Antecedents of turnover intention: A meta-analysis study in the United States

Abstract

The purpose of this paper was to determine the direction and effect size of the relationships between turnover intention and its main antecedents in the United States. The main predictors of turnover intention are chosen as job satisfaction, organizational commitment and empowerment. The studies which are published between 1998 and 2018 are reviewed. ScienceDirect, Scopus, and ProQuest databases were searched and 2.356 studies are screened. The meta-analysis software package, Comprehensive Meta-analysis Software (CMA), was used for the meta-analysis. 101 studies were suitable and the three data sets are formed: fi rst set included 312,261 subjects and 91 studies relating job satisfaction to turnover intention, the second set included 13,502 subjects and 29 studies relating organizational commitment to turnover intention, and the third set included 997 subjects and 5 studies relating empowerment to turnover intention. Each data was heterogeneous signifi cantly and the random effects model was used. Publication bias is analyzed for each data set and no evidence of publication bias was detected. The results revealed that the overall relationship between turnover intention and the selected three constructs was negative and signifi cant: the power of the job satisfaction's effect and organizational commitment's effect are almost the same (-.52), and the power of the empowerment's effect is weaker (-.22). Job satisfaction and organizational commitment have a large impact and empowerment has a small effect on turnover intention. The moderator analysis determined that type of industry and region are the moderators affecting the relationship between job satisfaction and turnover intention. The fi ndings also provide guidance for the managers working in the United States who need to keep turnover under control.

A meta‐analysis on predictors of turnover intention of hospital nurses in South Korea (2000–2020)

“…In contrast, factors that prevented nurses’ turnover intention were work engagement, job satisfaction and professional self‐concept. The finding that individual‐level predictors affect nurses’ turnover intention was consistent with findings in studies conducted in other countries (Chiang & Chang, 2012 ; Deery et al., 2011 ; Fasbender et al., 2019 ; Özkan et al., 2020 ; Simone et al., 2018 ; Zhao et al., 2018 ). In particular, job stress was shown to be prevalent in nursing positions and highly related to emotional exhaustion, depression, fatigue and burnout (Chiang & Chang, 2012 ; Trybou et al., 2014 ).…”

“…In the present review, 170 related factors were identified in 417 studies that investigated the predictors of the turnover intention in hospital nurses. In the past 10 years, the number of relevant stud- Han & Jekel, 2011; Hayes et al, 2006;Oyeleye et al, 2013;Özkan et al, 2020;Simone et al, 2018;Yin & Yang, 2002).…”

“…This indicates that several important factors influencing turnover intention in hospital nurses have been measured repeatedly. As such, even though the effect sizes may differ, predictors that were repeatedly measured, such as job satisfaction, job stress, burnout, organizational commitment, emotional labour, work environment, job embeddedness and resilience, have been shown to be statistically significant predictors of nurses’ turnover intention worldwide (Brewer et al., 2012 ; Chiang & Chang, 2012 ; Deery et al., 2011 ; Foster et al., 2020 ; Han & Jekel, 2011 ; Hayes et al., 2006 ; Oyeleye et al., 2013 ; Özkan et al., 2020 ; Simone et al., 2018 ; Yin & Yang, 2002 ).…”

“…Second, the present study included research that used instruments with multiple items measured using continuous variables and had proven reliability in the analysis, without limiting turnover intention scales. However, most previous meta‐analysis studies did not limit turnover intention scales either (Özkan et al., 2020 ; Yin & Yang, 2002 ). In fact, a random‐effects model was used because the effect sizes of most predictors were heterogeneous, regardless of whether a specific scale was used (Lee & Kang, 2018 ) or the scales used were not limited (Özkan et al., 2020 ).…”

“…However, most previous meta‐analysis studies did not limit turnover intention scales either (Özkan et al., 2020 ; Yin & Yang, 2002 ). In fact, a random‐effects model was used because the effect sizes of most predictors were heterogeneous, regardless of whether a specific scale was used (Lee & Kang, 2018 ) or the scales used were not limited (Özkan et al., 2020 ). The range of the confidence interval for the random‐effects model increases because the model reflected heterogeneity among studies, and weights for differences in the number of samples in each study are relaxed.…”