Social support theory provides the foundation for understanding supervisor support mechanisms and provides guidance for tools and strategies supervisors can use to bolster employee well-being. Generally speaking, the social support we receive from others in our social networks, including the people with whom we work, is vital to our health and well-being (e.g., Cohen & Wills, 1985) including emotions (Pierce & Quiroz, 2019). Perceptions of workplace social support involve the perception that one is a valued employee and the access to resources that help meet role demands (e.g., Kossek, Pichler, Bodner, & Hammer, 2011); such resources include emotional and instrumental or tangible assistance (e.g.,Cutrona, 1989).According to a recent comprehensive meta-analysis of the multidisciplinary social support literature, social support plays multiple distinct roles in shaping well-being (Bavik et al., 2020). Of most relevance to our investigation is the role of social support as positivity catalyst and negativity buffer (Bavik et al., 2020).

Positivity catalyst describes how the exchange of resources over time fosters high quality social relationships and is built on the social norm of reciprocity that governs all relationships (e.g., Gouldner, 1960). When social support acts as a positivity catalyst, an employee provided with a supportive benefit engenders their own positive behaviour (e.g., employee engagement). While this role of social support as a positivity catalyst is consistent with social exchange theory (Blau, 1964), it also suggests that the provision of support serves as a resource, as described in the conservation of resources (COR) theory (Hobfoll, 1989) and the job demands-resources (JD-R) model (Demerouti, Bakker, Nachreiner, & Schaufeli, 2001). COR theory and the JD-R model articulate how employees are motivated to accumulate and maintain highly valued resources, yielding work and well-being benefits. This view of social support most closely aligns with the direct effect model of social support (Cohen & Wills, 1985), in which positive benefits are derived directly from one’s socially rewarding social roles and embeddedness in a larger social network, and not necessarily in response to a real or potential stressor.

The most dominant perspective of social support roles is that of the negativity buffer (Bavik et al., 2020). In this role, the potential impact of a stressor on one’s health and well-being is a function of one’s appraisal of a perceived stressor and one’s adequacy (or inadequacy) of resources to meet the needs of that stressor (Lazarus & Folkman, 1984). In the workplace context, support can also be viewed as a resource, which can reduce the negative effects of stressors and job demands on well-being, consistent with the JD-R model (Demerouti et al., 2001). This support role is most consistent with the stress-buffering function of social support (Cohen & Wills, 1985). Although received support can be miscarried, support that is given when needed is associated with positive perceptions of support and mental well-being (Melrose, Brown, & Wood, 2015). As a negativity (or stress) buffer, social support can influence the appraisal of a potential stressor, thereby preventing a stress response. Similarly, a key benefit of resource accumulation, as described above, is that it prevents the occurrence of many stressors.

Supervisors are particularly valuable source of support for employees in the workplace, as the support they give can serve as both a positivity catalyst and negativity buffer. Supervisors are able to cultivate a culture of support for employees; for example, workplace programmes encouraging family supportive supervisor behaviours (FSSB) lead to improved employee health, well-being, and work outcomes (Hammer & Perry, 2019). Supervisors also hold positions enabling them to provide critical resources to offset work-related stress (Hammer,Kossek,Zimmerman,& Daniels, 2007). They provide direction on performing the job, provide rewards for performance, and are gatekeepers to top organizational leaders and to accessing organizational programmes and policies that can support employees both in times of need and normalcy.

Consequently, supervisor support benefits employee health and well-being (for reviews, see Crain & Stevens, 2018; Kelloway & Barling, 2010), impacting both positive and negative psychological health outcomes, including stress and strain (Kelloway & Barling, 2010). During times of high need (e.g., organizational crisis), employee positive and negative emotions mediate the relationship between perceptions of supervisor support and cynicism (Cole, Bruch, & Vogel, 2006). Daily supervisor support protects against work–family conflict-associated declines in negative affect (Almeida et al., 2016). Further, FSSB training, which is conceptually related to the present study intervention, has been shown to improve job satisfaction among employees experiencing significant work–family conflict (Hammer et al., 2011); job satisfaction being a facet of subjective well-being, along with emotions (e.g., Diener, 2009; Judge & Locke, 1993). Despite evidence on the benefits of supervisor support (e.g., Hammer et al., 2011; Odle-Dusseau, Hammer, Crain, & Bodner, 2016), little information in the literature instructs supervisors on how to enact supportive behaviours. Recently, research has begun to demonstrate the positive effects of SST interventions that target enactment of supportive behaviours on worker health and well-being outcomes (e.g., Crain et al., 2019; Kossek et al., 2019).

The workplace SST developed for this study targets support provisions that are relevant to all employees, while also incorporating veteran-specific support that targets the civilian workplace reintegration of veterans and active reservists (e.g., National Guard service members). Our approach helps address the call for the inclusion of more diversity of employees in organizational research, including current and former service members (i.e., veterans; Colella et al., 2017), in order to better understand the workplace context for that segment of the workforce. Evidence suggests that challenges in work and family domains may be particularly concerning for reintegrating veterans who report difficulty adjusting to civilian life post-military deployment (44%) and experience strains in family life (48%; Tanielian, Batka, & Meredith, 2017). In addition, veterans face many potential work-related difficulties, including a lack of positive support interactions and stigma in the workplace (Keeling, Kintzle, & Castro, 2018). Many veterans face additional stressors, such as cynicism that carry over from military deployment and issues affecting work and family functioning (Smith, Weisenbach, & Jones, 2018). Thus, workplace supports developed to address work–family issues may be especially beneficial to military veterans’ emotional well-being.

The SST examined in the present study (Hammer et al., 2019) is based on evidence-based FSSB training (Hammer, Demsky, Kossek & Bray, 2016) and targets four types of support behaviours, defined as (1) emotional, or behaviours that demonstrate a worker is cared for and their feelings are being considered; (2) instrumental, or behaviours that help workers manage schedules and solve schedule conflicts; (3) role modelling, or behaviours that show how a supervisor is taking care of their own work/life challenges; and (4) creative work–family management, or behaviours aimed at redesigning work to support conflicting employee work-life demands in a manner that is a ‘win-win’ for both employees and employers. Job performance support training was also included covering topics associated with general supervisor skills that address task accomplishment, including measurement, direction, feedback, coaching, providing resources, and health protection, to complement the FSSB training (Hammer et al., 2016). Our intervention also included information regarding mental and physical health effects that could result from military combat, including clarifying that the incidence of PTSD in the military is lower than public perceptions (Mahar et al., 2017), while offering tips on warning signs, thereby educating civilian supervisors and destigmatizing such experiences.

Thus, we aimed to cultivate greater resources for the veterans through SST leading to improved well-being. The intervention also includes information that contributes to a supportive military culture by highlighting the contributions of service members such as high levels of dedication and strong leadership skills and abilities. Accordingly, the positive emotional enhancement brought about by the SST serves as a positivity catalyst and negativity buffer.

Positive and negative emotions are characterized as feeling pleasant or unpleasant across lay and academic perspectives (e.g., Beedie, Terry, & Lane, 2005; Davidson, 1994; Russell & Barrett, 1999) and comprise core components of subjective well-being, with the understanding that optimal well-being is represented by a high positive emotion to negative emotion ratio (Diener, 2000). Positive and negative emotions ebb and flow with everyday positive and negative events (Gable, Reis, & Elliot, 2000) including work stressors (e.g., Bolger, DeLongis, Kessler, & Schilling, 1989). Emotions and cognitions are thought to be primary responses to contextual factors (Lazarus, 1982; Zajonc, 1980), preceding behavioural outcomes. Positive and negative emotions function independently (Diener & Emmons, 1984), meaning individuals can experience both low negative and positive emotions simultaneously. Emotional well-being has most often been represented by a summary or composite of positive and of negative emotion scores to represent a general sense of pleasant or unpleasant emotional experience (e.g., Diener, 2000).

Because emotional states are dynamic and changeable (e.g., Larsen & Kasimatis, 1990), as are the relationships between day-to-day stressors and emotions, it is critical to capture these variables as they are experienced, as opposed to retrospectively (e.g., Tennen, Affleck, Armeli, & Carney, 2000). Importantly, a recent rigorous evaluation of global perceptions of well-being (as in traditional self-report surveys) revealed that global perceptions overestimate daily experiences of well-being when compared to daily reports of well-being (Newman, Schwarz, & Stone, 2020). Beyond reflecting recent experience, global reports also incorporate peak experiences and frequently occurring experiences. Newman et al. (2020) highlight the benefits of assessing well-being via aggregated daily emotions or ‘momentary averages’ of emotions as a more accurate means for assessing well-being as it is experienced in one’s daily life. We propose that this more accurate methodological approach should be incorporated into the context of measuring SST effectiveness. Thus, using a randomized controlled trial (RCT) design with 35 organizations, where organization was the unit of randomization, we examined the effects of an SST intervention on employee emotions experienced during the 32-day pre-intervention period and then again 6 months later (post-intervention) in a second 32-day diary. Our study offers a beneficial method for assessing training effectiveness on employee’s daily emotional experience.

One additional benefit to employing daily emotion reports is that it facilitates the differentiation of positive and negative emotions into discrete emotions based on the pleasant/unpleasant dimension and also the physiological arousal or activation associated with each, as described by the Circumplex Model of Emotion (Larsen & Diener, 1992). Bored emotion, for example, is considered the opposite of enthusiastic or lively emotions. Similarly, unactivated (or deactivated) unpleasant is considered the opposite of activated pleasant emotion, wherein each emotion type (e.g., unactivated pleasant, activated pleasant, unactivated negative, activated negative) is theoretically associated with distinct behavioural effects. FMRI research has provided evidence that valence, or the positive–negative dimension of emotion, is associated with a distinct neural substrate in the brain from the arousal dimension (Posner et al., 2009). Weiss and Cropanzano (1996) noted the potential value for considering discrete emotional states in workplace studies. A recent study demonstrated clear relationships between perceived supervisor support and both positive and negative emotions in the expected directions (Cole et al., 2006). Further, Lazarus (2006) articulated the importance of examining discrete emotions, which can hinder or enhance our interpersonal relationships and are a central feature of our ongoing relationships, in addition to general summaries (composite scores) of positive and negative emotions via daily diary reports. We proposed to examine emotions from both a dimensional view, which presumes there is an underlying commonality according to valence (i.e., positive vs. negative), as well as a discrete emotions view, whereby emotions ‘are assumed to be psychologically meaningful units that have unique organizational significance’ (Diener, Thapa, & Tay, 2020, p. 453).

Despite previous work demonstrating the beneficial effects of SST on employee health and work outcomes (Hammer et al., 2019) there is no published research establishing the effects of such an intervention on daily emotions. Considering that emotion is a core component of well-being, and the above literature ties supportive supervision to emotional well-being, we hypothesized that:

H 1. The SST would have a significant effect on veteran employee emotions, specifically improving positive emotions and decreasing negative emotions 6 months post-intervention both in terms of composite (average) and discrete positive and negative emotions.

Negative affective disturbance (e.g., loneliness, irritability) is a defining feature of PTSD, as are difficulties enjoying positive experiences (APA; American Psychiatric Association, 2013). According to the DSM-V, those with PTSD can experience ‘persistent negative emotional states’ as wellas ‘persistent inability to experience positive emotions’ (pp. 271–272). Consequently, veterans who are experiencing greater PTSD may experience greater negative emotions in their daily lives than those who do not have PTSD and thus may be in particular need of supportive intervention to help reduce such distress. Experimental research has shown that following exposure to emotion words in the laboratory, combat veterans with PTSD experienced less positive emotion (i.e., calm, pleased) and greater negative emotion than combat veterans without PTSD (Amdur, Larsen, & Liberzon, 2000). Yet, social support should buffer the effects of negative stimuli on individual well-being (Cohen& Wills, 1985). Correspondingly, post-deployment social support has been shown to reduce PTSD severity for veterans (Han et al., 2014). We proposed that although PTSD screening would have a positive relationship with daily negative emotions, the intervention would buffer the relationship between PTSD and daily negative emotions. It is less clear whether the intervention would boost positive emotions among those with PTSD, given the evidence of anhedonia or emotional numbing reported in (male) veterans with PTSD in terms of positive emotion (e.g., Kashdan, Elhai, & Frueh, 2007). Consequently, we did not propose a directional hypothesis for this test.

As previously reported (McCabe, Mohr, Hammer & Carlson, 2019), the estimate of positive PTSD screens from our survey sample was 17%. Such PTSD screening and cutoffs are defined based on empirically based levels that identify those veterans who are likely to be experiencing PTSD symptoms and in need of early intervention (Bliese et al., 2008). The use of screening tools with empirically validated cutoff values, such as the PTSD screening, has been widely encouraged as a way of reaching more members of the general public who may need support or intervention. Researchers have noted this as a significant development for public mental health, as these tools can be quickly and easily administered without a clinician and have been instituted broadly including in the military (e.g., Bliese, Adler, & Castro, 2011). The present study affords the opportunity to consider whether those above the PTSD screening cutoff receive particular benefit from the supervisor support intervention, relative to those below the PTSD screening cutoff.

H 2. PTSD scores would moderate the effects of the SST on employee negative emotions such that the intervention has a stronger (decreasing) effect on negative emotions among veteran employees with positive screens for PTSD, compared to those with negative PTSD screens.

RQ1: Do PTSD scores moderate the effect of the SST on employee positive emotions such that the intervention has a stronger (increasing) effect on positive emotions among veteran employees with negative screens for PTSD, compared to those with positive PTSD screens?

This study affirms the importance of workplace supervisor support provision to boost veteran affective experiences. The study contributes to workplace positive psychology intervention research by considering the effect of an SST on employees’ well-being in the form of daily emotions, which can shed light on positive shifts in one’s emotions in the immediate periods pre- to post-supervisor support intervention. Such information is qualitatively different from the information gleaned from global assessments of well-being (e.g., Newman et al., 2020). This study extends evidence on the effectiveness of SST to the domain of veteran emotional well-being as they transition into and thrive in the civilian workplace.

The SST intervention was demonstrated to have a significant effect on facets of emotional well-being. Although there were no direct effects of the intervention on composite positive and negative emotion averages across the 32-day diaries, pre- and post-intervention, the intervention improved relaxed and quiet emotions, whereby veterans in the intervention condition reported significantly higher 6-month relaxed and quiet emotions than those in the control condition, which represents greater calmness or ‘serenity’ (Watson et al., 1988) or lower arousal (e.g., Feldman, 1995). An inspection of condition means suggests the intervention may have stabilized or increased those emotions for intervention condition veterans (relative to the control condition which appeared to show decrements over time). The significant moderating effect of positive PTSD screening suggests that the intervention improved composite positive emotion at 6 months among those who were screened as PTSD negative (i.e., most participants). For those screened positive for PTSD, results suggest the intervention was effective at reducing composite negative emotion at 6-month follow-up. Because of the small sample of PTSD-positive participants, these results should be interpreted with caution.

Our results revealed the insight gleaned from studying discrete emotions, whereby the supervisor intervention was effective for all veterans in promoting calm and quiet emotions, but not necessarily the more activated positive emotions such as happy emotion. Such findings may have important health outcomes, given meta-analysis reports showing that unactivated positive emotion induction (e.g., calm, relaxed) was related to improved immune and neuroendocrine functioning (Pressman & Cohen, 2005). There is also evidence of work performance-related outcomes, whereby Bissing-Olson, Iyer, Fielding, and Zacher (2013) found that unactivated emotions (i.e., calm, relaxed and content) were uniquely linked to a greater likelihood that employees engaged in work tasks in an environmentally friendly way.

Composite positive emotion, which includes activated positive emotions like happy, was increased among veterans who did not screen positive for PTSD (i.e., most veterans). According to Seo, Barrett, and Bartunek (2004), positive affect is associated with greater focus on goal attainment, generativity, and intensity and persistence towards goal attainment. Seo et al. also argue that activated, but not deactivated (unactivated) affect enhances effort, though they do not differentiate deactivated positive (e.g., relaxed) from negative (e.g., lonely, bored), which our results indicate would be important. We encourage future work on the influence of employee emotional well-being, in terms of activated and unactivated positive and negative emotions and work-related outcomes. Researchers may also consider differentiating emotions at work from emotions at home (e.g., Sonnentag & Grant, 2012) to better capture affective experiences in work and non-work (family) domains.

For veterans who screened positive for PTSD, SST was associated with a reduction in negative emotions, but activated positive emotions like happiness did not increase. These results need to be replicated with a larger sample of veterans with positive PTSD screens before conclusions can be drawn. A preliminary interpretation could suggest a particular challenge for them in fully enjoying the benefits of the supervisor intervention – potentially due to consequences of PTSD, which impairs the ability to enjoy positive experiences (APA, 2013). Conversely, our intervention did not reduce negative emotions, except for among the PTSD-positive-screening veterans. Given that individuals who experience PTSD are more likely to have stronger associations with negative affect (e.g., downward spirals of emotions; for overview see Garland et al., 2010), this finding may indicate that those downward spirals were weakened by the SST. This outcome is consistent with the negativity buffering model of social support (Bavik et al., 2020), in which veterans with positive PTSD screens may benefit from enhanced supervisor support resulting in a reduction of negative emotions. Veterans without PTSD experienced lower rates of negative emotions in our study (e.g., mean negative emotion ranged from 1.39 to 1.43 depending on condition, on a 1-to-5 scale), as with other non-clinical populations (e.g., Crawford & Henry, 2004), wherein negative emotion may be too low to necessitate intervention.

Observed increases in positive emotions resulting from the SST are consistent with the positivity catalyst role of social support (Bavik et al., 2020), in which support leads to generation of resources. Theoretically, the positive emotions engendered by the intervention may reflect improved relationships with the supervisor, which should lead to positive employee behaviour (e.g., employee engagement) via reciprocity (Bavik et al., 2020). Accordingly, positive emotions as a component of employee well-being are fundamental to employee engagement (Robertson & Cooper, 2010). The intervention was also designed to alleviate or buffer employee stress, described as a negativity buffer (Bavik et al., 2020). To probe further, future work should examine how supervisor social support influences responses to employee stress in daily work interactions, for example as described in the JD-R model (Demerouti et al., 2001). As employees perceive stressor demands exceeding capacity or resources, supervisor support should enhance resources or offset demands to alleviate stress. Future work should also consider the potential for supervisor support to alter the appraisal of a potential stressor, intervening early in the stress response, which would also be consistent with stress and negativity buffering models. This study presents a methodological contribution to the literature by demonstrating the value of daily diary methods to capture ecological momentary emotion assessments before and after the intervention to facilitate the assessment of intervention effectiveness. The combination of multi-wave daily assessment and RCTs is rare. Although the length of time between daily diary intervals (6 months) might seem too long to detect intervention effects, other non-daily intervention analyses emerging from our larger study reveal that effects not present at a 3-month interval were significant at 9-month follow-up (Hammer et al., 2019). As static surveys, the 3- and 9-month surveys were not designed to measure the everyday emotions captured in the 6-month daily diaries. The 6-month daily diary, then, may have captured an emerging effect on emotions of the SST that was building over time. One potential explanation for the delayed treatment effect is that supervisors need time to practice in order to provide effective support.

Our study was robust in the length of the daily surveys, spanning up to 32 days of reports in each wave for a total of 64 possible reports. Daily diaries are particularly useful in measuring the ebb and flow of emotions as they occur while minimizing concerns about retrospective bias (e.g., Newman et al., 2020; Reis & Gable, 2000). Our approach provides a solid foundation for launching more complex and sophisticated inquiries into employee emotional experiences. We look forward to future work examining more complex within-person changes over time as a function of the intervention, modelling processes specified by positivity catalyst and negativity buffer models (Bavik et al., 2020).