Predicting Organizational Actual Turnover Rates in the U.S. Federal Government

ABSTRACT This study expands the level of analysis on turnover in public administration, especially in the U.S. federal government, from the individual level to the organizational level by using organizational actual turnover rates. Some scholars argue that public employees' turnover intention generally reflects actual turnover. However, very little empirical evidence supporting this argument has been provided in public administration, especially in a broad array of public agencies. This study has rejected this argument by showing insignificant or weakly significant correlations between organizational actual turnover and weighted turnover intention rates. In addition, overall, the two regression results for organizational actual turnover and turnover intention also show different results from those in the existing literature on individual-level turnover intention. The significant predictors of organizational actual turnover rates are goal ambiguity, pay satisfaction, and diversity policy satisfaction. The correlation and regression results imply that research on predictors of turnover may need to consider the differences that may result from using different units of analysis and to make a distinction between turnover intention and actual turnover. Notes Note. N = 176 (federal agencies). *p < .05; **p < .01; ***p < .001. Note. Sample size = 176. *p < .10; **p < .05; ***p < .01; ****p < .001. This study recognizes some empirical studies examining the relationship between turnover intention and actual turnover at the individual level in the context of one public agency (e.g., Alley and Gould Citation1975; Mowday, Koberg, and McArthur Citation1984). However, the results were not consistent and most of them did not report a high correlation or a significant regression coefficient. Therefore, these results are not sufficient to support the assumption that public employees' turnover intention reflects actual turnover at the individual and agency levels. Regarding this issue, this study should only discuss empirical studies predicting turnover intention, though for the hypotheses there are a few studies predicting actual turnover. Considering the possibility that organizational (weighted) average turnover intention rates can be a predictor of organizational actual turnover rates or a mediator between the main independent variables and actual turnover rates, this study also ran the regression model including (weighted) average turnover intention rates as an independent variable of actual turnover rates. The results were very similar to the results of organizational actual turnover rates of Table 3. Specifically, statistical significance, influence directions, and influence magnitudes of all the independent variables still remained almost the same as those in the model presented in Table 3. In addition, weighted turnover intention did not have a significant relationship with organizational actual turnover rates. For these two models, the data are tested for assumptions of normality, multicollinearity, homoscedasticity, and model target-specification. The test results show that the assumptions of normality, homoscedasticity, and no-omitted variables are met for the models. In addition, all the Variance Inflation Factors (VIF) exist within the acceptable degrees (the largest VIF is 3.58), which indicates that there are no multicollinearity problems among the independent variables (Chun and Rainey Citation2005b). These findings might result from ecological inference, although this analysis uses survey weights to control for the biased estimates. Additional informationNotes on contributorsChan Su Jung Chan Su Jung (oraenby@gmail.com) is Visiting Assistant Professor in the Department of Public and Social Administration at the City University of Hong Kong. He received his PhD in the Department of Public Administration and Policy at the University of Georgia. His research interests include organizational goal properties such as goal ambiguity, performance measurement and management, turnover, motivation, and leadership in public organizations.