| Passive aggression (PA) is considered to be the most common form of aggression, and yet the most difficult to identify (Hershcovis et al., 2007). This study aims to validate an Organizational Passive-aggressive Scale (OPAS) as a three-factor model with a confirmation factor analysis (CFA) (Thomson, 2010). This research also investigates the relationship of organizational passive aggression with emotional intelligence and workplace deviance for construct validity.;For this study, data was collected with an online survey on the OPAS, Emotional Quotient Inventory: Short (EQ-i:S), Workplace Deviance Scale and demographical information from a sample of 198 working adults. A CFA with a three-factor model of organization passive aggression yielded a poor fit in the calibration and validation sample. Through a refinement process, two contributing factors (verbal aggression and defiance) revealed a good model fit with eight items: (chi2(19) = 28. 451, p = .052, CFI = .98, RMSEA = .052) on the calibration sample (n = 198). Additionally, a two-factor eight item OPAS provided good reliability (alpha = .84) as a new developed measure. However, cross-validation of the two-factor model with six items was not supported (chi2(8) = 22.704, p< .05, CFI = .88, RMSEA = .133) on the validation sample (n = 104) due to sample variance.;Construct validity for the OPAS was established when organizational passive-aggressive behavior was negatively related to emotional intelligence (r = -.356, p < .001, two-tailed) and positively related to work place deviance (r = .525, p < .001, two tailed). As predicted in hypotheses testing, the OPAS had small to negative effects on each of the sub components of EQ-i:S; adaptability (r = -.239, p < .001), general mood (r = -.269, p <.001), interpersonal skills (r = -.225, p < .001), intrapersonal (r = -.207, p < .001), and stress management (r = -.441, p < .001). Findings in this study advance our understanding of PA and suggest that the OPAS is a reliable and valid tool for researchers; however, additional research is encouraged since the two-factor model was not successfully cross-validated. |
