.

We chose to limit our sample to not-for-profit organizations because recent studies suggest that this sector, which has grown significantly in most Western countries over recent years (e.g. Lewis, 2004), faces challenges similar to those faced by the for-profit sector, making the two sectors more similar than previously thought. For example, many not-for-profit organizations respond to environmental changes in a way similar to their for-profit counterparts, by adopting a strategic approach to various functions including planning, operations and service management (e.g. Hume et al., 2006). In this context, issues such as emotional job demands, EL, and emotional exhaustion are very relevant to the functioning of not-for-profit employees, and warrant further investigation.

Data were eventually received from exactly 200 employees of not-for-profit service organizations in the Netherlands. Four respondents failed at closer inspection to meet the inclusion criteria and were therefore excluded from our analyses. An additional 26 observations were excluded owing to either suspect or excessive missing data (e.g. 33% or more uncompleted items). The final sample thus included 170 participants – an effective response rate of 66.9 percent – of which 45.9 percent worked in the health sector, 27.3 percent in the education sector, 9.8 percent in the government sector, 6.7 percent in leisure and cultural organizations (e.g. museums) and 10.3 percent in other organizations. Thanks to our balanced sampling method, participants were almost evenly distributed in terms of gender and age. The sample included a slightly higher number of women (56%), most likely because more women than men work in the not-for-profit service sector in the Netherlands (Statistics Netherlands, 2009). The mean age was approximately 40.8 years (SD = 12.45) – a figure in line with Dutch labour force statistics (Statistics Netherlands, 2009). The average percentage of time spent in interacting with customers was 79.4 percent (SD = 15.3). Thirty-four percent of the respondents had what is known in the Netherlands as a ‘low educational level’ (i.e. had some schooling, graduated from high school or graduated from a lower vocational/technical college) and 66 percent had a ‘middle or high educational level’ (i.e. graduated from a middle or higher vocational/technical college or had a university-level degree).

Procedure

We used two instruments for data collection. Participants were instructed to first complete a general questionnaire, and then complete a diary survey for three consecutive working days, beginning three or four days after completing the general questionnaire. Thus, completion of both research tasks (general questionnaire and diary) took place over a week or eight days. The three-to-four-day lag between the questionnaire and diary was long enough that participants would not accurately recall the contents of the former while filling in the latter, yet short enough to reduce drop-out rates. This is in line with prior studies using the diary technique (e.g. Sonnentag, 2003).

The questionnaire, diary and instructions were either handed out personally or sent by mail. Instructions regarding both research tasks were also provided to all participants via a phone call or face-to-face meeting. The instruction letter explained how to fill in the questionnaire and diary, guaranteed anonymity, highlighted the importance of the research and provided contact information in case of questions. Special emphasis was given to the time issue; participants were instructed to complete the questionnaire before the diary, and then, three to four days later, to fill in the diary surveys over three consecutive workdays, at the end of the day. The average time needed to complete the general questionnaire was 17 minutes. Filling in the diary survey took approximately 5 minutes a day. Completed questionnaires and diaries were collected personally or sent back to the research assistants by mail (stamped return envelopes were provided). After two weeks, respondents were reminded by telephone or email.

Critics of diary methods have raised questions about the extent to which participants comply with researchers’ instructions, particularly with regard to the timing of diary reports (e.g. Bolger et al., 2003). A study by Green et al. (2006) suggests that two factors play an important role in the degree to which participants comply with diary study instructions: the burden created by the study and the degree to which participant motivation is fostered. Our study involved a five-minute end-of-day diary to be completed within one hour of bedtime. The one-hour requirement was flexible in that entries completed two hours before bedtime were likely to serve the purpose of the study almost equally well. However, we regarded filling in a questionnaire the next morning and providing retrospective information about the previous day as clearly detrimental to the study’s aims, and participants were made aware of this. To reduce biased responses owing to experiences that intervened between the time participants were at work and the time they went to bed, the wording of the diary items clearly emphasized that participants should think about the specific work day when responding. The burden for participants was quite low, as participants provided only a single entry each day, and the average time for completing the survey was short. Furthermore, given our convenience sampling method, which was based on personal contacts, the research assistant–participant rapport was good. According to Green et al. (2006), this is likely to have resulted in high motivation among our participants. Considering the low burden and high motivation we believe that the rate of faked compliance, such as hoarding, was minimal.

Day-level (diary) measures

In each of the three diary surveys participants reported on their emotional job demands and emotional exhaustion with regard to the specific workday.

Emotional job demands were measured using five items from the Dutch Questionnaire on the Experience and Assessment of Work (known by its Dutch initials as the VBBA; see van Veldhoven et al., 2002). Participants responded using a four-point Likert scale ranging from (1) ‘never’ to (4) ‘always’. This scale has been used in other studies and has shown good psychometric properties (e.g. Bakker et al., 2004). Items were adapted to the day level: for example, ‘Did you have to interact with difficult customers at work today?’ and ‘Did you have to convince/persuade other people at work today?’. A pilot study (n = 5) designed to elicit potential difficulties in relating these items to the day level found no problems with the scale. Cronbach’s alpha for days 1, 2 and 3 was .82, .81 and .84, respectively.

Emotional exhaustion was measured using a scale from the ‘Utrecht Burnout Scale – general version’ (UBOS-A: Schaufeli and van Dierendonck, 2000). This is a Dutch version of the well-established ‘Maslach Burnout Inventory – general version’ of Maslach et al. (1996). The five-item scale was answered on a five-point Likert scale ranging from (1) ‘never’ to (5) ‘always’. Items were adapted to the day level: for example, ‘I feel emotionally drained from this working day’ and ‘I feel “used up” by this working day’. Here too, a pilot study identified no problem with the scale. Cronbach’s alpha for days 1, 2 and 3 was .89, .87 and .85, respectively.

Person-level (questionnaire) measures

The questionnaire measured the three EL-related variables (surface acting, deep acting and psychological flexibility) as well as the control variables.

Surface acting was measured using five items developed by Grandey (2003) and Brotheridge and Lee (1998). Participants answered on a five-point Likert scale ranging from (1) ‘strongly disagree’ to (5) ‘strongly agree’. Sample items included: ‘I fake a good mood when interacting with customers’; ‘I just pretend to have the emotions I need to display for my job’ (α = .91).

Deep acting was measured using three items developed by Grandey (2003) and Brotheridge and Lee (1998). Participants answered on a five-point Likert scale from (1) ‘strongly disagree’ to (5) ‘strongly agree’. Sample items included: ‘I try to actually experience the emotions that I must show to customers’; ‘I work hard to feel the emotions that I need to show to customers’ (α = .93).

Psychological flexibility was measured using the 10-item Acceptance and Action Questionnaire (AAQ-II: Bond et al., 2011). We adopted an existing Dutch translation of this scale, for which reliability and validity have been confirmed (Jacobs et al., 2008). Participants were asked to respond to the scale by thinking of how they deal with everyday situations (e.g. ‘It is OK if I remember something unpleasant’; see Appendix 1 for the complete list). They responded to each item on a seven-point Likert scale ranging from (1) ‘never true’ to (7) ‘always true’ (α = .75). Items were reversed-coded as appropriate such that higher scores indicated greater flexibility.

Bond et al. (2011) conducted a series of studies to investigate the psychometric properties of the AAQ-II as a unidimensional measure assessing the construct of psychological flexibility. First, they conducted an exploratory factor analysis (EFA) in order to establish a stable factor structure. This analysis revealed two distinct factors that accounted for a good proportion of the variance shared by the items. The results showed that the seven items that comprised Factor 1 were negatively worded, and the three that loaded on Factor 2 were positively worded (items 1, 6 and 10 in Appendix 1). This distinction in wording indicates that the two factors may not actually represent different constructs, but may have been produced by a method effect (Marsh, 1996). To confirm that the AAQ-II is, as hypothesized, a unidimensional scale assessing psychological flexibility, Bond et al. (2011) then conducted an EFA on four additional samples, and received similar results. Finally, they conducted a set of confirmatory factor analyses (CFA) using three new samples, in which they compared the relative fit of two different models. Consistent with Marsh (1996), the first was a one-factor model that did not specify the presence of those directional (positive/negative) method effects, while the second specified measurement error correlations among the three positively worded items. In all three samples, the second model showed a better fit to the data than the first model, indicating a method effect.

With respect to the three EL-related variables, we tested the goodness-of-fit of the hypothesized three-factor model in comparison to four other competing nested models (three with two factors and one with a single factor) through sequential chi-square difference tests. The results indicate the superiority of the three-factor model over the one-factor model (χ²(131) = 215.63 and Δχ²(4) = 890.58, respectively, both at p < .01; RMSEA = .06 and .19, respectively). All items loaded on the respective factors with values ranging from .53 to .92 and with no cross-loadings above .29.

Control variables. In order to rule out spurious relations, we controlled for gender and age (in years). These variables have been identified as related to emotional job demands and/or emotional exhaustion (e.g. Vermeulen and Mustard, 2000). In addition, in all analyses we controlled for person-level emotional exhaustion using the same scale by Schaufeli and van Dierendonck (2000) used in the diary survey (but with the original, general wording). Sample items included: ‘I feel emotionally drained from my work’ and ‘I feel ‘used up’ at the end of the work day’ (α = .85).

Analysis technique

Given that emotional job demands and emotional exhaustion were measured for three consecutive days, our data can be viewed as multilevel data with repeated measurements (days) nested within individuals. This leads to a two-level model with the days at the first level (level 1; N = 510 occasions) and the individual employees at the second level (level 2; N = 170 participants). We calculated the intra-class correlation (ρ) to examine whether there is enough variance at both the between- and within-person level in day-level emotional exhaustion. For day-level emotional exhaustion ρ was .21, suggesting that 21 percent of the variance is attributable to between-person fluctuations and 79 percent to within-person fluctuations. These results support the use of multilevel modelling in our study. Multilevel models were estimated using PROC MIX in SAS.

Multilevel analysis allows modelling simultaneously within- and between-group variance in the examined factors (Dierdorff and Ellington, 2008). In the present study, the predictor variable at the day level (level 1; i.e. day-level emotional demands) was centred to the respective person mean to strictly reflect intra-individual processes. Predictor variables at the person level (level 2; i.e. surface acting, deep acting and psychological flexibility) were centred around the grand mean. Centring day-level emotional demands at the person mean allows removal of the between-person variance from this variable. This eliminates the role of stable differences in explaining participants’ day-level emotional demands from the analyses (see also Sonnentag and Niessen, 2008; Tims et al., 2011).

Results

Means, standard deviations and correlations among the variables are displayed in Table 1. The bivariate results indicate a positive relationship between day-level emotional job demands and day-level emotional exhaustion (r = .33, p < .01). The results also indicate a positive correlation between surface acting and daily exhaustion (r = .52, p < .01) and an inverse correlation between psychological flexibility and daily exhaustion (r = −.43, p < .01).

Table 1

Means, standard deviations and correlations (Pearson) of the measured variables (N = 170 employees)

	Variable	M	SD	1	2	3	4	5	6	7
1	Gender^a	1.56	.50	–
2	Age^a	40.81	12.45	−.34**	–
3	Emotional exhaustion^a	2.46	.76	−.19*	−.04	–
4	Emotional job demands^b (3-day mean)	2.36	.49	−.10	−.08	.16*	–
5	Surface acting^a	2.09	.76	−16*	−.06	.33**	.15	–
6	Deep acting^a	2.60	.92	.04	.07	−.13	−.14	.07	–
7	Psychological flexibility^a	5.33	.61	−.04	.03	−.38**	−.16*	−.22**	−.05	–
8	Emotional exhaustion^b (3-day mean)	1.96	.74	−.05	−.21**	.56**	.33**	.52**	−.11	−.43**

p < .05, ** p < .01.

Person level; ^bday level.

Response scale information: Emotional exhaustion 1–5; Emotional job demands 1–4; Surface and deep acting 1–5; Psychological flexibility 1–7.

The results of our multilevel analyses testing Hypothesis 1 (which specified that day-level emotional job demands are positively associated with day-level emotional exhaustion) are presented in Table 2 (Model 2). The results support Hypothesis 1 (B = .32, SE = .10, t = 3.2, p < .01).

Table 2

Multilevel estimates of models predicting day-level emotional exhaustion (N = 170 employees and N = 510 days)

Model Variable	(1) Control model			(2) Main effect model			(3) Model including surface/deep acting			(4) Model including psychological flexibility			(5) Interaction (full) model
	B	SE	t	B	SE	t	B	SE	t	B	SE	t	B	SE	t
Gender	−.04	.10	−.41	−.00	.09	−.02	.07	.09	.78	.03	.08	.32	.01	.08	.14
Age	−.01	.00	−3.7**	−.01	.00	−3.3**	−.01	.00	−2.9**	−.01	.00	−3.0**	−.01	.00	−3.4**
Emotional exhaustion	−.54	.06	9.0**	.51	.06	8.8**	.41	.06	7.3**	.34	.06	5.8**	.37	.06	6.6**
Emotional job demands				.32	.10	3.2**	.27	.09	2.8**	.24	.09	2.6*	.18	.09	2.0*
Surface acting							.31	.05	6.0**	.29	.05	5.8**	.10	.20	.46
Deep acting							−.03	.04	−.74	−.05	.04	−1.1	−.18	.19	.95
Psychological flexibility										−.22	.06	−3.5**	−.33	.27	−1.2
Surface acting × Emotional job demands													.18	.08	2.1*
Deep acting × Emotional job demands													−.10	.08	−1.3
Psychological flexibility × Emotional job demands													−.24	.11	2.2*
Model summary
−2 log	768.3			753.4			720.7			709.2			693.2
Δ−2×log	76.6**^a			14.9**			32.7**			11.5**			16**
Level 1 (within-person) variance	R² = .18			R² = .20			R² = .22			R² = .27			R² = .30
Level 2 (between-person) variance	R² = .36			R² = .41			R² = .52			R² = .55			R² = .59

p<0.05; **p<0.01.

Relative to the null model.

The results presented in Table 2 also support our Hypothesis 2 (positing that, even after controlling for person-level surface and deep acting, person-level psychological flexibility is associated with lower levels of daily emotional exhaustion). Specifically, as Model 3 indicates, person-level surface acting was found to be associated with higher levels of daily exhaustion, and person-level deep acting was not related to daily exhaustion (B = .31, SE = .05, t = 6.0, p < .01 and B = −.03, SE = .04, t = −.74, p > .05, respectively). Importantly, as Model 4 indicates, person-level psychological flexibility was associated with lower levels of daily exhaustion (B = −.22, SE = .06, t = −3.5, p < .01).

In order to test Hypotheses 3a (positing that person-level surface acting and, to a lesser extent, deep acting strengthen the positive association between day-level emotional job demands and day-level emotional exhaustion) and 3b (positing that person-level psychological flexibility attenuates that association), the interaction terms, based on the centred variables, were incorporated into the full model. As shown in Model 5 of Table 2, the positive association between day-level emotional demands and day-level emotional exhaustion was found to be amplified as a function of person-level surface acting (B = .18, SE = .08, t = 2.1, p < .05). Person-level deep acting showed no synergistic effect with daily emotional job demands (B = −.10, SE = .08, t = −1.3, p > .05). Not only do these effects have opposing signs, but they also differ in magnitude (t = 2.18 (d.f. = 159), p < .05; the Satterthwaite approximation for the denominator degrees of freedom was used; Littell et al., 1996). Accordingly, Hypothesis 3a was partly supported; the moderation effect of surface acting was, as predicted, stronger than the moderation effect of deep acting, yet the latter was insignificant. In addition, Model 5 provides support for Hypothesis 3b; the demands–exhaustion association was found to be attenuated as a function of psychological flexibility (B = −.24, SE = .11, t = 2.2, p < .05).

To further examine the effect of surface acting and psychological flexibility on the link between emotional job demands and emotional exhaustion, we calculated the regions of significance of the simple slopes of emotional job demands conditioned on surface acting and psychological flexibility, utilizing the Rweb application developed by Preacher et al. (2004; 2006). The Rweb application is intended to depict the confidence bands of a simple slope for any value of the moderator (the X-axis, ranging, as a default, between −10 and +10). Thus, Figures 1 and 2 show the value of the simple slope (the Y-axis) for each value of the moderator, and whether the simple slope is significant for each value of the moderator. When the confidence band contains zero, the simple slope is not significant. The results suggest that for all surface acting values greater than −0.2388 (conditional values at the mean level of psychological flexibility) the simple slope of emotional job demands is positively significant. Further, Figure 1 suggests that when surface acting is low (below −0.2388) the simple slope is insignificant (i.e. the relationship between emotional demands and exhaustion is insignificant). In addition, the results suggest that for all psychological flexibility values greater than 0.2022 (conditional values at mean level of surface acting) the simple slope of emotional job demands is insignificant. Finally, Figure 2 suggests that when psychological flexibility is low (below 0.2022) the simple slope is positively significant (i.e. the relationship between emotional demands and exhaustion is positive and significant).

Figure 1

Plot of confidence bands for observed sample values of surface acting.

Figure 2

Plot of confidence bands for observed sample values of psychological flexibility.

Discussion

Our study extends previous research by demonstrating the potential advantages of a cognitive-behavioural pattern, psychological flexibility, which enables service employees to use a form of EL that does not mandate the active regulation of emotions – namely, acceptance and commitment. Specifically, we found that while emotional exhaustion was positively associated with surface acting, it was negatively associated with psychological flexibility. These results are particularly important in the context of service delivery, as service providers are obliged to continuously display certain emotions during their interactions with customers (Morris and Feldman, 1996). By enabling employees to accept unpleasant emotions (for example, those arising during interactions with difficult or irate customers), psychological flexibility redirects energy and attention away from the intensive regulation of emotions to align them with display rules, as in surface acting, and toward other, more professionally relevant and less psychologically consuming tasks (Bond et al., 2006; Hayes et al., 2006). Although, in the current study, we did not directly test whether the level of effort invested in regulating spontaneous emotions explains the different associations between surface acting/deep acting/psychological flexibility and employees’ subsequent emotional exhaustion, a recent study by Martínez-Iñigo et al. (2007) may offer some initial insights on this mechanism – at least with respect to surface acting. On the basis of data drawn from 345 general practitioners in Spain, these researchers concluded that the positive relationship between surface acting and emotional exhaustion was partially mediated by participants’ psychological effort.

Our study responds to calls to explore the role of personal resources as buffering variables in stressor–strain relations (e.g. Bakker et al., 2004; Jex et al., 2001). In this respect, it is worth noting the body of literature that has examined the relationship between EL and emotional intelligence, and the role of the latter in the demand–exhaustion relationship (Johnson and Spector, 2007; Wong and Law, 2002). Psychological flexibility and emotional intelligence are both meta-cognitive constructs that emphasize the extent to which people can perceive their thoughts and feelings (Goleman, 1998; Mayer and Salovey, 1997; Mayer et al., 2000). However, the two constructs differ in their understanding of how meta-cognition is used to promote well-being and performance outcomes. It is in the regulation (or control) of emotions that the key distinction between the two can be found. Psychological flexibility, as a concept, is based on the idea that emotions should be accepted, not regulated, and that regulating emotions will diminish well-being and performance. Emotional intelligence, in contrast, is defined as the ability to identify, assess and regulate emotions, with the latter seen as a mechanism for enhancing well-being and performance (Donaldson-Fielder and Bond, 2004; George, 2010; Vincent, 2011).

The results of Hypothesis 3a indicate that surface acting may contribute to (i.e. exacerbate) the development of emotional exhaustion under conditions of high emotional job demands, whereas deep acting seems to have no meaningful influence on the demand–exhaustion association. These results mirror the mixed findings reported by others (e.g. Brotheridge and Lee, 2002; Gross and John, 2003) regarding the potential effects of the two commonly used forms of EL. That deep acting did not moderate emotional demands might reflect two opposing processes. Hypothesis 3a was based on the idea that deep acting required some degree of psychological effort in response to emotional demands. On the other hand, because deep acting is linked to feelings of self-efficacy, personal accomplishment, status and the like, employees using this form of EL may feel less vulnerable, and more competent, to handle job demands (e.g. Brotheridge and Lee, 2002; Grandey, 2003). At the same time, however, our results concerning Hypothesis 3b suggest that psychological flexibility helps compensate for the exhaustion associated with emotional job demands. As Figure 2 illustrates, emotional job demands were positively associated with exhaustion when psychological flexibility was low, but not when it was high. Consistent with the JD-R model (Demerouti et al., 2001), the underlying mechanism of this moderation effect may be that psychological flexibility enables employees to be less disturbed by potentially stressful conditions, including emotional job demands. As a result, they use their energy more effectively in doing their job.

The mean psychological flexibility score for the present study (5.33) may provide indirect support for this notion. This relatively high mean score may have to do with the nature of our sample. People who work for not-for-profit organizations – unlike service employees in for-profit firms – may tend to focus less on short-term goals or targets and more on professional and organizational long-term objectives. They may be able to see each day’s work as part of a pattern of progress towards a specific, long-term contribution to important goals for society (good quality of health care, education, etc.). This sensitivity to a long-term goal-related context may help them focus on the opportunities inherent in everyday situations rather than on the potentially difficult or frustrating aspects (Bond et al., 2010; Bond et al., 2008; Hayes et al., 2006).

Taken together, our results suggest that psychological flexibility holds promising opportunities for research in work and organizational settings. But how does psychological flexibility relate to surface acting and deep acting in predicting such outcomes as exhaustion? Can different forms of EL be present in the same individual? If so, can they be present at the same time? The latter seems unlikely. Given that surface and deep acting involve efforts to actively suppress or change feelings, whereas acceptance involves willingness to experience feelings, it does not seem logical that these ways of managing feelings could occur at the same point in time. That is, when handling a specific situation, a person is likely to engage in only one of the three: suppressing, changing or accepting the feelings that arise.

On the other hand, individuals can generally score high overall in two or even three modes of EL. In our sample, 18 respondents scored high (+SD) in both psychological flexibility and surface acting, 17 in both psychological flexibility and deep acting, 18 in both surface acting and deep acting, and 5 in all three measures. This is consistent with the modest correlations documented between the three constructs (see Table 1). All in all, there is little support for an interpretation of psychological flexibility as necessarily ruling out surface or deep acting in the same individual. This relates to the broader question of whether individuals differ systematically in their use of (regulation/acceptance-based) EL – a question still under debate (Ashkanasy and Humphrey, 2010; Gross and John, 2003). While EL is not a personality trait, but rather a relatively malleable characteristic that is open to change and may vary, to a certain extent, within the same person across situations (e.g. Judge et al., 2009), individuals may habitually use certain types of EL. For example, men tend to use more surface acting than women (Gross and John, 2003); employees high in negative affectivity tend to use more surface acting (e.g. Brotheridge and Lee, 2002); and employees high in emotional intelligence tend to use more deep acting (Cheung and Tang, 2009).

Along these lines, the findings of the current study suggest that while the EL literature has tended to refer to surface and deep acting in the narrow sense of ways of coping, it might be useful to see them as elements in a wider cognitive-behavioural repertoire (along with acceptance and commitment, among others) over which the employee exercises a certain degree of mastery, and which allows the individual to (learn to) better handle environmental contingencies, such as emotion-generating situations, and better achieve work-related goals.

Limitations and suggestions for future research

The findings of this study should be considered in light of its limitations, which may also offer research opportunities. First, our sample was fairly homogeneous in that it comprised only non-managerial service personnel working in not-for-profit organizations. Although this sample minimizes potential confounding factors such as job level or job type, our findings may not be generalizable to other occupations or branches of industry (in particular, for-profit firms). Future studies may wish to investigate samples drawn from other work populations.

Second, in our analyses we controlled for gender and age as potentially confounding variables. Previous studies have also accounted for the effect of dispositional variables (e.g. negative affect; Bond and Bunce, 2003), but there are doubtless other relevant variables that should be examined. In particular, future research may consider employees’ perceptions regarding predefined display rules or the degree to which employees are committed to such rules. This may be especially interesting in light of recent studies (e.g. Wright and Nishii, 2008) suggesting that employees often act according to what they perceive are the interests and goals of the organization.

Third, the research data were self-reported and thus may be biased by perceived social desirability. However, by controlling for person-level emotional exhaustion and using person-level data for EL, we can indirectly rule out this potential threat. In addition, we conducted Harman’s single factor test to assess the potential risk of common method bias (Podsakoff et al., 2003). We did this separately for the general questionnaire and for each of the three diary surveys. Item-level principal component analyses with varimax rotation resulted in the expected factors, while in all four analyses the first factor never explained the majority of the total explained variance. These results can at least partially allay concerns about method variance in this study. Future research may further address the issue of method variance by, for example, examining perceptions of job demands from the perspective of others, in particular co-workers and supervisors, and by considering objective measures (e.g. blood pressure or cardiovascular problems) that can be related to emotional exhaustion.

Implications

Despite these limitations, our findings have several implications for research and practice. With regard to research, our findings shed light on psychological flexibility as an important predictor of emotional exhaustion, even when accounting for surface and deep acting. This suggests that Hayes’ (1987) theory of psychological acceptance, which considers mental health from the perspective of clinical psychology, may enhance our understanding of this outcome in the context of organizational behaviour.

In terms of practical implications, some firms go further than prescribing display rules, also advising employees to engage in certain forms of EL in order to meet service requirements. These prescriptions may be based on inaccurate assumptions, ignoring the potential maladaptive outcomes of certain types of EL to employee well-being. Given that psychological flexibility is a cognitive-behavioural pattern with some stability and generalizability, but also a pattern that is malleable and can be developed through training, service firms may benefit from assessing and enhancing psychological flexibility. This implication is consistent with several studies that examined acceptance-focused interventions aimed at teaching service employees how to become more ‘accepting’ in executing their work. For example, Bond and Bunce (2003) described a worksite training program that successfully increased psychological flexibility among call-centre operators (from a baseline measurement), resulting in improved mental health and performance. Organizational efforts to increase psychological flexibility may constitute an important opportunity to sustain well-being and performance among service providers, and thereby may indirectly buttress branches of industry that are under increasing societal pressure to achieve higher-quality service and care at lower costs.

Footnotes

Appendix

Appendix 1

Acceptance and Action Questionnaire AAQ-II (Bond et al., 2011)

1. It is OK if I remember something unpleasant.

2. My painful experiences and memories make it difficult for me to live a life that I would value. (R)

3. I am afraid of my feelings. (R)

4. I worry about not being able to control my worries and feelings. (R)

5. My painful memories prevent me from having a fulfilling life. (R)

6. I am in control of my life.

7. Emotions cause problems in my life. (R)

8. It seems like most people are handling their lives better than I am. (R)

9. Worries get in the way of my success. (R)

10. My thoughts and feelings do not get in the way of how I want to live my life.

Acknowledgements

The authors are extremely grateful for the assistance of Etti Doveh from the Technion Statistical Laboratory, Haifa, Israel. Special thanks also go to C Lentz, M Onwezen, J Meijer, D Stoepker, N Curiel and K Vernaeve for their assistance in data collection.

Funding

This study was supported by a Site Visit Grant Award from the Industry Studies Association.

Michal Biron is a lecturer at the Graduate School of Management, University of Haifa, Israel, and a visiting researcher in the Department of Human Resource Studies at Tilburg University, The Netherlands. Michal’s research focuses on the nature and implications of peer relations, the role of HRM in organizations, and employee misbehaviours. Her articles have been published in Journal of Applied Psychology, Organizational Behavior and Human Decision Processes, Journal of Vocational Behavior and Journal of Occupational Health Psychology.

Marc van Veldhoven is Professor in the Department of Human Resource Studies at Tilburg University, The Netherlands. His main interest is in building bridges between research on occupational health psychology and HRM. He has published in Work & Stress, Occupational & Environmental Medicine, International Journal of Stress Management, Journal of Personnel Psychology and European Journal of Work and Organizational Psychology. Marc is an Associate Editor for the Journal of Occupational and Organizational Psychology.

References

Ashforth

Humphrey

(1993) Emotional labor in service roles: The influence of identity. Academy of Management Review 18(1): 88–115.

Ashkanasy

Humphrey

(2010) Current emotion research in organizational behavior. Emotion Review 3(2): 214–224.

Bakker

Demerouti

(2007) The job demands-resources model: State of the art. Journal of Managerial Psychology 22(3): 309–328.

Bakker

Demerouti

(2008) Towards a model of work engagement. Career Development International 13(3): 209–223.

Bakker

Demerouti

Taris

Schaufeli

Schreurs

(2003) A multi-group analysis of the job demands-resources model in four home care organizations. International Journal of Stress Management 10(1): 16–38.

Bakker

Demerouti

Verbeke

(2004) Using the job demands-resources model to predict burnout and performance. Human Resource Management 32(1): 83–104.

Bakker

Demerouti

Euwema

(2005) Job resources buffer the impact of job demands on burnout. Journal of Occupational Health Psychology 10(2): 170–180.

Bolger

Davis

Rafaeli

(2003) Diary methods: Capturing life as it is lived. Annual Review of Psychology 54(1): 579–616.

Bond

Bunce

(2003) The role of acceptance and job control in mental health, job satisfaction, and work performance. Journal of Applied Psychology 88(6): 1057–1067.

10.

Bond

Flaxman

(2006) The ability of psychological flexibility and job control to predict learning, job performance, and mental health. Journal of Organizational Behavior Management 26(1–2): 113–130.

11.

Bond

Hayes

(2002) ACT at work. In: Bond

Dryden

(eds) Handbook of Brief Cognitive Behaviour Therapy. Chichester: Wiley, 117–140.

12.

Bond

Hayes

Barnes-Holmes

(2006) Psychological flexibility, ACT, and organizational behavior. Journal of Organizational Behavior Management 26(1–2): 25–54.

13.

Bond

Flaxman

Bunce

(2008) The influence of psychological flexibility on work redesign: Mediated moderation of a work reorganization intervention. Journal of Applied Psychology 93(3): 645–654.

14.

Bond

Flaxman

van Veldhoven

Biron

(2010) The impact of psychological flexibility and Acceptance and Commitment Therapy (ACT) on health and productivity at work. In: Houdmont

Leka

(eds) Contemporary Occupational Health Psychology: Global Perspectives on Research and Practice. Chichester: Wiley, 296–313.

15.

Bond

Hayes

Baer

Carpenter

Orcutt

Waltz

Zettle

(2011) Preliminary psychometric properties of the Acceptance and Action Questionnaire – II: A revised measure of psychological flexibility and acceptance. Behavior Therapy 42(4): 1–38.

16.

Brotheridge

Lee

(1998) On the dimensionality of emotional labor: Development and validation of an emotional labor scale. Paper presented at the First Conference on Emotions in Organizational Life, San Diego, CA, August.

17.

Brotheridge

Lee

(2002) Testing a conservation of resources model of the dynamics of emotional labor. Journal of Occupational Health Psychology 7(1): 57–67.

18.

Butler

Grzywacz

Bass

Linney

(2005) Extending the demands–control model: A daily diary study of job characteristics, work–family conflict and work–family facilitation. Journal of Occupational and Organizational Psychology 78(2): 155–169.

19.

Chebat

Kollias

(2000) The impact of empowerment on customer contact employees’ role in service organizations. Journal of Service Research 3(1): 66–81.

20.

Cheung

Tang

(2009) The influence of emotional intelligence and affectivity on emotional labor strategies at work. Journal of Individual Differences 30(2): 75–86.

21.

Conway

Briner

(2002) A daily diary of affective responses to psychological contract breach and exceeded promises. Journal of Organizational Behavior 23(3): 287–302.

22.

Cote

Morgan

(2002) A longitudinal analysis of the association between emotional regulation, job satisfaction, and intentions to quit. Journal of Organizational Behavior 23(8): 947–962.

23.

D’Ausilio

(1997) The impact of conflict management on customer service delivery. Customer Relationship Management 2(4): 20–23.

24.

Deery

Iverson

Walsh

(2002) Work relationships in telephone call centres: Understanding emotional exhaustion and employee withdrawal. Journal of Management Studies 39(4): 471–496.

25.

Demerouti

Bakker

Nachreiner

Schaufeli

(2001) The job demands-resources model of burnout. Journal of Applied Psychology 86(3): 499–512.

26.

Diefendorff

Gosserand

(2003) Understanding the emotional labor process: A control theory perspective. Journal of Organizational Behavior 24(8): 945–959.

27.

Diefendorff

Greguras

(2009) Contextualizing emotional display rules: Examining the roles of targets and discrete emotions in shaping display rule perceptions. Journal of Management 35(4): 880–898.

28.

Diefendorff

Richard

(2003) Antecedents and consequences of emotional display rule perceptions. Journal of Applied Psychology 88(2): 284–294.

29.

Diefendorff

Croyle

Gosserand

(2005) The dimensionality and antecedents of emotional labor strategies. Journal of Vocational Behavior 66(2): 339–357.

30.

Dierdorff

Ellington

(2008) It’s the nature of work: Examining behavior-based sources of work-family conflict across occupations. Journal of Applied Psychology 93(4): 883–892.

31.

Donaldson-Fielder

Bond

(2004) The relative importance of psychological acceptance and emotional intelligence to workplace well-being. British Journal of Guidance & Counselling 32(2): 187–203.

32.

George

(2010) Emotions and leadership: The role of emotional intelligence. Human Relations 53(8): 1027–1055.

33.

Goleman

(1998) Working with Emotional Intelligence. London: Bloomsbury.

34.

Grandey

(2000) Emotion regulation in the workplace: A new way to conceptualize emotional labor. Journal of Occupational Health Psychology 5(1): 95–110.

35.

Grandey

(2003) When ‘The show must go on’: Surface acting and deep acting as determinants of emotional exhaustion and peer-rated service delivery. Academy of Management Journal 46(1): 86–96.

36.

Grandey

Fisk

Steiner

(2005) Must ‘service with a smile’ be stressful? The moderating role of personal control for U.S. and French employees. Journal of Applied Psychology 90(5): 893–904.

37.

Grandey

Kern

Frone

(2007) Verbal abuse from outsiders versus insiders: Comparing frequency, impact on emotional exhaustion, and the role of emotional labor. Journal of Occupational Health Psychology 12(1): 63–79.

38.

Green

Rafaeli

Bolger

Shrout

Reis

(2006) Paper or plastic? Data equivalence in paper and electronic diaries. Psychological Methods 11(1): 87–105.

39.

Gross

John

(2003) Individual differences in two emotion regulation processes: Implications for affect, relationships, and well-being. Journal of Personality and Social Psychology 85(2): 348–362.

40.

Hayes

(1987) A contextual approach to therapeutic change. In: Jacobson

(ed.) Psychotherapists in Clinical Practice: Cognitive and Behavioral Perspectives. New York: Guilford Press, 327–387.

41.

Hayes

Strosahl

(2004) A Practical Guide to Acceptance and Commitment Therapy. New York: Springer, Inc.

42.

Hayes

Strosahl

Wilson

(1999) Acceptance and Commitment Therapy: An Experiential Approach to Behavior Change. New York: Guilford.

43.

Hayes

Luoma

Bond

Masuda

Lillis

(2006) Acceptance and commitment therapy: Model, processes and outcomes. Behaviour Research and Therapy 44(1): 1–25.

44.

Higgins

McClean

Conrath

(1985) The accuracy and biases of diary communication data. Social Networks 7(2): 173–187.

45.

Hobfoll

(2002) Social and psychological resources and adaptation. Review of General Psychology 6(4): 307–324.

46.

Hochschild

(1979) Emotion work, feeling rules, and social structure. The American Journal of Sociology 85(3): 551–575.

47.

Hochschild

(1983) The Managed Heart: Commercialization of Human Feelings. Berkeley: University of California Press.

48.

Hume

Mort

Liesch

Winzar

(2006) Understanding service experience in non-profit performing arts: Implications for operations and service management. Journal of Operations Management 24(4): 304–324.

49.

Ilies

Schwind

Heller

(2007) Employee well-being: A multilevel model linking work and nonwork domains. European Journal of Work and Organizational Psychology 16(3): 326–341.

50.

Jacobs

Kleen

de Groot

A-Tjak

(2008) Het meten van experiëntiële vermijding: De Nederlandstalige versie van de Acceptance and Action Questionnaire-II (AAQ-II) [Measuring experiential avoidance: The Dutch version of the Acceptance and Action Questionnaire-II (AAQ-II)]. Gedragstherapie 41: 349–361.

51.

Jex

Bliese

Buzzell

Primeau

(2001) The impact of self-efficacy on stressor-strain relations: Coping style as an explanatory mechanism. Journal of Applied Psychology 86(3): 401–409.

52.

Johnson

Spector

(2007) Service with a smile: Do emotional intelligence, gender, and autonomy moderate the emotional labor process? Journal of Occupational Health Psychology 12(4): 319–333.

53.

Judge

Fluegge-Woolf

Hurst

(2009) Is emotional labor more difficult for some than for others? A multilevel, experience-sampling study. Personnel Psychology 62(1): 57–88.

54.

Kanfer

Kantrowitz

(2002) Emotional regulation: Command and control of emotion in work life. In: Lord

Klimoski

Kanfer

(eds) Emotions in the Workplace. San Francisco, CA: Jossey-Bass, 433–472.

55.

Karasek

(1979) Job demands, job decision latitude, and mental strain: Implications for job redesign. Administrative Science Quarterly 24(2): 285–308.

56.

Karasek

Theorell

(1990) Healthy Work. New York: Basic Books.

57.

Larsen

Prizmic

(2004) Affect regulation. In: Baumeister

Vohs

(eds) Handbook of Self-Regulation: Research, Theory, and Applications. New York: Guilford, 40–61.

58.

Lazarus

(1966) Psychological Stress and the Coping Process. New York: McGraw-Hill.

59.

Lee

Wolpin

(2006) Intersectoral labor mobility and the growth of the service sector. Econometrica 74(1): 1–46.

60.

Lewis

(2004) The state and the third sector in modern welfare states: Independence, instrumentality, partnership. In: Evers

Laville

(eds) The Third Sector in Europe. Cheltenham, UK: Edward Elgar, 169–187.

61.

Littell

Milliken

Stroup

Wolfinger

(1996) SAS System for Mixed Models. Cary, NC: SAS Institute Inc.

62.

Marsh

(1996) Positive and negative global self-esteem: A substantively meaningful distinction or artifactors? Journal of Personality and Social Psychology 70(4): 810–819.

63.

Martin

Knopoff

Beckman

(1998) An alternative to bureaucratic impersonality and emotional labor: Bounded emotionality at The Body Shop. Administrative Science Quarterly 43(2): 429–469.

64.

Martínez-Iñigo

Totterdell

Alcover

Holman

(2007) Emotional labour and emotional exhaustion: Interpersonal and intrapersonal mechanisms. Work & Stress 21(1): 30–47.

65.

Maslach

(2003) Job burnout: New directions in research and intervention. Current Directions in Psychological Science 12(5): 189–192.

66.

Maslach

Jackson

(1981) The Maslach Burnout Inventory. Palo Alto, CA: Consulting Psychologists Press.

67.

Maslach

Jackson

Leiter

(1996) Maslach Burnout Inventory Manual, 3rd edn. Palo Alto, CA: Consulting Psychologists Press.

68.

Maslach

Schaufeli

Leiter

(2001) Job burnout. Annual Review of Psychology 52: 397–422.

69.

Mayer

Salovey

(1997) What is emotional intelligence? In: Salovey

Sluyter

(eds) Emotional Development and Emotional Intelligence: Educational Implications. New York: Basic Books, 3–31.

70.

Mayer

Salovey

Caruso

(2000) Models of emotional intelligence. In: Sterenberg

(ed.) Handbook of Intelligence. Cambridge: Cambridge University Press, 396–422.

71.

Morris

Feldman

(1996) The dimensions, antecedents, and consequences of emotional labor. Academy of Management Review 21(4): 986–1010.

72.

Muraven

Baumeister

(2000) Self-regulation and depletion of limited resources: Does self-control resemble a muscle? Psychological Bulletin 126(2): 247–259.

73.

Parkinson

Totterdell

(1999) Classifying affect-regulation strategies. Cognition and Emotion 13(3): 277–303.

74.

Peeters

MCW

Montgomery

Bakker

Schaufeli

(2005) Balancing work and home: How job and home demands are related to burnout. International Journal of Stress Management 12(1): 43–61.

75.

Podsakoff

MacKenzie

Lee

Podsakoff

(2003) Common method biases in behavioral research: A critical review of the literature and recommended remedies. Journal of Applied Psychology 88(5): 879–903.

76.

Preacher

Curran

Bauer

(2004) Simple intercepts, simple slopes, and regions of significance in MLR 2-way interactions. Available at: http://quantpsy.org/interact/hlm2.htm (accessed 3 November 2011).

77.

Preacher

Curran

Bauer

(2006) Computational tools for probing interaction effects in multiple linear regression, multilevel modeling, and latent curve analysis. Journal of Educational and Behavioral Statistics 31(1): 437–448.

78.

Schaufeli

van Dierendonck

(2000) UBOS: Utrechtse Burnout Schaal—Handleiding [UBOS: Utrecht Burnout Scale – Manual]. Lisse, The Netherlands: Swets & Zeitlinger.

79.

Simbula

(2010) Daily fluctuations in teachers’ well-being: A diary study using the Job Demands–Resources model. Anxiety, Stress & Coping: An International Journal 23(5): 563–584.

80.

Sonnentag

(2003) Recovery, work engagement, and proactive behavior: A new look at the interface between non-work and work. Journal of Applied Psychology 88(3): 518–528.

81.

Sonnentag

Niessen

(2008) Staying vigorous until work is over: The role of trait vigour, day-specific work experiences and recovery. Journal of Occupational and Organizational Psychology 81(3): 435–458.

82.

Sonnentag

Dormann

Demerouti

(2010) Not all days are created equal: The concept of state work engagement. In: Bakker

Leiter

(eds) Work Engagement: The Essential in Theory and Research. New York: Psychology Press, 25–38.

83.

Statistics Netherlands (2009) Labor and social security. Available at: http://www.cbs.nl/nl-NL/menu/themas/arbeid-sociale-zekerheid/nieuws/default.htm (accessed 29 January 2009).

84.

Taris

(2006) Is there a relationship between burnout and objective performance? A critical review of 16 studies. Work & Stress 20(4): 316–334.

85.

Thayer

Newman

McClain

(1994) Self-regulation of mood: Strategies for changing a bad mood, raising energy, and reducing tension. Journal of Personality and Social Psychology 67(5): 910–925.

86.

Tims

Bakker

Xanthopoulou

(2011) Do transformational leaders enhance their followers’ daily work engagement? The Leadership Quarterly 22(1): 121–131.

87.

Totterdell

Holman

(2003) Emotion regulation in customer service roles: Testing a model of emotional labor. Journal of Occupational Health Psychology 8(1): 55–73.

88.

van Veldhoven

de Jonge

Broersen

Kompier

Meijman

(2002) Specific relationships between psychosocial job conditions and job related stress: A three-level analytic approach. Work and Stress 16(3): 207–228.

89.

Vermeulen

Mustard

(2000) Gender differences in job strain, social support at work, and psychological distress. Journal of Occupational Health Psychology 5(4): 428–440.

90.

Vincent

(2011) The emotional labour process: An essay on the economy of feelings. Human Relations 64(10): 1369–1392.

91.

Wenzlaff

Wegner

(2000) Thought suppression. Annual Review of Psychology 51(3): 59–91.

92.

Wharton

(1993) The affective consequences of service work. Work and Occupations 20(2): 205–232.

93.

Wong

Law

(2002) The effects of leader and follower emotional intelligence on performance and attitude: An exploratory study. The Leadership Quarterly 13(3): 243–274.

94.

Wright

Nishii

(2008) Strategic HRM and organizational behavior: Integrating multiple levels of analysis. In: Guest

Paauwe

Wright

(eds) Human Resource Management and Performance: Progress and Prospects. Oxford: Blackwell Publishing.

95.

Xanthopoulou

Bakker

Demerouti

Schaufeli

(2007) The role of personal resources in the job demands-resources model. International Journal of Stress Management 14(2): 121–141.

96.

Zapf

(2002) Emotion work and psychological well-being: A review of the literature and some conceptual considerations. Human Resource Management Review 12(2): 237–268.