“Your help isn’t helping me!” Unhelpful workplace social support, strain, and the role of individual differences

Abstract

While the reception of social support at work is generally considered a net positive for employees, researchers have identified that particular kinds of social support, such as unhelpful workplace social support (UWSS), tend to evoke stress and contribute to strain for recipients. Although (Gray et al. Work and Stress, 34(4), 359–385, 2020), when validating the novel UWSS measure, uncovered relations between UWSS and various outcomes, more research is needed to further understand the impacts of UWSS. Furthermore, the extant social support literature is currently lacking in its understanding of how individual differences strengthen or weaken the relations such support has with strain. Drawing from the Theory of Stress as Offense to Self (Semmer et al. Occupational Health Science, 3(3), 205–238. 10.1007/s41542-019-00041-5, 2019), we, through two studies (N₁ = 203, N₂ = 277), further explore the relations of UWSS, focusing on behavioral and psychological strain, and examine how these relations are influenced by relevant individual differences (e.g., Big Five traits). Results from our first study replicate key findings from (Gray et al. Work and Stress, 34(4), 359–385, 2020), providing additional validity evidence for the novel measure of UWSS, and demonstrate that UWSS is related to various types of behavioral strain. Our second study shows that the strength of these deleterious relations varies based on characteristics of the recipient of UWSS. Altogether, the present research contributes to the literature on social support as a stressor by elucidating further the effects of UWSS, and, perhaps more importantly, for whom UWSS is particularly deleterious.

Social support, which broadly refers to “psychological or material resources that are provided to a focal individual by partners in some form of social relationship” (Jolly et al., 2020, p. 229), is a dynamic, exchange-based process widely researched across psychological disciplines (e.g., French et al., 2018; Schwarzer & Leppin, 1989). Generally, the reception of social support is considered a net positive for individuals, both in a work context and in general. With regard to work, meta-analytic research demonstrates that the reception of both emotional (e.g., one colleague being allowed to vent to another about a personal problem) and instrumental (e.g., a colleague helping another with a task-related problem) support share positive relations with desirable job attitudes (e.g., job satisfaction, organizational commitment), negative relations with stressors (e.g., role stress) and strains (e.g., burnout), and can often mitigate the relations between stressors and strains (e.g., Kossek et al., 2011; Mathieu et al., 2019). Altogether, the reception of social support is generally a resource for employees, one which enriches the psychological context of work and erodes the effects of stress.

Recent research, however, has helped to elucidate where and when social support exchanges—even the reception of social support—have the potential to yield negative effects for employees (e.g., Deelstra et al., 2003; Gallagher et al., 2021; Hughes, 2021). The present research focuses on broadening our understanding of one such example of the “dark side” of social support: the reception of unhelpful workplace social support (UWSS). Defined as “any action taken by a supervisor and/or colleague that the recipient believes was intended to benefit him or her but is perceived as unhelpful or harmful” (Gray et al., 2020, p. 376), UWSS is a novel framing of a type of social support exchange likely to be lamented by employees—that which amplifies and contributes to stress. Indeed, when validating their measure of UWSS, Gray et al. (2020) found that UWSS was related to numerous kinds of employee strain, ranging from emotional exhaustion to negative mood, in addition to exacerbating the impacts of stressors (e.g., organizational constraints) on strains (e.g., physical symptoms). Their study echoes and extends past research on social support as a potential stressor (e.g., Beehr et al., 2010) and provides scholars with a psychometrically valid measure to continue to investigate such deleterious support exchanges.

Empirical findings, although few and far between, suggest that UWSS is a stressor for employees (e.g., Beehr et al., 2010; Gray et al., 2020). What remains unclear, however, is for whom these exchanges are most deleterious for. That is, the existing literature on unhelpful social support at work is void of knowledge surrounding the impact of individual differences on relations between such support and strain. It seems intuitive that UWSS would be more salient, and therefore more damaging, to certain kinds of employees, making individual differences of heightened importance for this kind of negative social exchange. Indeed, Bowling & Jex (2013) suggest that a better understanding of the relations between stressors and strains can be obtained by exploring the influencing role of individual differences. Moreover, Semmer et al. (2019) note, in their updating of the Theory of Stress as Offense to Self (SOS model; Semmer et al., 2019), that additional research is needed to determine the impact of individual differences on boosts and, importantly for UWSS, threats to the self.

To address this gap in the literature on workplace social support as a stressor and build upon extant research, the present study had several goals. Drawing from the SOS model (Semmer et al., 2019), we first, using Gray et al.’s (2020) novel measure of UWSS, establish relations between UWSS and behavioral strain (i.e., counterproductive work behavior, a noted form of behavioral strain, e.g., Fox et al., 2001; Penney & Spector, 2005; Welbourne & Sariol, 2016), as well as replicate strong main effects from previous research. In addition, we delineate UWSS’ place among related negative workplace exchanges (e.g., interpersonal conflict, incivility) in the prediction of well-being and attitudinal outcomes. Finally, we examine relevant individual differences (e.g., Big Five traits) that may work to strengthen or buffer the associations between UWSS and strain. Taken together, the present research contributes to the nascent literature on workplace social support as a stressor by establishing novel UWSS relations (e.g., behavioral strain) and focusing on how individual differences influence these and other important (e.g., psychological strain) relations. These findings contribute to theoretical models of such support by linking these exchanges to behavioral strain and establishing for whom such exchanges are particularly impactful. Furthermore, we continue to build upon the needed research surrounding individual differences and stressor-strain frameworks (Bowling & Jex, 2013).

When Support Can be Stressful

Although social support has long been discussed as a resource that buffers or mitigates the effects of stress (e.g., Cobb, 1976), there are numerous instances where social support can contribute to and magnify stress. Research on the stressful impacts of workplace social support can be traced back several decades to when scholars began to uncover what are known as “reverse buffering effects” of social support (e.g., Kaufmann & Beehr, 1986, 1989). In instances of reverse buffering, social support will strengthen, as opposed to alleviate, the relations between workplace stressors and consequent strain. For example, a study from LaRocco et al. (1980) found that supervisor support enhanced the relation between workload stress and somatic complaints and workload dissatisfaction. The authors suggested that support from supervisors has the potential to make unappealing elements of the work context more salient to recipients, thus amplifying the impact of stress caused by workload. In the years following, other authors put forth alternative explanations for reverse buffering effects of social support. Chisholm et al. (1986) and Karasek et al. (1982) found reverse buffering effects in samples of low stress employees, proposing that social support, while a net positive for work units overall, may add extra burden to those who are not “in need” of such resources.

With the goal of clarifying precisely when social support is stressful, Beehr et al. (2010), drawing from the research on the reverse buffering effects of social support, found that social support was appraised as stressful if: a) it made draining aspects of the environment more salient (e.g., consoling a dissatisfied colleague while validating and expanding their perception of the work environment as unfulfilling), b) was not wanted by the recipient (e.g., helping another when not asked), or c) made the recipient feel incompetent (e.g., helping another by doing their task for them). The authors found, in these instances, social support will contribute to workplace strain, such as emotional exhaustion and physical symptoms. This study, in tandem with prior research on reverse buffering effects, helped establish the foundation for Gray et al.’s (2020) conceptualization of UWSS.

Unhelpful Workplace Social Support

Borne out of the aforementioned research on the receipt of social support as a catalyst for stress (e.g., Beehr et al., 2010), Gray et al. (2020), acknowledging a gap in the stressful social support literature, sought to construct a measure that captures social support exchanges that contribute to and amplify the impact of stress on various types of strain. Using qualitative data as a guide during their measurement development process, the authors verified a seven-factor model of unhelpful workplace social support (UWSS) exchanges from the perspective of the recipient of such support. The types of UWSS noted by the authors are as follows: critical social support, imposing social support, partial social support, undependable social support, shortsighted social support, uncomforting social support, and conflicting social support (Gray et al., 2020).

It is important to note, UWSS is not the absence of helpful social support; as demonstrated by Gray et al. (2020), UWSS is a unique exchange where support is provided to recipients, but in ways that are not actually helpful or are draining (e.g., rude or uncourteous). Also of importance is the distinction between UWSS and other negative social exchanges, such as incivility and interpersonal conflict. Incivility, defined as “rude and discourteous behavior that lacks a clear intent to harm” (Cortina et al., 2013, p. 1580), and interpersonal conflict (i.e., conflict between colleagues, ranging from minor disagreements all the way to physical assault; Spector & Jex, 1998) are two of the more researched negative social exchanges in the organizational sciences, contributing to a bevy of negative outcomes on behalf of employees (e.g., emotional exhaustion) and organizations (e.g., turnover intentions) (e.g., Nixon et al., 2011; Yao et al., 2021). While UWSS appears similar to these exchanges in nomological network, UWSS is distinct in that instrumental, emotional, appraisal, or informational resources (House, 1981) are provided to the recipient from a provider—just in a manner that is stressful. In contrast, resources are not necessarily provided in instances of incivility and interpersonal conflict. For example, while critical social support is akin to incivility and interpersonal conflict in terms of display valence (i.e., all are rude or hostile), recipients of critical social support are still receiving resources of value from providers. Even in instances where incorrect or faulty resources are provided (e.g., conflicting social support, undependable social support), there is an effort on the part of the provider to bestow resources to a recipient—an effort not made when simply behaving uncivil or hostile towards others.

Drawing from prior research on social support as a stressor (e.g., Beehr et al., 2010; Deelstra et al., 2003), Gray et al. (2020) contended that UWSS would have significant negative impacts for recipients in that these exchanges may generate negative affect and negative self-evaluations as well as impede their ability to complete work tasks. In their exploration of main effects, the authors found preliminary support for these notions, with UWSS sharing positive relations with work-related burnout, interpersonal conflict, job-related negative affect, organizational frustration, organizational constraints, and negative mood, and negative relations with coworker satisfaction, organizational self-esteem, and helpful social support. Before extending research on UWSS, given the novelty of the construct and its measurement, we attempt to replicate strong relations (r > .50) observed by Gray et al. (2020); in particular, relations between UWSS, job-related negative affect, emotional exhaustion, interpersonal conflict, and coworker satisfaction. The authors expected UWSS to relate to these outcomes due to UWSS’ potential for self-esteem threat (e.g., Deelstra et al., 2003), theoretical rationale that we expand upon in the following section.

• Hypothesis 1a-d: UWSS will be positively related to a) job-related negative affect, b) emotional exhaustion, and c) interpersonal conflict, and negatively related to d) coworker satisfaction.

Unhelpful Workplace Social Support as a Threat to the Self

To meaningfully contribute to the nascent literature on workplace social support as a stressor beyond Gray et al.’s (2020) findings, we draw from the SOS model (Semmer et al., 2019) to develop hypotheses focused on UWSS and behavioral strain, as well as on the role individual differences may play in influencing the relations between UWSS and key strain. The SOS model is useful for constructing hypotheses surrounding the relations UWSS shares with strain, as both functional and dysfunctional social support are constructs integral to the theoretical model.

The SOS model of stress positions self-esteem as a core contributor to stress and well-being. More specifically, the SOS model contends that there are two kinds of self-esteem: personal self-esteem, which concerns one’s self-evaluation in terms of aspired (e.g., competence, attractiveness) and intrinsic (e.g., dignity) qualities, and social self-esteem, which concerns the degree to which one feels “esteemed, acknowledged, and appreciated by significant others” (Semmer et al., 2019, p. 208), and is strongly dependent on one’s social environment. Threats to personal or social self-esteem (e.g., interpersonal stressors) engender physiological, behavioral, and psychological strain in focal individuals, while boosts to self-esteem (e.g., appreciation, social support) help to foster well-being.

We contend that UWSS is likely to be a significant threat to recipients’ personal and social self-esteem in spite of the fact that UWSS is social support, an exchange that often boosts self-esteem (Semmer et al., 2019). As described earlier, although resources are provided (or an attempt is made to provide resources) during instances of UWSS (marking a distinction from other negative exchanges), they are provided in a way that engenders stress—likely due to such support being perceived as a threat to both kinds of self-esteem. For example, a recipient of critical social support, while acknowledging that resources were indeed provided, may feel that the provider of the support does not respect them due to the providers’ rudeness (threat to social self-esteem), which may deflate one’s self-evaluations of competence or worthiness (threat to personal self-esteem). Similarly, a recipient of undependable social support may, while acknowledging that an attempt was made to provide instrumental resources, feel worse about their level of competence once the exchange has concluded (threat to personal self-esteem), as now they must worry about correcting the mistakes of another (which they may or may not know how to do) while further attempting to understand the task in question.

The SOS model posits that behavioral strain, in addition to psychological and physiological strain, is a likely outcome of self-esteem threat (Semmer et al., 2019). Behavioral strains, such as counterproductive work behavior (CWB; i.e., voluntary behaviors that harm the organization, its members, or its stakeholders; Marcus et al., 2016), are often spurred on by interpersonal stressors (threats to self-esteem), such as interpersonal conflict or incivility (Berry et al., 2012; Han et al., 2020). CWB is a family of behaviors of heightened importance for organizations as it is estimated that 33 to 75% of all employees engage in CWB at some point in their career (Robinson & Bennett, 1995). Further, in the United States alone, it is estimated that 2.8 million productive workdays are lost due to absenteeism (Ones & Dilchert, 2013) and theft and fraud cost businesses up to $50 billion annually (Coffin, 2003).

It is important to note that employees do not engage in such harmful acts in a vacuum. Organizational scientists have concluded that, in general, employees engage in CWB as a reaction to negative environmental and emotional stimuli (e.g., Neuman & Baron, 2005). To that end, research drawing from the SOS model (Semmer et al., 2019) has found that CWB is a probable response to threats to self-esteem (e.g., Semmer et al., 2010; Zhou et al., 2018). Given our positioning of UWSS as a threat to the self, building from this research, we anticipate UWSS will relate to various forms of CWB such as production deviance (e.g., purposefully working slowly), withdrawal (e.g., arriving late to work without permission), and abuse (e.g., verbally threatening someone at work). These types of CWB have been identified as forms of emotion-focused coping an employee may leverage to deal with the stress engendered by negative social exchanges (e.g., Krischer et al., 2010; Shoss et al., 2016).

• Hypothesis 2a-c: UWSS will be positively related to a) abuse, b) withdrawal, and c) production deviance.

UWSS Among Related Exchanges

Before transitioning into our hypothesis development surrounding individual differences and UWSS, it is important to discuss UWSS’ place among related constructs. UWSS seems at home among a variety of interpersonal workplace stressors, such as incivility and interpersonal conflict, previously discussed. Both incivility and interpersonal conflict are relatively well understood by organizational scholars (e.g., Hershcovis et al., 2007; Schilpzand et al., 2016). What is unclear, however, is how UWSS fits in with these constructs: does UWSS have explanatory power in the prediction of employee outcomes over incivility and interpersonal conflict? We will take an exploratory approach to this question due to the novelty of the UWSS construct, comparing UWSS with incivility and interpersonal conflict in the prediction of well-being (i.e., emotional exhaustion) and attitudinal (i.e., satisfaction with coworkers, job-related negative affect) outcomes.

• Research Question: Does UWSS have incremental validity over incivility and interpersonal conflict in the prediction of well-being and attitudinal outcomes?

The Moderating Role of Individual Differences

When concluding their update of the SOS model, Semmer et al. (2019) note that “it is unlikely that threats and boosts to the self are the same for everyone” (p. 226), and encourage future research to explore the role of individual differences as boundary conditions for the impacts of boosts and threats to the self. We answer this call and add to the literature by investigating the role of individual differences in influencing the UWSS-CWB and UWSS-emotional exhaustion relations. We focus on these forms of behavioral and psychological strain, respectively, as they are of heightened importance for organizations. CWB is, as mentioned previously, particularly costly (e.g., U.S. Chamber of Commerce, 2002); similarly, emotional exhaustion contributes to massive revenue loss for organizations: a survey conducted by the American Psychological Association in 2017 found that U.S. businesses lose up to 300 billion dollars yearly because of workplace stress, a precursor to emotional exhaustion and strain.

We contend that the Big Five traits of extraversion, agreeableness, conscientiousness, and neuroticism (McCrae & Costa, 1987),1 and the social exchange disposition of negative reciprocity beliefs (Eisenberger et al., 2004) are likely important in determining the extent to which an interpersonal stressor such as UWSS is appraised as a threat to personal or social self-esteem. These individual differences have been found to influence interpersonal stressor-strain relations (e.g., Gallagher & Hughes, 2020; Welbourne et al., 2020), and may be of importance here.

Neuroticism

Neurotic employees can best be described as worrisome, insecure, temperamental, and anxious (McCrae & Costa, 1987). Individuals who are high in neuroticism often display consistent negative reactions to stressors (McCrae & Costa, 1987), tend to over-think events or stimuli that could be considered threatening (McCrae & Costa, 1987), and usually maladaptively cope with stress (e.g., Gross, 1998). In addition, neuroticism is also a crucial moderator for stressor-strain relations (e.g., Kennedy & Hughes, 2004; Mohiyeddini et al., 2015; Welbourne et al., 2020; Zhou et al., 2014a, b).

Neurotic employees are likely to appraise UWSS as a significant threat to personal and social self-esteem. This is due to their general tendency to perceive (or identify behaviors as) stressors in their environment (e.g., Gallagher, 1990; Sliter et al., 2015): neurotic employees may view UWSS as an attack on their competence (if the social support is unwanted or controlling), their time (if the social support is incomplete or incorrect), or their character (if the social support is rude). Appraising UWSS in such ways would likely contribute to employees feeling poorly about themselves (personal self-esteem) and the way others perceive them (social self-esteem). Neurotic employees, then, are likely to hyper fixate on such appraisals and events, boosting the relations between UWSS and emotional exhaustion. Moreover, the relations between UWSS and behavioral strains, such as CWB, are likely to be stronger for those who are more neurotic. This is because, as previously described, neurotic employees tend to maladaptively cope with stress as a result of feeling overwhelmed (e.g., Gross, 1998).

Agreeableness

Agreeable employees can be described as selfless, trusting, helpful, and forgiving (McCrae & Costa, 1987). They tend to cultivate and utilize social support processes (e.g., Zellars & Perrewé, 2001), meaning that these individuals often have social resources available to buffer the impacts of stressors (e.g., Swickert et al., 2010). Thus, it is perhaps not surprising that agreeableness tends to reduce the likelihood of maladaptive coping in response to workplace stressors (e.g., Yang & Diefendorff, 2009; Zhou et al., 2014a, b).

Important for UWSS, though, research has shown that agreeableness tends to negatively relate with primary appraisal (i.e., whether a person perceives a situation or stimuli to be threatening; Lazarus & Folkman, 1984) of stressors. That is, agreeable people tend to give others “the benefit of the doubt”; this phenomenon was observed by Sliter et al. (2015). When agreeable employees encounter UWSS, they may be more likely to forgive or disregard any rudeness accompanying the support and chalk up incomplete, partial, or incorrect support as simply a mistake on behalf of the provider. That is, agreeable employees are likely to not perceive UWSS as a threat to personal or social self-esteem—meaning that psychological or behavioral strain outcomes are less likely.

Extraversion

Extraverted employees are those who are sociable, friendly, and talkative (McCrae & Costa, 1987). These individuals like to invest time and energy into social interactions, or events that otherwise attract social attention (Ashton et al., 2002). Moreover, extraverted individuals tend to be high in positive affectivity (e.g., Meyer & Shack, 1989), meaning that they are often jovial, confident, and optimistic. The impacts of extraversion on the stressor-strain process have been researched for some time. Importantly, Milam et al. (2009), drawing from research on traits and temperament (e.g., Watson & Clark, 1992), suggested that extraverted employees may be less likely to perceive minor breaches in social norms as a result of their often positive outlook; indeed, other authors, such as Sliter et al. (2015) have made similar propositions. Both Milam et al. (2009) and Sliter et al. (2015), though, did not find support for these predictions, as their results surrounding incivility perceptions and extraversion were null.

It is important to note that these studies focused on incivility. As described previously, incivility is meaningfully distinct from UWSS in that no resources are (or are attempted to be) provided. Incivility, characterized by ambiguity and low-level hostility, may not activate elements of positive affect potentially necessary to give instigators the “benefit of the doubt” (akin to agreeableness). It may be that extraverted employees are, as a result of their positive and social disposition, less likely to emphasize the negative elements of UWSS, though, since—opposed to incivility—resources are (or are attempted to be) provided during UWSS exchanges. Instead, extraverted employees may focus on the positive element of the UWSS exchange (the attempted support), meaning that UWSS is not likely to be perceived as a threat to personal or social self-esteem for these individuals (i.e., psychological and behavioral strain outcomes are less likely).

Conscientiousness

Conscientious employees are those who are orderly, hard-working, thorough, and reliable (McCrae & Costa, 1987). As a result of such qualities, conscientious employees are often better performers, with higher reported task and citizenship performance (e.g., Debusscher et al., 2017; Mount & Barrick, 1998). Furthermore, they are often firm rule followers who are sensitive to violations of formal expectations and norms (McCrae & Costa, 1987).

We anticipate, building from the SOS model (Semmer et al., 2019) and the notion that UWSS is perceived as a threat to personal and social self-esteem, that conscientiousness will differentially impact the UWSS-emotional exhaustion and UWSS-CWB relations. Conscientious employees, in addition to being firm rule followers, tend to hold others to the same standard of rule adherence (McCrae & Costa, 1987). Consequently, these employees are typically more sensitive to violations of social norms during interpersonal exchanges (Coyne et al., 2000; Sliter et al., 2015). UWSS represents a family of exchanges that can be rude, passive aggressive, or irritating (e.g., incomplete, incorrect, or conflicting support), all of which would violate the “good colleague” norm inherent to most workplaces (Stryker & Vryan, 2006). As a result, we anticipate UWSS—esteem-threatening exchanges that violate interpersonal norms in most work settings—to be particularly draining for conscientious employees, with conscientiousness strengthening the UWSS-emotional exhaustion relation. Conversely, we expect conscientiousness to weaken the UWSS-CWB relation, even if conscientious employees are more likely to detect and appraise UWSS as a threat to the self. This is because conscientious employees tend to avoid emotion-focused coping behaviors, such as CWB (e.g., Krischer et al., 2010), opting instead to engage in problem-focused coping (e.g., Bartley & Roesch, 2011). Even if conscientious employees are, as suggested by previous research (e.g., Sliter et al., 2015), more likely to detect norm violations (inherent to UWSS), they may be more likely regulate their behavior, address any concerns surrounding the UWSS exchange with the provider of the support, and avoid lashing out at others or the organization as a result of threats to the self generated by such exchanges.

Negative Reciprocity

Reciprocity is a key component of social exchanges (Cropanzano & Mitchell, 2005) which revolves around the idea that actions from a distributing party create obligations for and prompt actions from a receiving party. People who believe that negative actions should be returned in kind are described as possessing negative reciprocity beliefs (Eisenberger et al., 2004). In essence, negative reciprocity beliefs detail the possession of “an eye for an eye, tooth for a tooth” mentality. To provide an organizational example, an employee high in negative reciprocity beliefs who experiences mistreatment from a focal employee is likely to return mistreatment to “get even with” the focal employee.

Although not yet assessed in an SOS framework, we contend that those higher in negative reciprocity beliefs are likely to perceive UWSS as a salient threat to personal and social self-esteem. Prior research has found that those who hold negative reciprocity beliefs are, perhaps intuitively, hot-headed, and quick to attribute hostility when confronted with negative events (e.g., Hoobler & Brass, 2006). It is likely that social support that is not actually helpful, and is often rude, incomplete, or ambiguous, is more likely to be viewed as hostile, and a threat to the self by those with such beliefs. If UWSS is perceived as a threat to personal and social self-esteem, recipients who hold negative reciprocity beliefs may feel that such support is worthy of “getting even” over, specifically in the form of abuse. We feel that negative reciprocity beliefs are likely to strengthen the UWSS-abuse relation in particular, as UWSS is an interpersonal stressor, and abusive behaviors are hostile, interpersonally targeted behaviors. That is, people higher in negative reciprocity beliefs may be more likely to return perceived interpersonal mistreatment with acts of interpersonal mistreatment. This pattern of returned interpersonal mistreatment has been observed in a number of studies (e.g., Gallagher & Hughes, 2020; Wu et al., 2014).

In sum, the following individual difference related hypotheses are proposed:

• Hypothesis 3a-d. UWSS will interact with a) neuroticism, b) conscientiousness, c) agreeableness, and d) extraversion to predict emotional exhaustion, such that higher levels of neuroticism and conscientiousness will strengthen the relation between UWSS and emotional exhaustion, and higher levels of agreeableness and extraversion and will buffer the relation between UWSS and emotional exhaustion.

• Hypothesis 4a-d. UWSS will interact with a) neuroticism, b) conscientiousness, c) agreeableness, and d) extraversion to predict production deviance, such that higher levels of neuroticism will strengthen the relation between UWSS and production deviance, and higher levels of conscientiousness, agreeableness, and extraversion will buffer the relation between UWSS and production deviance.

• Hypothesis 5a-d. UWSS will interact with a) neuroticism, b) conscientiousness, c) agreeableness, and d) extraversion to predict withdrawal, such that higher levels of neuroticism will strengthen the relation between UWSS and withdrawal, and higher levels of conscientiousness, agreeableness, and extraversion will buffer the relation between UWSS and withdrawal.

• Hypothesis 6a-e. UWSS will interact with a) neuroticism, b) conscientiousness, c) agreeableness, d) extraversion, and e) negative reciprocity beliefs to predict abuse, such that higher levels of neuroticism and negative reciprocity beliefs will strengthen the relation between UWSS and abuse, and higher levels of conscientiousness, agreeableness, and extraversion will buffer the relation between UWSS and abuse.

Overview of Studies

We conducted two studies to test our hypotheses and research question. Our first study replicates key findings from Gray et al. (2020), establishes relations between UWSS and behavioral strain (CWB), and assesses the incremental validity of UWSS over incivility and interpersonal conflict in the prediction of relevant workplace outcomes. Our second study replicates UWSS-CWB relations from Study 1 and examines the influencing role of individual differences on UWSS-Emotional exhaustion and UWSS-CWB relations. Taken together, these studies, drawing from the SOS model (Semmer et al., 2019), help broaden and build upon the growing research on social support as a stressor, and further contribute to the literature on individual differences in stressor-strain relations.

Study 1 Method

Participants and Procedure

Data for this study were collected using Amazon’s Mechanical Turk (MTurk) during the COVID-19 pandemic. Due to the pandemic’s potential effects on the focal constructs in this study (e.g., experiencing UWSS and other behaviors), data were screened meticulously. To participate in our survey, participants had to first pass a CAPCHA (to prove they were not bots) and a 4-item English language competency test (e.g., “Choose the best synonym (words of similar meaning) for “bold”). Then, they had to meet criteria such as being employed outside of MTurk (and working 20 non-MTurk hours per week), working on-site at least sometimes, and interacting on-site with at least 1 coworker. In total, 1061 participants consented to participate in this study, however, only 307 were eligible to participate. Of those 307, 104 participants were removed for inattentive responding (i.e., missing 2 or more of 4 attention check items; N = 30), taking longer than 30 min to complete the survey (N = 20), or identically responding to 14 or more consecutive UWSS items (half of the scale; N = 54), resulting in a final sample of 203 (M_age = 37, SD_age = 11; 57% male). Participants worked an average of 38 non-MTurk hours per week (SD = 8.5) and the majority (75%) worked on-site over half the time. This sample was representative of a variety of occupations, with the most prevalent being Health Care and Social Assistance (13%), Finance and Insurance (11%), and Manufacturing (11%).

Measures

Unhelpful Workplace Social Support

To measure UWSS, the 28-item UWSS measure from Gray et al. (2020) was used. Responses are quantified on a 1 to 6 scale (1 = Never, 6 = Very frequently). Although we model this construct as unidimensional (i.e., general unhelpful support) to test our hypotheses (similar to Gray et al. (2020) multivariate analyses), we will present the 7 subscales in our correlation matrix in order to observe potential differences in bivariate relations with the other constructs in this study. The UWSS measure reported a coefficient alpha value of .98.

Incivility

To measure workplace incivility, the 4-item Workplace Incivility measure from Matthews and Ritter (2016) was used. Responses to items are quantified on a 1 to 5 scale (1 = Never, 5 = Many times). The coefficient alpha value for this measure was .88.

Interpersonal Conflict at Work

To measure interpersonal conflict at work, the 4-item Interpersonal Conflict at Work Scale from Spector and Jex (1998) was used. Responses are quantified on a 1 to 5 scale (1 = Less than once per month or never, 5 = Several times per day). The coefficient alpha value for this measure was .94.

Counterproductive Work Behavior

To measure abuse, production deviance, and withdrawal, we used the 17, 3, and 4-item subscales (respectively) from the 32-item Counterproductive Work Behavior Checklist (CWB-C; Spector et al., 2006). Responses to items are quantified on a 1 to 5 scale (1 = Never, 5 = Everyday). The coefficient alpha values for the production deviance, withdrawal, and abuse scales were .94, .89, .99.

Job-related Negative Affect

To measure job-related negative affect, 10-items from the Job-Related Affective Well-Being scale (Van Katwyk et al., 2000) were used. Responses are quantified on a 1 to 5 scale (1 = Never, 5 = Extremely often). The coefficient alpha value for this measure was .94.

Emotional Exhaustion

To measure emotional exhaustion, 8 items from the 16-item Oldenburg Burnout Inventory (OLBI; Demerouti & Bakker, 2008) were used. Responses to items are quantified on a 1 to 4 scale (1 = Strongly disagree, 4 = Strongly agree). The coefficient alpha value for this measure was .68.

Satisfaction with Coworkers

To measure satisfaction with coworkers, the 4-item Satisfaction with Coworkers subscale taken from the 36-item Job Satisfaction Survey (Spector, 1985) was used. Responses are quantified on a 1 to 6 scale (1 = Disagree very much, 6 = Agree very much). The coefficient alpha value for this measure was .71.

Study 1 Results

Descriptive statistics and bivariate correlations for each variable can be found in Table 1. Before testing our hypotheses, we tested several measurement models to ensure construct validity of our variables. Our utilized 9-factor measurement model, estimated using diagonally weighted least squares estimation demonstrated excellent fit using fit indices recommended by Kline (2015): χ²(3203) = 2326.93, p > .01; CFI = 1.00; RMSEA = .00; SRMR = .06. This model also demonstrated significantly better fit compared to a 7-factor model that combined UWSS, incivility, and interpersonal conflict (i.e., negative social exchanges that were highly correlated), as well as to a 5-factor model that combined these exchanges and collapsed the CWB dimensions (all highly correlated) into a single factor: χ²(15) = 285.84, p < .01; χ²(26) = 313.78, p < .01, respectively. These results help provide construct validity evidence for our variables.

Table 1.

Means, standard deviations, and correlations with confidence intervals for study 1

Variable	M	SD	1	2		3	4	5		6
1. UWSS (General)	3.02	1.27	.98
2. UWSS (Critical)	2.56	1.58	.90** [.86, .92]	.95
3. UWSS (Imposing)	3.17	1.32	.92** [.90, .94]	.83** [.78, .87]		.90
4. UWSS (Partial)	3.28	1.30	.89** [.85, .91]	.72** [.65, .78]		.80** [.74, .84]	.91
5. UWSS (Undependable)	3.14	1.34	.91** [.88, .93]	.74** [.67, .80]		.79** [.73, .84]	.81** [.75, .85]	.91
6. UWSS (Shortsighted)	2.95	1.43	.90** [.88, .93]	.83** [.78, .87]		.86** [.82, .89]	.72** [.65, .78]	.75** [.68, .80]		.94
7. UWSS (Uncomforting)	2.89	1.49	.91** [.88, .93]	.80** [.75, .85]		.80** [.75, .85]	.75** [.69, .81]	.82** [.77, .86]		.79** [.73, .83]
8. UWSS (Conflicting)	3.08	1.39	.93** [.91, .94]	.75** [.68, .80]		.79** [.73, .83]	.83** [.78, .87]	.86** [.83, .90]		.79** [.73, .84]
9. Incivility	2.40	1.13	.84** [.79, .87]	.80** [.74, .84]		.76** [.70, .81]	.71** [.63, .77]	.77** [.71, .82]		.73** [.66, .79]
10. Interpersonal Conflict	2.12	1.23	.82** [.76, .86]	.86** [.81, .89]		.74** [.68, .80]	.66** [.58, .74]	.71** [.64, .77]		.76** [.69, .81]
11. Abuse	1.94	1.16	.84** [.79, .88]	.85** [.80, .88]		.75** [.68, .80]	.67** [.59, .74]	.74** [.66, .79]		.77** [.70, .82]
12. Production Deviance	1.95	1.16	.78** [.72, .83]	.78** [.72, .83]		.69** [.61, .75]	.63** [.54, .71]	.68** [.60, .75]		.71** [.64, .78]
13. Withdrawal	2.28	1.08	.76** [.70, .81]	.72** [.65, .78]		.67** [.59, .74]	.63** [.54, .71]	.72** [.64, .78]		.68** [.60, .75]
14. JAWS	2.41	1.00	.73** [.66, .79]	.66** [.57, .73]		.65** [.56, .72]	.66** [.57, .73]	.72** [.65, .78]		.59** [.50, .68]
15. Emotional Exhaustion	2.34	0.46	.41** [.29, .52]	.30** [.17, .42]		.36** [.23, .48]	.43** [.31, .54]	.43** [.31, .54]		.31** [.18, .43]
16. Coworker Satisfaction	4.16	0.96	−.55** [−.64, −.45]	−.47** [−.58, −.36]		−.50** [−.60, −.39]	−.52** [−.61, −.41]	−.57** [−.66, −.47]		−.40** [−.51, −.27]
Variable	7	8	9	10	11	12	13	14	15	16
7. UWSS (Uncomforting)	.92
8. UWSS (Conflicting)	.84** [.80, .88]	.94
9. Incivility	.78** [.72, .83]	.76** [.69, .81]	.88
10. Interpersonal Conflict	.74** [.67, .79]	.70** [.62, .76]	.82** [.76, .86]	.94
11. Abuse	.80** [.75, .85]	.75** [.68, .80]	.84** [.79, .87]	.88** [.85, .91]	.99
12. Production Deviance	.73** [.66, .79]	.70** [.62, .76]	.80** [.74, .84]	.82** [.77, .86]	.93** [.91, .95]	.94
13. Withdrawal	.70** [.62, .76]	.70** [.62, .76]	.78** [.72, .83]	.77** [.71, .82]	.87** [.83, .90]	.87** [.83, .90]	.89
14. JAWS	.70** [.62, .77]	.66** [.57, .73]	.80** [.74, .84]	.71** [.64, .78]	.73** [.66, .79]	.71** [.63, .77]	.75** [.68, .80]	.94
15. Emotional Exhaustion	.38** [.25, .49]	.40** [.28, .51]	.41** [.28, .52]	.31** [.18, .43]	.22** [.09, .36]	.22** [.08, .35]	.28** [.14, .40]	.53** [.42, .62]	.68
16. Coworker Satisfaction	−.53** [−.62, −.42]	−.53** [−.62, −.42]	−.53** [−.62, −.42]	−.50** [−.60, −.39]	−.45** [−.56, −.33]	−.43** [−.54, −.31]	−.47** [−.57, −.35]	−.53** [−.63, −.42]	−.58** [−.67, −.48]	.71

N = 203. Values in square brackets indicate the 95% confidence interval for each correlation. The confidence interval is a plausible range of population correlations that could have caused the sample correlation (Cumming, 2014). Alpha values are in bold along the diagonal. UWSS:Unhelpful workplace social support; JAWS Job-related negative affect

* p < .05, ** p < .01

Hereon, we report the 95% confidence intervals for our correlations, denoted as “95% CI”, with the corresponding lower- and upper-bound confidence intervals following suit.

Hypothesis 1 Supported

UWSS was related to increased a) job-related negative affect (r = .73, 95% CI[.66, .79], p < .01), b) emotional exhaustion (r = .41, 95% CI[.29, .52], p < .01), and c) interpersonal conflict at work (r = .82, 95% CI[.76, .86], p < .01). UWSS was also related to decreased d) satisfaction with coworkers (r = −.55, 95% CI[−.64, −.45], p < .01).

Hypothesis 2 Supported

UWSS was related to a) abuse (r = .84, 95% CI[.79, .88], p < .01), b) withdrawal (r = .76, 95% CI[.70, .81], p < .01), and c) production deviance (r = .78, 95% CI[.72, .83], p < .01).

Research Question

Delineating UWSS’ place among related workplace exchanges is important, particularly with a construct as novel as UWSS. We took a research question approach, using hierarchical regression, assessing the predictive validity of UWSS compared to incivility and interpersonal conflict in predicting well-being (i.e., emotional exhaustion) and attitudinal (i.e., satisfaction with coworkers, job-related negative affect) outcomes. UWSS demonstrated incremental validity over incivility and interpersonal conflict in the prediction of satisfaction with coworkers (ΔR² = .03, F(3,194) = 30.3, β = −.33, t(194) = −2.77, p < .01) and emotional exhaustion (ΔR² = .03, F(3,192) = 15.4, β = .32, t(192) = 2.47, p < .05), but not in the prediction of job-related negative affect.

Study 1 Discussion

The purpose of this study was to replicate findings from Gray et al. (2020) and extend these findings by focusing on behavioral strain in the form of CWB. In addition, we sought to determine the incremental validity of UWSS over incivility and interpersonal conflict—similar but distinct negative social exchanges. Results fully supported our hypotheses. To begin, UWSS was related to engaging in production deviance, withdrawal, and abusive behavior at work, which is in line with the SOS model’s proposition that behavioral strain is a likely outcome of threats to self-esteem. Next, we replicated several strong (i.e., r > .50) significant relations from Gray et al. (2020), focused on interpersonal conflict, emotional exhaustion, (dis)satisfaction with coworkers, and job-related negative affect, with each replicated bivariate relation providing additional validity support for Gray et al.’s (2020) novel measure of UWSS—and the latter three providing support for UWSS as a threat to the self per the SOS model (Semmer et al., 2019). Lastly, our incremental validity analyses revealed that UWSS has explanatory power above and beyond incivility and interpersonal conflict in the prediction of emotional exhaustion and satisfaction with coworkers, established outcomes of both interpersonal stressors. Moreover, incivility, interpersonal conflict, and UWSS were all highly correlated (see Table 1); this was expected, as each of these experiences would be categorized as an interpersonal workplace stressor. Our findings clearly indicate that UWSS, even when compared to established negative workplace experiences, is a potent stressor for employees—one that warrants further investigation. Moving forward, we focus on the moderating effects of individual differences on the impacts of UWSS.

Study 2 Method

Participants and Procedure

A total of 1035 employees consented to complete a survey hosted on CloudResearch containing measures of UWSS, emotional exhaustion, incivility, CWB, and our individual differences variables. Participants were required to meet all the same eligibility criteria as Study 1. After eligibility screening and data cleaning (data were cleaned akin to Study 1; eligibility, N = 629; attention checks, N = 0; consecutive responses, N = 109; over 30-min survey completion time, N = 20), which removed 758 participants, the final sample consisted of 277 employees (M_age = 40, SD_age = 12; 52% female). Participants worked an average of 39 non-MTurk hours (SD = 7) and the majority (64%) worked on-site over half the time. This sample was representative of a variety of occupations, with the most prevalent being Education (16%), Health Care and Social Assistance (14%), and Retail Trade (9%).

Measures

Unhelpful workplace social support (α = .95), emotional exhaustion (α = .84), abuse (α = .94), production deviance (α = .58),2 and withdrawal (α = .76) were assessed using the same measures and response formats found in Study 1.

Personality

To assess the focal FFM personality domains, 32 items from the 40-item Big Five Mini-Markers measure (Saucier, 1994) were used. Responses are quantified on a 1 to 9 scale (1 = Extremely inaccurate, 9 = Extremely accurate). The coefficient alphas for each of the Big Five domains were as follows: extraversion (.86), agreeableness (.91), conscientiousness (.88), and neuroticism (.85).

Incivility

Given the strong correlation between UWSS and incivility in Study 1 (r = .85[.81,.88], p ≤ .01), we opted to assess incivility using a more comprehensive measure: the 12-item workplace incivility measure from Cortina et al. (2013). Responses are quantified on a 1 to 5 scale (1 = Never, 5 = Many times). The coefficient alpha for this measure was .95.

Negative Reciprocity

To measure negative reciprocity beliefs, the 14-item negative reciprocity measure from Eisenberger et al. (2004) was used. Responses to items are quantified on a 1 to 7 scale (1 = Strongly disagree, 7 = Strongly agree). The coefficient alpha for this measure was .95.

Study 2 Results

Descriptive statistics and bivariate correlations for each variable can be found in Table 2. As in Study 1, we fit our intended measurement model (estimated using diagonally weighted least squares estimation) and compared this model to alternative models. Our utilized 11-factor measurement model demonstrated acceptable fit: χ²(6730) = 6993.33, p < .05; CFI = .99; RMSEA = .01; SRMR = .07. This model also demonstrated significantly better fit compared to a 10-factor measurement model that combined UWSS and incivility (i.e., negative social exchanges that were highly correlated): χ²(10) = 569.78, p < .01. These results help provide construct validity evidence for our variables.

Table 2.

Means, standard deviations, and correlations with confidence intervals for study 2

Variable	M	SD	1	2	3	4	5	6
1. UWSS (General)	2.24	0.80	.95
2. UWSS (Critical)	1.57	0.89	.73** [.67, .78]	.87
3. UWSS (Imposing)	2.42	1.01	.73** [.67, .78]	.48** [.39, .57]	.83
4. UWSS (Partial)	2.65	1.09	.82** [.77, .85]	.50** [.40, .58]	.51** [.42, .59]	.88
5. UWSS (Undependable)	2.39	1.08	.77** [.72, .82]	.50** [.40, .58]	.41** [.31, .50]	.56** [.47, .64]	.90
6. UWSS (Shortsighted)	2.13	1.04	.70** [.64, .76]	.46** [.36, .55]	.62** [.54, .69]	.45** [.35, .54]	.37** [.27, .47]	.93
7. UWSS (Uncomforting)	1.98	1.07	.73** [.67, .78]	.55** [.47, .63]	.44** [.34, .53]	.51** [.42, .60]	.56** [.47, .64]	.42** [.32, .52]
8. UWSS (Conflicting)	2.44	1.14	.84** [.80, .87]	.49** [.39, .57]	.46** [.36, .55]	.74** [.68, .79]	.67** [.59, .73]	.44** [.34, .53]
9. Incivility	1.51	0.69	.66** [.59, .72]	.67** [.60, .73]	.42** [.32, .51]	.49** [.40, .58]	.52** [.43, .60]	.40** [.29, .49]
10. Extraversion	5.34	1.52	−.20** [−.31, −.08]	−.11 [−.22, .01]	−.20** [−.31, −.09]	−.10 [−.22, .02]	−.17** [−.28, −.05]	−.14* [−.25, −.02]
11. Agreeableness	7.03	1.39	−.23** [−.34, −.12]	−.24** [−.35, −.13]	−.09 [−.21, .03]	−.18** [−.29, −.06]	−.25** [−.36, −.14]	−.08 [−.20, .04]
12. Conscientiousness	7.21	1.27	−.23** [−.34, −.11]	−.28** [−.39, −.17]	−.18** [−.29, −.07]	−.16** [−.27, −.04]	−.05 [−.17, .07]	−.20** [−.31, −.08]
13. Neuroticism	3.58	1.44	.32** [.21, .42]	.31** [.20, .41]	.24** [.12, .34]	.20** [.09, .31]	.21** [.09, .32]	.19** [.07, .30]
14. Negative Reciprocity	2.73	1.12	.18** [.06, .29]	.20** [.09, .31]	.12 [−.00, .23]	.07 [−.04, .19]	.18** [.06, .29]	.08 [−.04, .20]
15. Abuse	1.18	0.33	.38** [.28, .48]	.46** [.36, .55]	.26** [.15, .37]	.23** [.12, .34]	.33** [.23, .43]	.28** [.17, .39]
16. Production Deviance	1.29	0.43	.35** [.24, .45]	.34** [.23, .44]	.21** [.09, .32]	.27** [.16, .38]	.27** [.16, .38]	.22** [.10, .33]
17. Withdrawal	1.74	0.62	.28** [.17, .39]	.19** [.07, .30]	.13* [.01, .24]	.25** [.14, .36]	.28** [.16, .38]	.15* [.03, .27]
18. Emotional Exhaustion	2.32	0.52	.36** [.26, .46]	.28** [.16, .38]	.21** [.10, .32]	.29** [.18, .39]	.26** [.15, .37]	.19** [.08, .30]
Variable		7	8	9	10	11	12	13
7. UWSS (Uncomforting)		.90
8. UWSS (Conflicting)		.52** [.43, .60]	.94
9. Incivility		.50** [.41, .58]	.54** [.45, .62]	.95
10. Extraversion		−.17** [−.29, −.06]	−.15** [−.27, −.04]	−.13* [−.24, −.01]	.86
11. Agreeableness		−.13* [−.24, −.01]	−.24** [−.35, −.13]	−.31** [−.41, −.20]	.36** [.25, .46]	.91
12. Conscientiousness		−.18** [−.29, −.07]	−.18** [−.29, −.06]	−.21** [−.32, −.09]	.16** [.05, .27]	.46** [.36, .55]	.88
13. Neuroticism		.31** [.19, .41]	.27** [.16, .38]	.30** [.19, .41]	−.33** [−.43, −.22]	−.46** [−.55, −.36]	−.53** [−.61, −.43]	.85
14. Negative Reciprocity		.07 [−.05, .19]	.21** [.10, .32]	.22** [.11, .33]	−.14* [−.25, −.02]	−.44** [−.53, −.34]	−.31** [−.41, −.20]	.28** [.17, .39]
15. Abuse		.22** [.10, .33]	.27** [.16, .38]	.46** [.36, .55]	−.06 [−.18, .06]	−.40** [−.50, −.30]	−.29** [−.40, −.18]	.27** [.16, .38]
16. Production Deviance		.20** [.08, .31]	.32** [.21, .42]	.31** [.20, .41]	−.16** [−.27, −.04]	−.40** [−.50, −.30]	−.37** [−.46, −.26]	.30** [.19, .40]
17. Withdrawal		.21** [.09, .32]	.26** [.15, .37]	.19** [.07, .30]	−.05 [−.17, .07]	−.18** [−.29, −.07]	−.30** [−.40, −.19]	.14* [.02, .25]
18. Emotional Exhaustion		.31** [.20, .41]	.37** [.27, .47]	.42** [.32, .52]	−.33** [−.43, −.22]	−.26** [−.37, −.15]	−.31** [−.41, −.20]	.44** [.33, .53]
Variable		14		15	16	17		18
14. Negative Reciprocity		.95
15. Abuse		.26** [.15, .37]		.94
16. Production Deviance		.20** [.08, .31]		.56** [.47, .64]	.58
17. Withdrawal		.09 [−.02, .21]		.38** [.28, .48]	.43** [.33, .52]	.76
18. Emotional Exhaustion		.22** [.11, .33]		.15* [.03, .26]	.26** [.14, .36]	.18** [.07, .29]		.84

N = 277. Values in square brackets indicate the 95% confidence interval for each correlation. The confidence interval is a plausible range of population correlations that could have caused the sample correlation (Cumming, 2014). Alpha values are in bold along the diagonal. UWSS Unhelpful workplace social support

* p < .05, ** p < .01

Hypotheses were tested using Hierarchical Moderated Multiple Regression (HMMR). Hierarchical regression is an analytic technique, guided by a set of theoretical assumptions, that allows for a predictor(s) to account for variance beyond another predictor(s); this approach is often employed to investigate multivariate effects (Aiken et al., 1991). In step 1, continuous predictors (e.g., UWSS) were entered into the regression equation, followed by interaction terms (e.g., UWSS x Conscientiousness) in step 2. For increased interpretability, predictors and interaction terms were mean centered (Dalal & Zickar, 2012). Each interaction was probed using simple slopes analyses to determine which levels of each respective moderator had significant slopes. See Table 3 for complete HMMR results. Interactions were deemed meaningful in magnitude if the range in slopes (difference between +1 SD and − 1 SD) was greater than .1 as all interactions below this range had slopes whose confidence intervals overlapped entirely across all ranges of UWSS.

Table 3.

Hierarchical moderated multiple regression tests for Hypotheses 3 through 6

Hypothesis 3 (EE)

Hypothesis 4 (PD)

Hypothesis 5 (WD)

Predictor

b

β

Fit

Difference

b

β

Fit

Difference

b

β

Fit

Difference

Int.

2.3**

1.0**

1.3**

UWSS

.13**

.23

.13**

.24

.19**

.24

Neuroticism

.14**

.26

.00

.01

−.06

−.10

Extraversion

−.11**

−.20

.01

.02

.02

.03

Agreeableness

.01

.03

−.11**

−.26

−.03

−.05

Conscientiousness

−.05

−.10

−.08**

−.19

−.16**

−.28

R2 = .28**

R2 = .26**

R2 = .15**

Int.

2.2**

1.1**

1.3**

UWSS

.14**

.23

.08*

.15

.20**

.26

Neuroticism

.09*

.18

.01

.03

−.03

−.04

Extraversion

−.10**

−.18

−.01

−.03

.02

.03

Agreeableness

.00

.00

−.08**

−.20

.00

.00

Conscientiousness

−.10**

−.21

−.05

−.11

−.17**

−.29

UWSSxN

.11**

.21

.01

.03

−.10

−.17

UWSSxE

−.02

−.03

.06*

.12

−.01

−.02

UWSSxA

.03

.06

−.07*

−.20

−.07

−.14

UWSSxC

.11**

.25

−.04

−.12

.01

.02

R2 = .32**

ΔR2 = .04**

R2 = .32**

ΔR2 = .06**

R2 = .16**

ΔR2 = .01

Hypothesis 6 (Abuse)

Predictor

b

β

Fit

Difference

Int.

1.1**

UWSS

.11**

.30

Neuroticism

.01

.02

Extraversion

.05*

.14

Agreeableness

−.10**

−.31

Conscientiousness

−.02

−.08

NR

.02

.06

R2 = .28**

Int.

1.1**

UWSS

.08**

.20

Neuroticism

.02

.07

Extraversion

.05**

.15

Agreeableness

−.10**

−.32

Conscientiousness

.03

.08

NR

−.01

−.03

UWSSxN

−.03

−.09

UWSSxE

−.02

−.05

UWSSxA

−.00

−.00

UWSSxC

−.08**

−.29

UWSSxNR

.07**

.21

R2 = .38**

ΔR2 = .10**

N = 277. Predictors centered at mean. A significant b-weight indicates the beta-weight is also significant. b represents unstandardized regression weights. β indicates the standardized regression weights. Int. represents the intercept term at each step. Confidence intervals (CI) are at the 95% level. * indicates p < .05. ** indicates p < .01. UWSS Unhelpful workplace social support; N Neuroticism; E Extraversion; A Agreeableness; C Conscientiousness; NR: Negative reciprocity; EE Emotional Exhaustion; PD Production deviance; WD Withdrawal

The observed relations in Study 1 between UWSS and abuse (r = .38, 95% CI[.28, .48], p < .01), production deviance (r = .35, 95% CI[.24, .45], p < .01), and withdrawal (r = .28, 95% CI[.17, .39], p < .01) were replicated here. These results provide additional empirical evidence for the ties between UWSS and CWB. In addition, UWSS and incivility were once again highly correlated (r = .66, 95% CI[.59, .72], p < .01).

Hypothesis 3 Partially Supported

Agreeableness (β = .06, t(267) = .75, p = .46) and extraversion (β = −.03, t(267) = −.56, p = .57) did not significantly interact with UWSS in the prediction of emotional exhaustion. Conscientiousness (β = .25, t(267) = 2.9, p < .01; see Fig. 1) and neuroticism (β = .21, t(267) = 2.7, p < .01; see Fig. 2), however, did significantly interact with UWSS in the prediction of emotional exhaustion (ΔR² = .04, F(9, 267) = 14.08, p < .01). The relation between UWSS and emotional exhaustion was strengthened by higher levels of conscientiousness and neuroticism.

Fig. 1.

Slopes of effect of UWSS on emotional exhaustion for low, average, and high conscientiousness. Note. Low/High Conscientiousness groups −1/+1 standard deviation. UWSS: Unhelpful Workplace Social Support

Fig. 2.

Slopes of effect of UWSS on abuse for low, average, and high neuroticism. Note. Low/High Neuroticism groups −1/+1 standard deviation. UWSS: Unhelpful Workplace Social Support

Hypothesis 4 Partially Supported

Neuroticism (β = .03, t(267) = .34, p = .73) and conscientiousness (β = −.12, t(267) = −1.4, p = .17) did not significantly interact with UWSS in the prediction of production deviance. Agreeableness (β = −.20, t(267) = −2.6, p < .05; see Fig. 3) and extraversion (β = .12, t(267) = 2.0, p < .05; see Fig. 4), however, did significantly interact with UWSS to predict production deviance (ΔR² = .06, F(9, 267) = 13.95, p < .01). The relation between UWSS and production deviance was weakened by higher levels of agreeableness and strengthened by higher levels of extraversion.

Fig. 3.

Slopes of Effect of UWSS on Production Deviance for low, average, and high Agreeableness. Note. Low/High Agreeableness groups −1/+1 standard deviation. UWSS: Unhelpful Workplace Social Support

Fig. 4.

Slopes of effect of UWSS on abuse for low, average, and high extraversion. Note. Low/High Extraversion groups −1/+1 standard deviation. UWSS: Unhelpful Workplace Social Support

Hypothesis 5 Not Supported

None of the Big Five traits interacted with UWSS to predict withdrawal: neuroticism (β = −.17, t(267) = −1.9, p = .06), conscientiousness (β = .02, t(267) = .17, p = .86), agreeableness (β = −.14, t(267) = −1.6, p = .11), and extraversion (β = −.02, t(267) = −.28, p = .78).

Hypothesis 6 Partially Supported

Extraversion (β = −.05, t(265) = −.87, p = .38), neuroticism (β = −.09, t(265) = −1.2, p = .24), and agreeableness (β = −.00, t(265) = −.04, p = .97) did not significantly interact with UWSS in the prediction of abuse. Negative reciprocity beliefs (β = .21, t(265) = 3.3, p < .01; see Fig. 5) and conscientiousness (β = −.29, t(265) = −3.5, p < .01; see Fig. 6), however, did significantly interact with UWSS in the prediction of abuse (ΔR² = .11, F(11, 265) = 14.94, p < .01). The relation between UWSS and abuse was strengthened by higher levels of negative reciprocity beliefs and lower levels of conscientiousness.

Fig. 5.

Slopes of effect of UWSS on Abuse for low, average, and high Negative Reciprocity. Note. Low/High Negative Reciprocity groups −1/+1 standard deviation. NR: Negative Reciprocity; UWSS: Unhelpful Workplace Social Support

Fig. 6.

Slopes of effect of UWSS on Abuse for low, average, and high Conscientiousness. Note. Low/High Conscientiousness groups −1/+1 standard deviation. UWSS: Unhelpful Workplace Social Support

Exploratory Analysis

Although the purpose of this paper is to examine UWSS holistically, it is likely that the unique forms of UWSS differentially contribute to strain. To test this assumption, we conducted a series of dominance analyses. Dominance analysis allows researchers a glimpse into the importance of each predictor in terms of R² contribution (Azen & Budescu, 2003), and is useful for addressing questions surrounding the relative explanatory importance of each form of UWSS. We focused on our focal outcomes (emotional exhaustion, production deviance, abuse, withdrawal), and found differences among forms of UWSS in predictor importance: critical support (social support that is delivered in a rude or hostile manner) was the most important predictor of both abuse and production deviance, conflicting social support (social support that conflicts with advice given previously by others) was the most importance predictor of emotional exhaustion, and undependable social support (social support that is low quality, unreliable, or delayed) was the most important predictor of withdrawal. See Table 4 for a complete summary of our dominance analyses.

Table 4.

Summary of exploratory dominance analyses

Predictor	Average R2 Contribution for EE	Average R2 Contribution for Abuse	Average R2 Contribution for PD	Average R2 Contribution for Withdrawal
Critical Social Support	.02	.13	.05	.01
Imposing Social Support	.01	.02	.01	.00
Partial Social Support	.02	.01	.02	.02
Undependable Social Support	.02	.04	.02	.03
Shortsighted Social Support	.01	.02	.01	.00
Uncomforting Social Support	.03	.01	.01	.01
Conflicting Social Support	.06	.02	.03	.02

N = 277. The greatest average R² contribution for each outcome is bolded. EE Emotional exhaustion; PD Production deviance

Study 2 Discussion

The purpose of this study was to investigate the moderating effects of individual differences on the UWSS-emotional exhaustion and UWSS-CWB relations, as well as replicate the behavioral strain findings from Study 1. Main effects between UWSS and production deviance, withdrawal, and abuse were replicated, providing further support for the notion of behavioral strain as an outcome of threats to the self, and UWSS as such a threat. Our multivariate results suggest that specific individual differences play a role in determining the extent to which UWSS is appraised as a threat to self, though. More specifically, we found that conscientiousness, agreeableness, extraversion, neuroticism, and negative reciprocity beliefs play a role in influencing the strength of relations held between UWSS and emotional exhaustion and CWB—costly outcomes for organizations. Moreover, once again found incivility and UWSS to be highly related; we discuss theoretical explanations for these relations in a future section. Finally, we found that different forms of UWSS were of differential importance in predicting focal outcomes in our exploratory analyses.

General Discussion

The purpose of the present research was to extend the literature on workplace social support as a stressor (i.e., UWSS) by establishing novel—and replicating prior—main effect relations, as well as addressing the individual differences gap in said literature. Drawing from the SOS model (Semmer et al., 2019), we, through two studies, present findings that contribute to the literature on stressful social support by: a) establishing relations between UWSS and behavioral strain (i.e., CWB) and replicating prior main effects, b) determining how UWSS is situated among related exchanges and interactions (e.g., incivility and interpersonal conflict), and c) assessing how individual differences impact the strength of noteworthy UWSS relations. Furthermore, these findings provide additional validity support for the novel UWSS measure constructed to assess this space (Gray et al., 2020), as well as build upon the extant research drawing from the SOS model (Semmer et al., 2019).

The guiding principle of the SOS model (Semmer et al., 2019) is that personal or social self-esteem is threatened or boosted by external stimuli, and that threats to either kind of self-esteem work to engender stress and reduce well-being. In support of this premise, three of the four higher magnitude relations we replicated dealt with psychological strain (e.g., job-related negative affect), a common outcome of threats to personal and social self-esteem. We also observed relations between UWSS and CWB, a noteworthy form of behavioral strain; behavioral strains are also typical as a response to threats to the self (Semmer et al., 2019). Such results were anticipated, despite UWSS exchanges bestowing (or attempting to bestow) resources to recipients—resources absent in other negative social exchanges (e.g., incivility). This is because these resources are delivered in ways that are likely to threaten and consequently reduce social (e.g., “Do they think that I’m unintelligent?”; “Do they think that I’m not worth teaching?”) and personal self-esteem (e.g., “Maybe I am unintelligent.”; “Maybe I’m not actually worth teaching.”). It should be noted that UWSS appears to be on par with other social stressors in terms of impact, as we found that UWSS had incremental validity beyond incivility and interpersonal conflict in the prediction of emotional exhaustion and coworker (dis)satisfaction. This finding opens the door for interesting future research, as we will describe in a future section.

The results from our second study suggest, as noted by Semmer et al. (2019), that people are likely to vary in the extent to which UWSS may be appraised as a threat to the self, thus influencing such exchanges’ relations with strain. The impact of other negative interpersonal exchanges are moderated by individual differences (e.g., Gallagher & Hughes, 2020; Peng et al., 2020; Wu et al., 2014), and UWSS is no different. We found that negative reciprocity beliefs, extraversion, and agreeableness influenced particular UWSS-CWB relations, while neuroticism influenced the UWSS-emotional exhaustion relation. Moreover, conscientiousness uniformly influenced UWSS-emotional exhaustion and UWSS-CWB relations. Beginning with agreeableness, those who are more agreeable have been found to give people the “benefit of the doubt” often (Lazarus & Folkman, 1984), and we anticipated agreeableness to uniformly weaken UWSS-CWB relations because of UWSS not being perceived as a threat to the self in these individuals. This buffering effect, however, only held for production deviance. Production deviance behaviors represent purposeful attempts by an employee to incorrectly perform tasks and is a noted form of organizationally directed deviance. Perhaps agreeable employees, because of their kindness and compassion, feel that their employers need not suffer as a result of their employees’ unhelpful or ineffective support. Meanwhile, negative reciprocity beliefs strengthened the UWSS-abuse relation. Those who possess negative reciprocity beliefs, in addition to believing in “eye for an eye” sentiment, are also likely to attribute hostility to negative events (e.g., Hoobler & Brass, 2006), meaning they are likely to view UWSS, despite the attempt of or actual provision of resources, as a significant threat to the self—one worth interpersonally retaliating over. Finally, conscientiousness played a role in UWSS’ relations with both psychological (emotional exhaustion) and behavioral strain (abuse). We contended that conscientious employees would be, as suggested by prior research (e.g., Sliter et al., 2015), more likely to detect the social norm violations inherent to UWSS, making UWSS a more salient threat to the self. Indeed, conscientiousness strengthening the UWSS-emotional exhaustion relation supports this line of thinking. Importantly, though, conscientiousness employees are strict rule followers, and may want others to view them as such (e.g., Chang et al., 2012; Shaffer et al., 2015), helping to explain the buffering effect conscientiousness had on the UWSS-abuse relation in particular, as these behaviors are typically more visible (e.g., Bowling et al., 2020). Finally, we also found moderating effects for both neuroticism and extraversion for the relations between UWSS and emotional exhaustion and UWSS and production deviance, respectively. The former effect was expected, as neurotic employees are often sensitive to stressors (McCrae & Costa, 1987), and are likely to find UWSS particularly draining. Extraverted employees, though, were more likely to engage in production deviance—an unexpected finding. It may be that extraverted employees, as a result of their social intelligence (Chamorro-Premuzic et al., 2010), do perceive UWSS as a threat to the self, but only to the extent that they need to engage in production deviance behaviors (e.g., intentionally working slowly) to cope with them. We, however, advise that these specific effects be interpreted with caution, as the change in the slopes for UWSS at each level of neuroticism/extraversion were very small compared to our other interactions (less than a .06 change in slope between +1 and − 1 standard deviation of neuroticism/extraversion; our other interactions saw changes in slope greater than .1 for +1 and − 1 standard deviations of our moderators). Altogether, these findings begin to fill the individual difference-related void in the literature on workplace social support as a stressor, and further contribute to the needed research on the role of individual differences in stressor-strain relations (Bowling & Jex, 2013). We strongly recommend scholars continue to explore the impact of individual differences on these relations.

In both studies, we found incivility and UWSS to be strongly related (r > .65). Although our measurement models suggest superior fit when modeled as distinct, there are certainly conceptual similarities that should be highlighted between both constructs. Perhaps most of the conceptual overlap between UWSS and incivility concerns the valence and modality of both behaviors. Regarding the former, both UWSS and incivility are behaviors of a negative valence, as demonstrated by their nomological networks. The latter may be of particular importance, though. Take, for instance, critical social support; this type of UWSS concerns support that is provided in a way that is rude or hostile (e.g., “Insult me when providing advice”). Compare this to a common example of incivility: “Made insulting or disrespectful remarks about you” (e.g., Cortina et al., 2013). The modalities of these behaviors are rather similar, with both being rude or hostile, although UWSS exchanges bestow resources to recipients, while uncivil exchanges, by definition, do not. It is possible that employees may lump such exchanges together under a broad “interpersonal stressor” umbrella, rendering them highly correlated in both samples. The interconnectedness of these constructs demands future research.

Finally, our exploratory analyses revealed that the different forms of UWSS were of differential importance in predicting our focal outcomes. Although we focus primarily on UWSS holistically, as it is a relatively novel social exchange variable, we think it is important that researchers be aware of the differential explanatory power of each form. As we note in the following section, it may be that different forms of UWSS work through different mechanisms (e.g., cognitive depletion, negative affect). For instance, perhaps critical social support, due to the hostility of these exchanges, works primarily through affective mechanisms, similar to other forms of mistreatment (Su et al., 2021). Meanwhile, undependable social support, or that which is unreliable, low quality, or delayed, may work through more cognitive mechanisms, as recipients must expend cognitive resources to correct or work through the provider’s mistakes. Further delineating differences between the forms of UWSS would be beneficial for both theory and practice, as such findings would provide both a better understanding of social support as a stressor and may inform social support interventions in organizations.

Limitations and Future Research Directions

The present research is not without limitations. Perhaps the largest limitation is that both studies reported findings that were cross-sectional and monomethod in nature. This removes the ability to make inferences regarding the causality or temporal ordering of results observed and increases concerns over common method bias, which can inflate relations between constructs (Podsakoff et al., 2003); although our utilized measurement model demonstrated acceptable fit in Study 1, providing confidence in the construct validity of our measures, the correlations were high, possibly as a result of our design and common method bias. It should be noted, though, that significant multivariate effects, of which we report several in Study 2, tend to not spuriously appear as a result of a common method (Siemsen et al., 2010). Indeed, these type of effects are more prone to Type II error (e.g., Jaccard & Wan, 1995). Thus, while there are limitations in what can be drawn from some of our results, our interaction results can be interpreted with greater confidence.

From a theoretical standpoint, we want to note that we did not assess the relations between UWSS and self-esteem. Thus, while the nomological network of UWSS suggests that it is a threat to the self, future research should explicitly assess self-esteem alongside UWSS. Gray et al. (2020) uncovered a strong positive relation between the two constructs, but future research should consider assessing self-esteem loss as a mediator between UWSS and outcomes. Such research would add validity evidence in support of UWSS’ place within the SOS model (Semmer et al., 2019).

Another potential limitation with this research is both studies’ reliance on participants from MTurk—a crowdsourcing platform whose data quality has been questioned as a result of inhuman participation (i.e., bots), inattentive responding, and more (e.g., Chmielewski & Kucker, 2020). We kept these limitations in mind and used several screening processes to help ensure data quality. For instance, in both studies, CAPCHAs were used to prevent inhuman participants from gaining access to the survey, and participants were required to pass an English competency test to ensure a reading level high enough for participation. Furthermore, to prevent range restriction of our exchange-based variables as a result of the COVID-19 pandemic, we screened participants to ensure that they were having at least some on-site interaction with (at least) a single coworker. Finally, we also screened collected data using attention check items and indicators of careless responding (i.e., response time, consecutive responses) in an attempt to retain only responses of higher quality. In light of these concerns, it should be noted that MTurk does afford researchers some advantages, namely the ability to acquire samples diverse in age, occupation, and education (Keith et al., 2017). Altogether, while MTurk data quality can be questionable, we took several steps to verify data were appropriate for use in these studies.

Keeping these limitations in mind, future research should assess UWSS dynamics using more robust designs. For example, experience sampling methodology would allow researchers to determine which types of UWSS may be more distressing in the moment; here, we found, perhaps because of the designs used, that the pattern of relations across the different types of UWSS did not differ to a notable degree. Future research should empirically investigate the similarities between incivility and UWSS. We suggest that the similarities observed here may be due to similarity in valence and modality. A potentially interesting avenue for future research would be spiraling effects of UWSS. The spiraling effects of incivility have long been discussed (e.g., Andersson & Pearson, 1999), and repeated instances of UWSS may lead recipients to—in addition to engaging in CWB, which we observed—begin providing UWSS to perpetrators in an attempt to “get even” with them. As astutely noted by a reviewer, it is possible that the different kinds of UWSS relate to outcomes through different mechanisms. For example, undependable social support may be especially taxing for cognitive resources, as recipients must work to correct the mistakes of the provider while further trying to understand the task in question. Meanwhile, critical social support may primarily generate negative affect, as the provider in such exchanges is rude, discourteous, or hostile. In our tests of incremental assessment, we found that UWSS predicted above and beyond incivility and interpersonal conflict in the prediction of emotional exhaustion and coworker (dis)satisfaction, but not job-related negative affect. Future research should work to determine in what contexts UWSS is more distressing than other social stressors. Finally, more individual difference-based research surrounding unhelpful social support is needed. Individual differences are important for all other social exchanges, and UWSS is no different.

Conclusion

In closing, the present research, drawing from the SOS model (Semmer et al., 2019), builds upon the extant research surrounding social support as a stressor. More specifically, we establish (and replicate) relations between UWSS and behavioral strain, determine that UWSS has explanatory power over established interpersonal stressors in the prediction of detrimental outcomes, and, perhaps most importantly (and intuitively), find that relations between UWSS and strains are influenced by individual differences. We echo Gray et al. (2020) by suggesting that additional research on this type of support exchange be conducted, as UWSS appears to be a critical stressor for employees whose impacts need to be further understood. More broadly, we encourage future research continue to explore the role of individual differences in stressor-strain relations, as such exploration informs both theory and practice (see Bowling & Jex, 2013).

Data Availability (data transparency)

Data are available upon request; please contact the corresponding author.

Declarations

Ethics Declarations

Participants consented to participate in this research. This research was approved by the Institutional Review Board at Bowling Green State University. This research posed less than minimal risk to participants.

Conflict of Interest Statement

The authors declare that they have no conflict of interest.