Why and when cognitive job insecurity relates to affective job insecurity? A three-study exploration of negative rumination and the tendency to negative gossip

ABSTRACT

The job insecurity literature distinguishes between cognitive job insecurity and affective job insecurity where cognitive job insecurity reflects perceptions regarding the likelihood of total job loss or job features loss and affective job insecurity refers to emotional reactions to that potential loss. Indeed, affective job insecurity is demonstrated to be an affective reaction to cognitive job insecurity. However, the relationship between cognitive job insecurity and affective job insecurity may be neither direct nor unconditional. Drawing from cognitive appraisal theory, this research takes a nuanced approach to exploring the mediating role of negative work rumination and the moderating role of the tendency to negative gossip in the relationship between cognitive job insecurity and affective job insecurity. We examined our hypotheses using three time-lagged survey studies with employees recruited from the U.S. and China. These studies found that negative work rumination mediated the relation between cognitive job insecurity and affective job insecurity (Studies 1–3) and the tendency to negative gossip attenuated the positive relation between cognitive job insecurity and affective job insecurity (Studies 1 and 2). Thus, this research advances the job insecurity literature by identifying a mediator and a moderator in the process of how employees may experience job insecurity.

KEYWORDS:

• Job insecurity
• cognitive job insecurity
• affective job insecurity
• rumination
• gossip

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

1. All three studies presented here were part of larger data collection efforts.

2. In all analyses below, we had attempted to control for participant gender and age. Yet, these demographic variables had no significant impacts on our outcome and therefore did not change the conclusion about our results. On the basis of Becker’s (2005) and Bernerth and Aguinis’s (2016) recommendation that the inclusion of unnecessary control variables reduces statistical power and yields biased estimates, we excluded these variables from our final analyses reported in the paper. Results with control variables are available upon request.

3. Because most measurements in Study 3 were the same as in Study 1, and Studies 1 and 3 recruited an American and a Chinese sample, respectively, we conducted measurement invariance tests using the procedure proposed by Cheung and Rensvold (2002). Specifically, we assessed configural invariance (form), metric invariance (loadings), and scalar invariance (intercepts). Invariance was assumed to hold if there was minimal difference in their fit by comparing the constrained model with the baseline model (Widaman & Reise, 1997). Because ΔCFI was less than 0.01 in the comparison of configural invariance and metric invariance tests (Cheung & Rensvold, 2002), we found evidence of configural invariance and metric invariance. Evidence of configural invariance indicates that both samples used the same conceptual frame of reference (i.e., exhibited the same factor structure), whereas evidence of metric invariance means that in addition to an invariant factor pattern, the loadings were also equivalent (i.e., each item contributes to the latent construct to a similar degree across samples). However, we did not find evidence of scalar invariance; that is, factor intercept differences were demonstrated across two countries. As such, we only found weak invariance but not strong invariance (Boer, Hanke, & He, 2018).

Additional information

Funding

The last author’s contribution is supported by the National Natural Science Foundation of China (Project No.71701074; 71832004)