臺大管理論叢第31卷第3期

73 NTU Management Review Vol. 31 No. 3 Dec. 2021 Second, our study contributes to the leadership literature by identifying one important antecedent of abusive supervision and LMX quality. Prior leadership research focuses on investigating the consequences of leadership while paying considerably less attention to the antecedents of leadership. Research on abusive supervision and research on LMX are by no means exceptional (cf. Gerstner and Day, 1997; Martinko et al., 2013; Schyns and Schilling, 2013; Tepper, 2007; Zhang and Bednall, 2016). Until now, quite limited studies have examined the effects of leaders’ assumptions on abusive supervision and LMX (e.g., Tepper et al., 2011; Whiteley et al., 2012). On the basis of the attitude-behavior consistency principle, we propose herein that the NIFTs held by leaders encourage them to display abusive supervisory behaviors toward employees (Haddock and Maio, 2004). We also propose that the NIFTs held by leaders impede their LMX relationships with employees because the leaders engage in an initial attribution of negative traits to employees regarding their behaviors and performance (Sy, 2010). In line with our predictions, the results show that leaders’ NIFTs led to abusive supervision and low-quality LMX. As many leaders rely on fragmented information to construct subordinates’ images and use simplified prototypes to categorize subordinates (Junker and Van Dick, 2014), leaders’ IFTs may be quite subjective and severely biased. When leaders hold NIFTs, employees undoubtedly have to endure a very unfair evaluation and the accompanying unfair treatment. Because abusive supervision creates multiple psychological and work problems for employees (e.g., depression, self-regulation impairment, and low levels of work morale) (Martinko et al., 2013; Tepper, Simon, and Park, 2017), and because lowquality LMX results in employees’ job dissatisfaction, turnover intention, and withdrawal behaviors (Gerstner and Day, 1997; Hsiung and Bolino, 2018), identifying the factors that influence abusive supervision and LMX quality should be an important task in the research area of abusive supervision and LMX. Noteworthily, previous literature has found a significant and negative correlation between LMX and abusive supervision (-.54; see Mackey, Frieder, Brees, and Martinko, 2017). It is plausible that LMX is an antecedent of abusive supervision and that leaders’ NIFTs may indirectly affect employees’ perceptions of abusive supervision via a poor LMX relationship.6 Hence, leaders’ NIFTs may have both a direct and an indirect relationship with abusive supervision. 6 After we control for the effect of LMX on abusive supervision (γ = -.36, p < .001), we find that leaders’ NIFTs still have a significant direct effect on abusive supervision (γ = .23, p < .001).

RkJQdWJsaXNoZXIy ODg3MDU=