Michelle K. Duffy, Daniel C. Ganster, Milan Pagon
9 min
While workplace research has long focused on the positive role of social support, this study introduces and empirically tests the concept of social undermining. Unlike broad categories of workplace deviance or aggression, social undermining specifically refers to subtle, persistent, and intentional behaviors intended to hinder a target's professional reputation, task success, or interpersonal relationships over time. The authors argue that because negative interactions are generally less frequent than positive ones, they carry greater emotional weight and salience.
To evaluate this, the authors surveyed 685 police officers in the Republic of Slovenia. They developed a measure of social undermining that differentiates between behaviors originating from supervisors and coworkers. The study analyzed the associations between these undermining behaviors and several outcomes: work-related self-efficacy, organizational commitment, counterproductive work behaviors (both active and passive), and somatic health complaints. Hierarchical regression was used to compare the impact of social undermining against that of social support.
The results show that social undermining is a consistent and strong predictor of negative employee outcomes. Notably, these effects are often stronger than the effects of social support. The researchers also identified a within-domain exacerbation effect: when a supervisor is perceived as providing high levels of both support and undermining, the employee suffers worse outcomes than if the supervisor were simply unsupportive. This suggests that the uncertainty caused by receiving conflicting signals—being supported and undermined by the same person—is particularly damaging. The authors conclude that organizations must pay greater attention to negative interpersonal dynamics, as they significantly impact employee well-being and organizational productivity.
Sam: How did they capture that combination without building the conclusion into the questions?
Alex: Officers completed questionnaires at randomly selected stations across the national police force. After missing data, the analysis sample contained 685 officers. The researchers used roughly half to develop the undermining measures, then tested those measures and their hypotheses in the remaining group.
Sam: So the people used to refine the questions were separate from those used to test the predictions. How large was that testing group?
Alex: The hypothesis tests used 343 officers. They reported how often their supervisor and closest coworker had engaged in particular behaviors during the previous month. One question asked whether the supervisor intentionally delayed work to make them look bad or slow them down.
Sam: That example connects the definition to something observable, although the intention still comes from the officer’s interpretation. What counted as support?
Alex: Support included actions intended to help with distress, share tasks, offer advice, teach skills, or provide material aid. The measures distinguished support from undermining, and supervisors from coworkers. The researchers then examined whether undermining added explanatory value beyond support and individual differences.
Sam: Which outcomes did they examine? A relationship could damage someone’s confidence without necessarily changing their behavior.
Alex: They assessed self-efficacy, meaning confidence in doing the job, and organizational commitment, meaning attachment to the employer. They also measured physical symptoms, including headaches. Finally, they examined counterproductive behaviors: active acts such as theft, and passive ones such as taking extended breaks.
Sam: Someone inclined toward negative feelings might report both worse relationships and more symptoms. Did the analysis address that possibility?
Alex: It controlled for general tendencies toward negative and positive feelings, as well as tenure. Undermining still explained additional variation beyond those factors and support. That reduces a particular alternative explanation, but it does not remove every possible reporting bias or unmeasured difference.
Sam: Let’s separate the findings. First, was undermining associated with worse outcomes even before considering the combination with support?
Alex: Generally, yes, but the source mattered. Supervisor undermining was associated with lower confidence and commitment; coworker undermining was not in those tests. Both sources were linked with active counterproductive behavior and physical complaints. Supervisor undermining was also associated with passive counterproductive behavior.
Sam: The authors also compare undermining with support. Was the negative side consistently more strongly related to outcomes?
Alex: For supervisors, undermining had the stronger association in four of the five comparisons. Commitment was the exception, with no significant difference between the relationships. The coworker comparison was less consistent: undermining was more strongly associated with active counterproductive behavior and physical complaints, but not the other outcomes.
Sam: Now the central result. Does saying the combination was worse mean support itself was generally harmful?
Alex: No. It means the relationship between undermining and outcomes depended on how much support came from that same source. For supervisors, high support combined with high undermining was associated with worse results across all the outcomes examined. The corresponding coworker pattern appeared only for confidence in doing the job.
Sam: The authors suggest conflicting messages could matter. Someone you rely on for help is also someone you perceive as damaging your position.
Alex: That is their interpretation, not a mechanism they demonstrated. They suggest uncertainty about the relationship may be difficult to manage. Undermining from an important supporter could also signal the possible loss of help the employee has come to depend on.
Sam: Could help from someone else compensate? A supportive coworker might offer something different from support supplied by the supervisor doing the undermining.
Alex: There were only two statistically supported examples of that protective pattern. Supervisor support weakened the link between coworker undermining and passive counterproductive behavior. Coworker support weakened the link between supervisor undermining and physical complaints. The authors describe this evidence as modest and urge caution.
Sam: How much explanatory weight did undermining carry overall? Statistical support alone does not tell us how much of workplace behavior this captures.
Alex: For active counterproductive behavior, undermining and its interactions explained an additional sixteen percent of variation after accounting for support and those individual differences. That makes it worth studying. But these were cross-sectional self-reports: relationships and outcomes were collected together, so the study cannot establish which came first.
Sam: Could employees’ behavior have helped shape the negative treatment they reported, rather than only resulting from it?
Alex: The authors explicitly raise that possibility, including counterproductive behavior preceding undermining. Independent reports would also help assess the measures. And generalization needs care: the sample was ninety-three percent male, within a national police force whose supervisor–subordinate power differences may be especially salient.
Sam: Given those limits, who should read the full document, and where should they start?
Alex: Researchers studying leadership, workplace relationships, or employee well-being should read it. Start with the definition of undermining and the measurement development. Then examine the same-source support comparisons alongside the limitations. For everyone else, remember this: support is not simply the opposite of undermining, and it should not be assumed to cancel it out.
Sam: Keep both sides of the relationship in view.