Serial multiple mediation of psychological empowerment and job burnout in the relationship between bullying experience and turnover intention among Chinese novice nurses in Triple A hospitals: A cross-sectional questionnaire survey
Description
To examine the direct and indirect effects of workplace bullying on turnover intention using the serial multiple mediators of psychological empowerment and job burnout. A convenience sample of 450 novice nurses from 5 Triple A hospitals in the S province of China were evaluated using self-reported questionnaires of the Negative Acts Questionnaire-Revised, the Psychological Empowerment Instrument, the Chinese Maslach Burnout Inventory, and the Turnover Intention Scale. The data analyses were performed using the SPSS software version 26.0 (SPSS Inc.) and the PROCESS macro. All the data sets were examined to ensure accuracy and completeness prior to any statistical analyses.
Files
Steps to reproduce
Descriptive statistics were performed to explain the general characteristics of the study participants. An independent-samples t-test and a one-way analysis of variance were calculated to control for the influence of any covariate (age, gender, education level, marital status, having children, academic title, and work experience) on the turnover intention. The relationships among the variables were assessed using Pearson correlation coefficients. Moreover, Mahalanobis distance values were employed to delete outliers exceeding the chi-square critical value. Hayes (2018) recommended that the assumptions of normality, linearity, and homoscedasticity should be checked before using the PROCESS macro; all three assumptions were satisfied after removing the outliers. The serial multiple mediation model 6 was performed to measure the effect of workplace bullying (independent variable) on turnover intention (dependent variable) through psychological empowerment and job burnout (two mediators). The bootstrap method was conducted to assess the significance of each indirect effect with 10,000 bootstrap samples and 95% percentile bootstrap confidence intervals (95% CI) not containing zero.