Does Turnover Intention Matter? Evaluating the Usefulness of Turnover Intention Rate as a Predictor of Actual Turnover Rate

Abstract

Turnover research has traditionally examined intention to turnover rather than actual turnover. Such studies assume that leave intent serves equally well as both a proxy for and predictor of employees’ actual turnover behavior. The purpose of this study is to provide an agency-level evaluation of the usefulness of turnover intention as a reliable proxy and predictor of actual turnover across 180 U.S. federal agencies, using hierarchical (stepwise) multiple regression. Our findings suggest that, at the organizational level, turnover intention and actual turnover are distinct concepts, predicted by different sets of variables. Based on these findings, we conclude that public managers tasked with retention might have better foresight concentrating on their agencies’ unique demographic characteristics and specific management practices, rather than on their employees’ self-reported aggregated turnover intention rate.

Keywords

turnover federal government human resource management

Introduction

Human capital planning in the federal government mostly relies on measuring employees’ future intention to leave (Broach & Dollar, 2006). However, studies that empirically examined the relationship between intention to turnover and actual turnover are scarce and demonstrate conflicting results in regard to the usefulness of intentions as a reliable proxy of behavior (Cho & Lewis, 2012; Jung, 2010; Kirschenbaum & Weisberg, 1990). In particular, some scholars have found that turnover intention is a poor predictor of actual turnover (e.g., Jung, 2010; Kirschenbaum & Weisberg, 1990; T. W. Lee & Mowday, 1987). In this study, we explore the relationship between turnover intention rates and actual turnover rates of U.S. federal agencies.

This research seeks to contribute to the emerging body of public employee turnover research by addressing a heretofore-neglected aspect of empirical study (Meier & Hicklin, 2008; Selden & Moynihan, 2000). Moreover, the vast majority of the already small number of public administration studies on turnover have mostly been using turnover intention as the dependent variable rather than actual turnover (Jung, 2010).

In addition, although previous turnover research has been mostly conducted at the individual level, very little empirical research has examined turnover at the organizational level. This lack of research is a matter of practical concern. For, although organizational withdrawal is a personal decision affected by socio-psychological considerations and each individual’s own unique circumstances, employee retention, recruitment, and training are strategic human resource management functions necessarily administered at the organizational level (Ingraham & Rubaii-Barrett, 2007; Perry, Hondeghem, & Wise, 2010; Van Marrewijk & Timmers, 2003). As Hausknecht and Trevor (2011) pointed out, turnover analysis at the organizational level is much more consistent with the way HR managers and leaders prefer to learn about turnover in their organizations (Gardner, Moynihan, & Wright, 2007). As a tenuous step toward the advancement of understanding organizational turnover, this study uses agency-level data for the analysis.

In this study, we explore the relationship between the turnover intention rate and actual rate of federal government agencies in the following ways. First, and most fundamentally, we assess whether agencies’ turnover intention rate and agencies’ actual turnover rate correlate within our sample. Second, we investigate the relative impacts of organizational-level perceptions toward HR practices on federal agencies’ actual turnover rates. The ultimate goals of HR practices are to increase organizational effectiveness and decrease actual employee turnover rate (Gardner et al., 2007; Gould-Williams, 2004), but limited number of studies have addressed this relationship in public administration research (Farnham & Giles, 1996; Hays & Kearney, 2001; Gould-Williams, 2004; Cho & Lewis, 2012). Thus, this research evaluates the impact of perceptions toward HR practices on agency turnover.

Third, we explore whether the organizational-level determinants that best explain agencies’ actual turnover rates also explain turnover intention rates. If intention rate is a reliable proxy for actual turnover rate, then the same patterns should hold for both dependent variables. Finally, we assess whether turnover intention rate actually predicts agencies’ actual turnover rate.

Antecedents of Employee Turnover Rates and Hypotheses

The “Actual–intention” link

Analysis of leave intentions has been a mainstay of the general turnover research since its advent (Cho & Lewis, 2012; Dalton, Johnson, & Daily, 1999; Kirschenbaum & Weisberg, 1990). The empirical turnover literature is replete with examples of turnover behavior that is inferred based on analyses of employees’ leave intentions and its correlates. Even as turnover models become increasingly sophisticated, this conceptual linkage between intent and actual turnover has remained cardinal. Such research is premised on the vital link between attitude and behavior, and thus, on the assumption that intent is the best predictor of actual turnover (e.g., Bertelli, 2007; Dalton et al., 1999; S. Y. Lee & Whitford, 2007; Steel & Ovalle, 1984; Tett & Meyer, 1993).

The rationale justifying intentions’ use as a turnover proxy is twofold. First, from a theoretical perspective, attitude theory generally supports the belief that intent is the best predictor of behavior (Kraut, 1975; Mobley, Horner, & Hollingsworth, 1978; Price & Mueller, 1981). As Fishbein and Ajzen (1975) wrote, “The best single predictor of an individual’s behavior will be a measure of his intention to perform that behavior” (p. 369). According to this line of research, turnover intention is expected to be the strongest predictor of actual turnover behavior (e.g., Currivan, 1999; Griffeth, Hom, & Gaertner, 2000; Hom, Griffeth, & Sellaro, 1984; S. Y. Lee & Whitford, 2007; Mobley, 1977; Vandenberg & Nelson, 1999). This theoretical expectation has empirical support. For instance, in a meta-analysis of job attitudes and behavior, Harrison, Newman, and Roth (2006) concluded that job attitudes, such as turnover intentions, reliably predict job behaviors, such as quitting.

Second, turnover scholars also rely on intentions for pragmatic reasons. As a surrogate, the intent construct is more amenable to research than actual turnover. It possesses desirable statistical qualities (i.e., easily scaled) and is more economic (Dalton et al., 1999). Conversely, the actual turnover construct is a dichotomous variable that generally requires costly longitudinal designs to fully assess. Most important, surveys are typically administered anonymously. Thus, connecting information gleaned from them to individuals’ actual behaviors is usually impossible and tends to be fraught by ethical implications (Dalton et al., 1999).

For these reasons, scholars commonly use turnover intention as a proxy of actual turnover (e.g., Bertelli, 2007; Kim, 2005; S. Y. Lee & Whitford, 2007; Pitts, Marvel, & Fernandez, 2011). This is true of turnover studies in general (Griffeth et al., 2000) and especially true of public-sector studies (Jung, 2010; Tett & Meyer, 1993).

Generally speaking, scholars have found that employees turnover intentions and quit behaviors to be statistically correlated. Findings as to the strength of the relationship, however, are inconclusive. Some studies report finding the constructs strongly and directly correlated (e.g., Griffeth et al., 2000; Hom et al., 1984; S. Y. Lee & Whitford, 2007; Mobley, 1977; Steel & Ovalle, 1984). For instance, based on a random sample drawn from the U.S. Office of Personnel Management’s (OPM) Central Personnel Data File, Cho and Lewis (2012) found a correlation of .80. Other studies, however, have found the relationship to be much weaker and even insignificant. T. W. Lee and Mowday (1987) found employees’ intentions explained only about 6% of turnover variance. Also, Kirschenbaum and Weisberg (1990) found a poor and non-significant relationship between intention to leave and actual turnover behavior. According to them, survey responses to whether one’s intent to leave his or her job or not, cannot actually attest to real future behavior. Finally, there is also research showing that the relationship between employees’ leave intentions and actual separation behaviors is incidental or even non-existent (cf. Jung, 2010; Kirschenbaum & Weisberg, 1990).

These finding suggest that although there is a general consensus among scholars about the positive relationship between quit intention and actual quitting, with regard to the strength of this relationship, alarming discrepancies exist. Therefore, this study examines the relationship between the turnover intention rates and actual turnover rates of U.S. federal agencies.

Hypothesis 1: An agency’s average turnover intention rate is positively associated with the average actual turnover rate of the agency.

Collective member perceptions

According to Hausknecht and Trevor (2011), a major category of turnover antecedents includes employees’ aggregated attitudes and perceptions of organizational characteristics, such as the quality management, HR practices, and organizational culture and climate. These predictors are well recognized in both the individual-level and organizational-level literature, which is generally based on turnover intention. In this study, we seek to explore whether these known predictors of turnover intention also apply to actual turnover in the organizational level.

Considering these various kinds of perceptional determinants of turnover, the collective perceptions regarding six types of organizational practices were chosen as independent variables for this research: telework, performance culture, pay satisfaction, advancement opportunities, workload, and flexible work schedule. These six practices were selected because they have been frequently identified within the public personnel management literature as factors related to employees’ perceptions of an organization’s HR management practices. The rationale underlying incorporation of these constructs is the theoretical expectation that effective HR management practices increase employees’ work motivation and thus reduce turnover (Wright & Boswell, 2002).

Telecommuting and telework describe remote working arrangements, where modern information technologies allow employees to perform tasks and fulfill transactional obligations away from centralized or physical organizational locations (Belanger & Collins, 1998). In recent years, due to enabling legislation and guidelines (e.g., OPM, 2005; U.S. Congress House, 1999), adoption of telework has become increasingly common across the U.S. federal government (Gajendran & Harrison, 2007). Organizational benefits accrued from the arrangement include reduced real estate expenses and reduced costs for compliance with regulations such as those associated with the Americans With Disabilities Act of 1990. Telecommuting benefits employees by affording them some measure of flexibility and control over when and from where they may fulfill their job requirements (Cayer, 2003; S. Y. Lee & Hong, 2011). Studies evaluating telework programs have generally found them to increase employees’ motivation and productivity while reducing absenteeism and turnover (Iscan & Naktiyok, 2005).

Hypothesis 2: An agency’s average level of satisfaction with telework practices is negatively associated with the actual turnover rate of the agency.

Performance management practices are implemented to connect employees’ performance to rewards through continuous feedback and ongoing evaluation (Kettl, 2005). G. Lee and Jimenez (2011) describe adoption of performance-oriented management practices as one of the most important HR developments to have occurred in public sector over the last two decades. Holding that employees value distributive justice, social exchange theory predicts that effectively administered management systems emphasizing performance-based rewards should be associated with reduced organizational turnover. Empirical studies supporting this posited linkage include Huselid (1995), who found significantly reduced turnover in organizations with strong performance practices, and Pitts et al. (2011) found enhanced job satisfaction, productivity, and retention in organizations valuing and rewarding performance.

Hypothesis 3: An agency’s mean level of satisfaction performance culture practices is negatively associated with the actual turnover rate of the agency.

A substantial body of individual-level empirical research indicates job satisfaction measures are among turnover’s strongest correlates (e.g., Bertelli, 2007; Bright, 2008; Carsten & Spector, 1987; Cotton & Tuttle, 1986; Lambert, Hogan, & Barton, 2001; Mobley, Griffeth, Hand, & Meglino, 1979; Porter & Steers, 1973). As increased job satisfaction is associated with individuals reduced propensity for withdrawal, the same conceptual logic extends to organizations. That is, as employees’ collective job satisfaction increases, organizational turnover decreases (Hausknecht & Trevor, 2011). This theoretical expectation is based, in part, on Hackman and Oldham’s (1976) classic model of work motivation. From this perspective, specific job elements (e.g., salary, benefits, opportunities, certain duties, and tasks) positively alter employees’ psychological state, thereby increasing work motivation, performance, job satisfaction, and their likelihood to stay in an organization.

Among satisfaction constructs, satisfaction with pay is one consistently associated with reduced voluntary turnover (Blau & Kahn, 1981; Cotton & Tuttle, 1986; Lambert et al., 2001; Park, Ofori-Dankwa, & Bishop, 1994; Shaw, Delery, Jenkins, & Gupta, 1998). Employees maximize self-interest through higher wages. Consequently, employees who perceive that their wages are highly satisfactory tend to remain in organizations (Shaw et al., 1998). Satisfaction with pay also reduces individuals’ financial anxieties (Lambert et al., 2001) and, thus, decreases job-search motivations (Blau & Kahn, 1981).

Hypothesis 4: An agency’s average level of pay satisfaction is negatively associated with the actual turnover rate of the agency.

Satisfaction with career advancement opportunities is also negatively associated with voluntary turnover (Cotton & Tuttle, 1986; Griffeth et al., 2000; Porter & Steers, 1973; Spector, 1985). Promotions are usually linked to salary increases, which, in turn, affect reduced exits (Johnston, Griffeth, Burton, & Carson, 1993). Observing rigid criteria complicate public-sector employee promotions, Pitts et al. (2011) argued that satisfaction with advancement opportunities is a key factor affecting federal employees’ overall job satisfaction and, consequently, turnover intentions.

Hypothesis 5: An agency’s average level of satisfaction with advancement opportunities is negatively associated with the actual turnover rate of the agency.

Work schedule and workload satisfaction also theoretically correlates with turnover. Where alternative work schedules enhance employees’ flexibility and sense of control, social exchange theory predicts increased commitment and reduced turnover (S. Y. Lee & Hong, 2011). Similarly, social exchange theory predicts that workload satisfaction is associated with higher moral and reduced turnover (cf. Anderson, Corazzini, & McDaniel, 2004; Banaszak-Holl & Hines, 1996).

Hypothesis 6: An agency’s average level of satisfaction with workload is negatively associated with the actual turnover rate of the agency.

Hypothesis 7: An agency’s average level of satisfaction with flexible work schedule is negatively associated with the actual turnover rate of the agency.

Collective member characteristics

Organization-level demographic characteristics strongly associate with organizational behavior such as employee turnover (Pfeffer, 1985). In some cases, demographic factors have been evaluated independently as explanatory variables (e.g., Cho & Lewis, 2012; Jung, 2010; Pitts et al., 2011), whereas, in other cases, they are operationalized as controls (Hausknecht & Trevor, 2011).

Age is inversely associated with turnover and is prominently featured within the empirical literature (e.g., Cho & Lewis, 2012; Jung, 2010; Kellough & Osuna, 1995; Pitts et al., 2011). Previous research has demonstrated that the relationship between age and turnover is curvilinear (e.g., Cho & Lewis, 2012). Turnover among young employees tends to be very high, but progressively decreases as employees age (Lewis, 1991; Lewis & Park, 1989). At the organizational level, in government, Kellough and Osuna (1995) observed proportions of younger employees within an agency’s workforce (i.e., less the 32 years old) correlate positively with agency’s quit rates. A common socio-psychological explanation for this relationship is that older employees tend to be more risk adverse, less mobile (Moynihan & Landuyt, 2008), and more constrained by family and financial obligations (Ippolito, 1987; O’Reilly, Chatman, & Caldwell, 1991).

Studies have also found that job tenure is inversely associated with turnover both at the individual level (e.g., Blau & Kahn, 1981; Lewis, 1991) and organization level (e.g., Bennett, Blum, Long, & Roman, 1993; Glebbeek & Bax, 2004; Hausknecht, Trevor, & Howard, 2009; Spell & Blum, 2005; Terborg & Lee, 1984; Trevor & Nyberg, 2008; Wiersema & Bird, 1993; Yanadori & Kato, 2007). By way of explanation, Farber (1999) argued that less tenured employees quit more frequently because they are younger and earn less. Conversely, Ippolito (1987) argued more tenured employees quit less frequently because they tend to be more vested in pension plans that are forfeited on withdrawal. Another explanation is that as employers invest in training and workers’ skills become increasingly firm-specific, comparable job alternatives become increasingly scarce.

Workforce diversity is another important determinant of organizational turnover. Theory holds that as demographic diversity increases, psychological attachment and group commitment decreases (cf. Greenhaus, Parasuraman, & Wormley, 1990; Sackett, DuBois, & Noe, 1991; Tsui, Egan, & O’Reilly, 1992). Turnover research evaluating gender differences has generally found that women, compared with men, are absent more frequently, have lower intentions to stay, and voluntarily terminate employment relationships more frequently (Bae & Goodman, 2014; Choi, 2009). Similarly, turnover has been linked to racial and minority status. Although the relationship is a complicated one, at the organizational level, the relationship between minority status and turnover has generally been found to be a positive one (Hausknecht & Trevor, 2011).

Data and Method

Data

This study combines and aggregates data from several sources, with the unit of analysis being the agency. Data for the independent variables are drawn from the 2010 Federal Employee Viewpoint Survey and the 2010 FedScope Employment Cube. The government-wide response rate for the survey is 52% (n = 263,475). Data for the dependent variables are drawn from the 2010 and 2011 FedScope Separation Cubes (www.fedscope.opm.gov) in the 12-month period immediately following the administration of the 2010 Federal Employee Viewpoint Survey. Overall, our sample includes data from 180 different U.S. federal agencies. Table 1 reports summary statistics for the dependent and independent variables of this study.

Table 1.

Descriptive Statistics of the Variables (N = 180).

Variables	Explanations	M	SD	Minimum	Maximum
Actual turnover rate	Proportion of workforce	0.017	0.011	0.002	0.065
Turnover intention rate	Weighted proportion (Q88)	0.020	0.012	0.002	0.106
Average tenure	Years	16.01	2.71	7.06	24.00
Population < 30 years old	Proportion of workforce	0.079	0.053	0.016	0.458
Population 30 to 49 years old	Proportion of workforce	0.477	0.079	0.301	0.787
Female	Proportion of workforce	0.451	0.145	0.124	0.738
Non-White minority	Proportion of workforce	0.337	0.130	0.050	0.906
Professional/administrative	Proportion of workforce	0.766	0.205	0.153	0.995
Telecommuters	Weighted proportion (Q72)	0.299	0.192	0.027	0.757
Performance culture	Weighted composite average	3.082	0.214	2.342	3.774
Workload satisfaction	Weighted Average (Q10)	3.369	0.199	2.806	4.004
Opportunity satisfaction	Weighted average (Q67)	3.089	0.200	2.389	3.697
Pay satisfaction	Weighted average (Q70)	3.724	0.194	3.020	4.331
Work schedule satisfaction	Weighted average (Q74)	3.794	0.383	2.423	4.580

Dependent Variable

Actual turnover rates are calculated as the total number of “quits” divided by the total number of full-time, permanent employees for each agency. The data have been log-transformed to create a normal distribution. In addition, actual turnover rate of an agency includes only voluntary withdrawals (i.e., “quits”). Involuntary terminations and voluntary separations for other causes, such as interagency transfers and retirements, are excluded.

Main Independent Variable

Our main independent variable, turnover intention rate, serves both as a proxy and predictor of agencies’ actual turnover rate. The question that addresses these variables in the survey is as follows: “Are you considering leaving your organization within the next year, and if so, why?” The possible answers are “No”; “Yes, to retire”; “Yes, to take another job within the federal government”; “Yes, to take another job outside the Federal Government”; and “Yes, other.” Agency’s turnover intention rate reflects the proportion of employees’ responding “Yes, to take another job outside the federal government.”¹ To compensate for under- and over-represented populations within the sample, all variables’ aggregations were calculated using OPM’s (2011) weighting methodologies.

Independent Variables

In addition to our main independent variable, the relationship between six theoretical constructs related to management practices and employees’ shared perceptions and actual turnover rate is tested. To measure these variables, we also aggregate employee responses to the agency level by using survey items from the 2010 Federal Employee Viewpoint Survey. Employee responses are gauged based on a Likert-style scale ranging from 1 (strongly disagree, very dissatisfied, very poor) to 5 (strongly agree, very satisfied, very good). To compensate for varying group response rates, individual ratings are aggregated using OPM’s weighting methodology. This procedure involves multiplying each response by its corresponding weight, summing the products, and dividing the result by the sum of each agency’s weights (OPM, 2011).

The measure of performance culture perception is constructed from four items: “Promotions in my work unit are based on merit”; “Awards in my work unit depend on how well employees perform their jobs”; “Employees are recognized for providing high quality products and services”; and “Pay raises depend on how well employees perform their jobs” (adopted from Fernandez & Moldogaziev, 2010). The Cronbach’s alpha of the four items is .95.

For workload satisfaction, opportunity satisfaction, pay satisfaction, and work schedule satisfaction, the following single items are used: “My workload is reasonable” (1 = strongly disagree to 5 = strongly agree); “How satisfied are you with your opportunity to get a better job in your organization”; “Considering everything, how satisfied are you with your pay”; and “How satisfied are you with the following Work/Life programs in your agency . . . Alternative Work Schedules (AWS)” (1 = very dissatisfied to 5 = very satisfied). For all these variables, employee responses are aggregated to the agency level using the survey weights.

The telecommuters variable is measured using the following survey item: “Please select the response below that BEST describes your teleworking situation.” The possible answers are as follows: “I telework on a regular basis (at least one entire work day a week),” “I telework infrequently (less than one entire work day a week),” “I do not telework because I have to be physically present on the job (e.g., Law Enforcement Officers, Park Rangers, Security Personnel),” “I do not telework because I have technical issues (e.g., connectivity, inadequate equipment) that prevent me from teleworking,” “I do not telework because I am not allowed to, even though I have the kind of job where I can telework,” and “I do not telework because I choose not to telework.” This variable reflects the proportion of agency respondents indicating the following: “I telework on a regular basis (at least one entire work day a week).”

Control Variables

This study controls for four collective member characteristics that might be antecedents of employee turnover: age, tenure, gender, and minority status. All the demographic controls, with the exception of minority status, are extrapolated from the raw individual-level FedScope employment data for March 2010 (www.opm.gov/data/Index.aspx). Proportions of “non-White minority” workers are taken directly from tables available on OPM’s FedScope website (see www.fedscope.opm.gov).

Employees’ precise ages (Age) are not available within the FedScope data; rather, they are specified by age group. Based on these data, we model two age cohorts: young employees, defined as those less than 30 years of age, and mid-career employees, defined as employees aged 30 to 49. The study’s reference group is thus composed of full-time, permanent employees aged 50 and older. Our average tenure variable is measured in years and operationalized as employees’ total combined years of service divided by agencies’ respective workforce populations in 2010. Gender is operationalized as the proportion of females in the workforce, and minority status is analyzed as proportions of non-White employees.

Statistical Approach

This study uses hierarchical (stepwise) multiple regression to model the variance of two continuous dependent variables, quit rate and quit intention rate, across 180 federal agencies. Hierarchical multiple regression is essentially a series of consecutive ordinary least squares (OLS) regression models, each adding a new set of predictors (Lewis, 2007). This statistical method allows us to assess the impact of each set of predictors above and beyond the previous set that was entered into the regression model (Lewis, 2007). Hierarchical regression is the practice of building successive linear regression models, each adding more predictors and is different from hierarchical linear modeling (HLM), which is the practice of using multi-level models.

To normalize distribution of residuals, log transformation was conducted for the turnover rates variables creating semi-elasticity (i.e., log–lin) models for this study. Because measurements of the study’s variables are mixed (i.e., years, proportions, and aggregated survey responses), for interpretive ease, we analyze the study’s variables as zero-centered measures. Meaning, the standardized regression coefficients (i.e., “beta”) represent constant proportional changes given the single standard deviation change of a respective regressor.²

This study’s ultimate objective is to assess the usefulness of agencies’ turnover intention rate as a proxy of agencies’ actual turnover rate. To accomplish this, a three-stage analytical procedure is performed. At the first stage, turnover intention rate operates as a sole independent variable so as to assess its direct contribution to predicting actual turnover rate (see Table 2).

Table 2.

Bivariate Regression Results for Actual Turnover Rate (logged; N = 180).

	β	SE
Turnover intention rate	.136*	0.046
Constant	−4.273	0.046
R ²	.047
Adjusted R²	.042
F(1, 178)	8.77
Prob > F	.004

Note. Standardized coefficient; SE = standard error.

p < .05.

At the second stage, two hierarchical regression models are executed (see Table 3). The first model (Model 1), comprised of two blocks of predictors (i.e., collective member characteristics; collective member perceptions), is regressed against actual turnover rate to determine its best predictors. In the second model (Model 2), the turnover intention variable is designated as the dependent variable with the purpose of evaluating whether the same predictors of actual turnover (i.e., the same collective member characteristics and collective member perceptions) also predict turnover intention rate.

Table 3.

Log–Lin Regression Results (N = 180).

	β	SE	β	SE
A.	Model 1a: Actual turnover rate		Model 2a: Turnover intention rate
Block 1
Average tenure	−.411***	0.045	−.076	0.053
Population < 30 years old	.034	0.040	.015	0.048
Population 30 to 49 years old	−.119***	0.040	.040	0.047
Female	.120**	0.044	.062	0.052
Non-White minority	−.012	0.041	−.117**	0.048
Professional/administrative	−.112**	0.039	.071	0.047
Constant	−4.273***	0.035	−4.047	0.041
R ²	.465		.080
Adjusted R²	.447		.049
Prob > F	25.09***		2.52**
	β	SE	β	SE
B.	Model 1b: Actual turnover rate		Model 2b: Turnover intention rate
Block 1
Average tenure	−.318***	0.044	−.040	0.058
Population < 30 years old	.100**	0.039	.060	0.051
Population 30 to 49 years old	−.062*	0.037	.065	0.049
Female	.096**	0.040	.021	0.053
Non-White minority	−.019	0.036	−.130***	0.048
Professional/administrative	−.149***	0.045	.098	0.060
Block 2
Telecommuters	.145***	0.045	.101*	0.060
Performance culture	.286***	0.060	−.005	0.079
Workload satisfaction	−.083**	0.038	−.117**	0.050
Advancement opportunity sat	−.150**	0.064	.009	0.085
Pay satisfaction	−.170***	0.052	−.127*	0.069
Flexible work schedule sat	−.130***	0.044	−.048	0.059
Constant	−4.273***	0.030	−4.047***	0.040
R ²	.618		.178
Adjusted R²	.591		.119
Prob > F	22.51***		3.02***
	β	SE	β	SE
C.	Model 1c: Actual turnover rate
Block 1
Average tenure	−0.317***	0.044
Population < 30 years old	0.093**	0.039
Population 30 to 49 years old	−0.068*	0.037
Female	0.094**	0.040
Non-White minority	−0.008	0.037
Professional/administrative	−0.153***	0.045
Block 2
Telecommuters	0.133***	0.046
Performance culture	0.279***	0.060
Workload satisfaction	−0.072*	0.038
Advancement opportunity sat	−0.144**	0.064
Pay satisfaction	−0.166***	0.052
Flexible work schedule sat	−0.116**	0.045
Block 3
Turnover intention rate	0.047	0.034
Constant	−4.273***	0.030
R ²		.622
Adjusted R²		.593
Prob > F		21.03***

Note. Beta represents standardized coefficients. SE = standard error.

p < .1. **p < .05. ***p < .01.

At the third stage, we elaborate Model 1 (the “actual turnover model”) by adding a turnover intention as a third block of independent variable to determine the marginal contribution of agencies’ turnover intention rate as a predicator of agencies’ future actual turnover rate, when holding all other factors constant. The next section brings the comparative results from the two models.³

Organizational-Level Results

The Direct Effect of Turnover Intention Rate on Actual Turnover Rate

As expected, agencies’ turnover intention rate and actual turnover rate correlate positively within our sample. Each standard deviation increase of intention to quit rate corresponds to a constant proportional quit-rate increase of 13.6% (i.e., 1.136 times). Table 2 provides the results from the bivariate regression between turnover intention rates and actual turnover rate. Although the bivariate regression model is statistically significant (p < .001), supporting Hypothesis 1, theory posits that the linkage between employees’ intent and actual withdrawals should be a strong one. In our sample, this is not the case as agencies’ turnover intention rate explains only 4.2% of quit rate variance across the federal government. This initial result does not bode well for intentions’ utility as a proxy, thus reinforcing this study’s approach to examine these constructs separately.

Collective Member Characteristics and Turnover Rate

Table 3 presents regression results for Model 1 and Model 2. In Model 1, actual turnover rate operates as the sole dependent variable, and in Model 2, turnover intention rate is designated as the dependent variable with the ultimate purpose of evaluating whether these two variables represent the same construct.

Due to the nature of our hierarchical models, we commence our discussion with the control variables (see Table 3, Part A). At the first stage, six collective demographic variables are introduced. For Model 1a (i.e., the actual turnover rate model), four collective demographic variables—average tenure, mid-career workforce, female, and occupation type—had a significant effect on agency actual turnover rate at. 05 level. Average tenure is most strongly associated with actual turnover rate (p < .000), and its effect is the greatest. With each standard deviation increase of average tenure, agencies’ predicted quit rates decrease by 41.1%, or are 4.11 times less likely to quit. Jointly, the six demographic factors specified within Model 1a explain about 45% of the variance in actual turnover across federal government agencies.

Our preliminary argument suggests that if turnover intention is a reliable proxy for actual turnover, then a set of variables explaining actual quit rates will similarly explain intention rates. However, when comparing Models 1a and 2a, the results differ from that which is expected. In Model 2a, although the variables in the model remain jointly significant (p = .023), only one demographic characteristic, non-White minority, is significant beyond a probability of .10 level. In fact, with the exception of non-White minority, all the variables’ effects are substantially diminished in Model 2a. As a result, the demographic features that explained 44.7% of agencies’ actual turnover rate in Model 1a explain less than 5% of agencies’ turnover intention rate in Model 2a. Interestingly, the member characteristics that explained a substantial portion of actual turnover rate variance in Model 1a explain employees’ turnover intention differently and less thoroughly in Model 2a. One implication of this comparison is that, at the organizational level, the linkage between employees’ aggregated intentions and their aggregated actual behaviors may be a dubious one. Also, researchers and practitioners using intention data in place of actual turnover data might erroneously conclude that these collective member characteristics are statistically unimportant, when in fact they are.

Collective Member Perceptions and Turnover Rate

Thus far, we find that turnover intention rate is not strongly correlated with actual turnover rate, and that member demographic characteristics explain actual turnover rate variance better than they explain turnover intention variance. Our next goal is to further evaluate the usefulness of turnover intention rate as a proxy of actual turnover rate by adding another set of turnover intention predictors—collective member perceptions—to the regression model. To do so, we add six collective member perceptions toward HR practices to the baseline model: telecommuters, performance culture, workload satisfaction, advancement opportunity satisfaction, pay satisfaction, and flexible work schedule satisfaction. As previously explained, the models similarly specified with the only difference between them being the dependent variable (i.e., actual turnover rate and turnover intention rate).

Part B of Table 3 presents the agency-level OLS regression results for actual turnover (Model 1b) and turnover intention (Model 2b) rates using survey weights. Overall, Models 1b and 2b are significant (p < .000 for actual turnover rate and p < .001 for turnover intention rate). Model 1b explains 59.1% of actual turnover variance across the federal government, whereas Model 2b combines to explain 11.9% of agencies’ turnover intention rate.

For Model 1b, as anticipated, agencies’ pay satisfaction, advancement opportunity satisfaction, workload satisfaction, and work schedule satisfaction are negatively correlated with the agencies’ actual turnover rate, supporting Hypotheses 4 through 7. Meaning, that agencies’ predicted quit rates decrease as agency employees’ satisfaction collectively increases.

Interestingly, although statistically significant, telecommuters and performance culture satisfaction do not correlate at the expected direction; thus, Hypotheses 2 and 3 are partially rejected. We expected actual turnover rate to decrease as telecommuting and performance culture perceptions increased. Instead, as Model 1b shows, both variables correlate positively with actual turnover rate. In other words, holding all other factors constant, a standard deviation increase in the proportion of telecommuters yields a quit-rate increase of 14.5%. Similarly, a standard deviation increase of the performance culture index corresponds to a predicted 28.6% (i.e., 1.286 times) increase in turnover. Although these results are unexpected, they are not necessarily surprising because both findings have theoretical explanations and empirical precedence within the turnover literature (see Cropanzano, Bowen, & Gilliland, 2007; Gajendran & Harrison, 2007). Possible theoretical explanations for these findings are further explained below.

When comparing Model 1b (i.e., actual turnover rate model) with Model 2b (i.e., turnover intention rate model), we notice several directional differences. The telecommuter variable is positively related to both dependent variables. Although advancement opportunity satisfaction is negatively correlated with agencies’ actual turnover rate (Model 1b), it is positively related to agencies’ turnover intention rate (Model 2b). Furthermore, the performance culture variable is positively correlated with actual turnover rate (Model 1b) but negatively related to turnover intention rate (Model 2b).

With regard to the collective demographic factors, in Model 1b, average tenure’s effect remains most substantial, whereas the influence of non-White minority representation is the weakest and remains statistically insignificant. In Model 2b, non-White minority representation is a significant demographic variable (p = .008).

These observed differences in the models suggest that the antecedent organizational perceptions that best explain agencies’ actual quit rates do not necessarily simultaneously explain agencies’ turnover intention rates. This validates the study’s preliminary assumption that turnover intention and actual turnover are distinctly different constructs, at least at the organizational level.

Turnover Intention Rate and Actual Turnover Rate

Last, we add the turnover intention rate variable to the model to evaluate the fully specified unrestricted model of actual turnover rate (i.e., Model 1c). OLS regression results for Model 1c are summarized in Table 3, Part C. The full model explains 59% of the variance in turnover rate.

Interestingly, turnover intention rate is not significantly associated with actual turnover rate once other job and personal characteristics are taken into account. These results are surprising considering the significant relationship between turnover intention rate and actual turnover rate in the bivariate regression model (see table 2). In other words, where actual turnover rate is otherwise explained, we reject Hypothesis 1, which states that agencies’ turnover intention rate is positively associated with the agencies’ actual turnover rate.

Taken together, our results provide strong evidence that turnover intention rate and actual turnover rate are indeed two distinct constructs, explained by two separate sets of determinants: Youthful workforce, proportion of females, telecommuting and agency’s satisfaction with performance culture practices, and advancement opportunities are found to influence actual turnover rate, but not turnover intention rate. The only variables to have a dual influence on both types of turnover rate are average tenure and agencies’ satisfaction with pay, telecommuting, and workload. In addition, we find that turnover intention rate is of little practical usefulness as predicator of agencies’ actual future turnover rate.

Discussion

The intention–actual turnover linkage has been a fixed assumption in most studies of voluntary turnover. This study has sought to evaluate the usefulness of turnover intention rate as a proxy and a predictor of actual turnover rate among U.S. federal government agencies. Several results observed here deserve further discussion.

First and most fundamentally, we found that turnover intentions have a direct effect on actual turnover, but so do other perceptual measures. As a perceptual variable, the turnover intention construct seems to be a less reliable predictor than it is initially assumed to be at the organization level. Turnover research posits that the linkage between turnover intentions to actual turnover should be a strong one. Our analysis illustrates that at the organizational level, this is not always the case. Within our sample, intentions explain less than 5% of turnover variance across the federal government. Although unanticipated, this result is not without precedent. Individual-level research reported intention–actual turnover correlations from .31 to .52 (Dalton et al., 1999). At the organizational level, Jung (2010) found no statistically significant correlation between agency-level turnover and sampled government employees’ weighted leave intentions.

It is true that at the aggregate level correlations are more likely to have a diminished statistical power due to linearity assumptions. However, these limitations alone do not explain why other same-source aggregate perceptual measures are more powerfully associated with actual turnover.

At a minimum, our initial finding thus serves as evidence that the level at which a phenomenon is observed matters. This observation is one of no small consequence to public management as strategic HR management is necessarily an organizational-level function. After all, public managers are mostly concerned with turnover and retention rates at their agencies rather than the individual reasons for employees’ turnover behavior (Gardner et al., 2007).

Second, if the conceptual construct of “turnover intention” is a reliable proxy of the concept “actual turnover,” then we would expect that the factors explaining the first one would similarly explain the latter. Within our sample, we found 12 organizational determinants explain more than 59% of actual turnover rate variance across the federal government. At the same time, however, these same factors simultaneously explain less than 12% of variance associated with turnover intention rate.

In addition to explaining intentions less thoroughly, the organizational determinants found to best explain actual turnover are not necessarily the ones that best explain turnover intention. Meaning, of the 11 variables that significantly explain actual turnover, only 3 are statistically associated with intentions (i.e., telecommuters, workload satisfaction, and pay satisfaction).

For example, average tenure and performance culture found to be most significantly and substantially associated with agency actual turnover are insignificant predictors of employees’ turnover intentions. Furthermore, non-White minority representation that is a consistently insignificant predictor of actual turnover is equally consistently statistically associated with turnover intentions.

Noteworthy is the relational shift occurring in some of the variables. We find that the proportion of an agency’s employees who telecommute as well as higher satisfaction with agencies’ performance cultures are each positively and significantly associated with agencies’ increased actual turnover. Both of these findings are contrary to those widely reported within the individual-level intentions-based empirical literature (cf. Gajendran & Harrison, 2007; Huselid, 1995; Pitts et al., 2011). However, testable hypotheses to explain these observations are available in the literature. For example, it is possible that the positive relationship between telecommuting and turnover does not necessarily reflect the attitudes of the telecommuters, but rather those of the non-telecommuters in agencies where telecommuting is practiced. In other words, it could be that quit rates are higher in agencies with higher proportions of telecommuters because of the negative impact they have on non-telecommuters employees who work in that agency. This idea that telework might have adverse consequences on non-teleworkers was also introduced by Golden (2007).

In addition, Baruch and Nicholson (1997) argued that telecommuting might have negative organizational consequences due to employees’ social isolation, perceived career stagnation, and family conflict. Likewise, employees might disproportionally withdraw from federal agencies where performance is emphasized due to the stressful nature of their work environments or simply because poor performers are encouraged to “voluntarily” resign.

Last, as a final stage of analysis, we elaborated our model to assess the influence of turnover intention as a predictor of actual turnover in the full model. Our purpose by so doing was to investigate whether turnover intention rate uniquely explains some aspect of actual turnover or whether controlling for intentions’ presence might enhance our model’s overall predictive power. We found it does neither. As an individual predictor in a model where agencies’ actual turnover rates are otherwise independently explained, we found intentions’ individual effect to be weak and statistically insignificant (p = .177).

Conclusion

Predicting employee turnover is an integral part of future organizational labor needs planning in the federal government, state governments, and local governments (Broach & Dollar, 2006; Goodman, French, & Battaglio, 2015; Jung, 2010; Price, 2004). The results of this study suggest that at the organizational level at least, agencies’ actual turnover rate and turnover intention rate are distinct and contrarily explained constructs. From a practical perspective, federal managers should be cognizant to the possibility that turnover intention may be a poor proxy for actual turnover and that its use as such is potentially yielding dubious results.

Our argument that these constructs may not be as closely related as some theories postulate is supported by methodological and conceptual rationales. For instance, actual turnover is a discernable and objectively measureable phenomenon. Employee turnover intentions, on the other hand, are indirectly and subjectively assessed. As an attitudinal construct, leave intentions is a process variable and, as such, is sensitive to intervening influences constantly in flux. Indeed, the intention to quit may vary as a result of a dispute with a supervisor, praise for a job well done, or rumors about future changes such as merging or downsizing (Kirschenbaum & Weisberg, 1990).

In addition, changing circumstances often prevent employees from putting their sincerest intentions into action. Such circumstances include macroeconomic conditions (Selden & Moynihan, 2000), health status (Price, 2004), family issues (Porter & Steers, 1973), or lack of alternative job opportunities (cf. Hom, Caranikas-Walker, Prussia, & Griffeth, 1992; Martin, 1979). It is therefore not surprising that where associations between employees’ stated intentions and quit behaviors have been observed, they have also often been found to diminish over time (Boe, Bobbitt, & Cook, 1997; LeCompte & Dworkin, 1991). Although these reasons by no means constitute a comprehensive list, they do serve to illustrate that the theoretical linkage between turnover intentions and actual turnover is an unsettled scholarship worthy of examination.

This study’s finding that demographic composition of work units portends actual quits is supported and has long standing in the empirical literature. Pfeffer (1985), for instance, demonstrated that differences in age, tenure, gender, and race in the work unit influence communication and networking patterns that, in turn, affect organizational phenomena such as turnover. Thus, we conclude that public managers tasked with retention might be better served concentrating on their agencies’ unique demographic characteristics, rather than on their employees’ self-reported leave intentions. Indeed, our study’s results might be applied to create preliminary profiles useful for this purpose.

Our goal in this study was to ask, “Does turnover intention matter?” We understand that this is a provocative question; one we do not claim to have answered. Nonetheless, we believe the question we ask is an important one and that our results, at a minimum, illustrate that the linkage between leave intentions and actual turnover at the organizational level is a tenuous one.

It should be noted that our analysis of cross-sectional data represents the experience of a particular set of agencies at a particular moment in time. In addition, data limitations prevented us from being able to test all factors discussed in the literature and to control for important constructs such as agency type and size.

Future research could improve this study’s weakness by collecting data on these omitted variables and investigate these testable propositions. Future research should also examine the relationship between individual-level data and organization-level turnover data to explore the differences in the concepts of actual turnover and turnover intention. By incorporating these two levels of data, important management perspectives may be gained, as cumulative differences might combine to point the way to a useful empirically based organizational-level theory of turnover.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Notes

Author Biographies

Galia Cohen is a senior lecturer and an associate director in the Justice Administration and Leadership program at the University of Texas at Dallas. She received her BA in psychology and her MA in organizational sociology from Bar-Ilan University, Israel. She holds a PhD in public affairs from the University of Texas at Dallas. Her scholarly interests include organizational behavior, human resource management, and leadership.

Robert S. Blake has a PhD in public affairs program from the University of Texas at Dallas. His research interests include public human resource management and effective administration of health care policies designed to promote institutional accountability and improved quality and patient safety. He is the director of strategic planning at the Medical Center of Plano.

Doug Goodman is a professor of public affairs and the Master of Public Affairs director in the School of Economic, Political, and Policy Sciences at The University of Texas at Dallas. His research and teaching interests include public human resource management, workforce planning, public management, and other topics in public administration. He has published in numerous public administration outlets.

References

Anderson

R. A.

Corazzini

K. N.

McDaniel

R. R.

(2004). Complexity science and the dynamics of climate and communication: Reducing nursing home turnover. The Gerontologist, 44, 378-388.

Bae

K. B.

Goodman

(2014). The influence of family-friendly policies on turnover and performance in South Korea. Public Personnel Management, 43, 520-542.

Banaszak-Holl

Hines

M. A.

(1996). Factors associated with nursing home staff turnover. The Gerontologist, 36, 512-517.

Baruch

Nicholson

(1997). Home, sweet work: Requirements for effective home working. Journal of General Management, 23, 15-30.

Belanger

Collins

R. W.

(1998). Distributed work arrangements: A research framework. The Information Society, 14, 137-152.

Bennett

Blum

T. C.

Long

R. G.

Roman

P. M.

(1993). A firm-level analysis of employee attrition. Group & Organization Management, 18, 482-499.

Bertelli

A. M.

(2007). Determinants of bureaucratic turnover intention: Evidence from the Department of the Treasury. Journal of Public Administration Research and Theory, 17, 235-258.

Blau

F. D.

Kahn

L. M.

(1981). Race and sex differences in quits by young workers. Industrial & Labor Relations Review, 34, 563-577.

Boe

E. E.

Bobbitt

S. A.

Cook

L. H.

(1997). Whither didst thou go? Retention, reassignment, migration, and attrition of special and general education teachers from a national perspective. Journal of Special Education, 30, 371-389.

10.

Bright

(2008). Does public service motivation really make a difference on the job satisfaction and turnover intentions of public employees? The American Review of Public Administration, 38, 149-166.

11.

Broach

Dollar

(2006). Comparison of intent-to-leave with actual turnover within the FAA. Oklahoma City, OK: Federal Aviation Administration Civil Aerospace Medical Institution.

12.

Carsten

J. M.

Spector

P. E.

(1987). Unemployment, job satisfaction, and employee turnover: A meta-analytic test of the Muchinsky model. Journal of Applied Psychology, 72, 374-381.

13.

Cayer

N. J.

(2003). Public employee benefits and the changing nature of the workforce. In Hays

S. W.

Kearney

R. C.

(Eds.), Public personnel administration: Problems and prospects (pp. 167-179). Upper Saddle River, NJ: Prentice Hall.

14.

Cho

Y. J.

Lewis

G. B.

(2012). Turnover intention and turnover behavior implications for retaining federal employees. Review of Public Personnel Administration, 32, 4-23.

15.

Choi

(2009). Diversity in the US federal government: Diversity management and employee turnover in federal agencies. Journal of Public Administration Research and Theory, 19, 603-630.

16.

Cotton

J. L.

Tuttle

J. M.

(1986). Employee turnover: A meta-analysis and review with implications for research. Academy of Management Review, 11, 55-70.

17.

Cropanzano

Bowen

D. E.

Gilliland

S. W.

(2007). The management of organizational justice. The Academy of Management Perspectives, 21(4), 34-48.

18.

Currivan

D. B.

(1999). The causal order of job satisfaction and organizational commitment in models of employee turnover. Human Resource Management Review, 9, 495-524.

19.

Dalton

D. R.

Johnson

J. L.

Daily

C. M.

(1999). On the use of “intent to.” variables in organizational research: An empirical and cautionary assessment. Human Relations, 52, 1337-1350.

20.

Farber

H. S.

(1999). Mobility and stability: The dynamics of job change in labor markets. In Ashenfelter

Card

(Eds.), Handbook of labor economics (Vol. 3, pp. 2439-2483). Amsterdam, The Netherlands: Elsevier Science B.V.

21.

Farnham

Giles

(1996). People management and employment relations. In Farnham

Horton

(Eds.), Managing the new public services (pp. 112-136). London, England: Macmillan.

22.

Fernandez

Moldogaziev

(2010). Empowering public sector employees to improve performance: Does it work? The American Review of Public Administration, 41, 23-47.

23.

Fishbein

Ajzen

(1975). Belief, attitude, intention and behavior: An introduction to theory and research. Reading, MA: Addison–Wesley.

24.

Gajendran

R. S.

Harrison

D. A.

(2007). The good, the bad, and the unknown about telecommuting: Meta-analysis of psychological mediators and individual consequences. Journal of Applied Psychology, 92, 1524-1541.

25.

Gardner

T. M.

Moynihan

L. M.

Wright

P. M.

(2007). The influence of human resource practices and collective affective organizational commitment on aggregate voluntary turnover. Retrieved from http://digitalcommons.ilr.cornell.edu/cahrswp/479

26.

Glebbeek

A. C.

Bax

E. H.

(2004). Is high employee turnover really harmful? An empirical test using company records. Academy of Management Journal, 47, 277-286.

27.

Golden

(2007). Co-workers who telework and the impact on those in the office: Understanding the implications of virtual work for co-worker satisfaction and turnover intentions. Human Relations, 60, 1641-1667.

28.

Goodman

French

P. E.

Battaglio

R. P.

(2015). Determinants of local government workforce planning. The American Review of Public Administration, 45, 135-152.

29.

Gould-Williams

(2004). The effects of “high commitment” HRM practices on employee attitude: The views of public sector workers. Public Administration, 82, 63-81.

30.

Greenhaus

J. H.

Parasuraman

Wormley

W. M.

(1990). Effects of race on organizational experiences, job performance evaluations, and career outcomes. Academy of Management Journal, 33(1), 64-86.

31.

Griffeth

R. W.

Hom

P. W.

Gaertner

(2000). A meta-analysis of antecedents and correlates of employee turnover: Update, moderator tests, and research implications for the next millennium. Journal of Management, 26, 463-488.

32.

Gujarati

D. N.

(2003). Basic econometrics (4th ed). Boston, MA: McGraw-Hill.

33.

Hackman

J. R.

Oldham

G. R.

(1976). Motivation through the design of work: Test of a theory. Organizational Behavior and Human Performance, 16, 250-279.

34.

Harrison

D. A.

Newman

D. A.

Roth

P. L.

(2006). How important are job attitudes? Meta-analytic comparisons of integrative behavioral outcomes and time sequences. Academy of Management Journal, 49, 305-325.

35.

Hausknecht

J. P.

Trevor

C. O.

(2011). Collective turnover at the group, unit, and organizational levels: Evidence, issues, and implications. Journal of Management, 37, 352-388.

36.

Hausknecht

J. P.

Trevor

C. O.

Howard

M. J.

(2009). Unit-level voluntary turnover rates and customer service quality: Implications of group cohesiveness, newcomer concentration, and size. Journal of Applied Psychology, 94, 1068-1075.

37.

Hays

S. W.

Kearney

R. C.

(2001). Anticipated changes in human resource management: Views from the field. Public Administration Review, 61, 585-597.

38.

Hom

P. W.

Caranikas-Walker

Prussia

G. E.

Griffeth

R. W.

(1992). A meta-analytical structural equations analysis of a model of employee turnover. Journal of Applied Psychology, 77, 890-909.

39.

Hom

P. W.

Griffeth

R. W.

Sellaro

C. L.

(1984). The validity of Mobley’s (1977) model of employee turnover. Organizational Behavior and Human Performance, 34, 141-174.

40.

Huselid

M. A.

(1995). The impact of human resource management practices on turnover, productivity, and corporate financial performance. Academy of Management Journal, 38, 635-672.

41.

Ingraham

P. W.

Rubaii-Barrett

(2007). Human resource management as a core dimension of public administration. Foundations of Public Administration Series. Retrieved from http://www.aspanet.org

42.

Ippolito

R. A.

(1987). Why federal workers don’t quit. Journal of Human Resources, 22, 281-299.

43.

Iscan

O. F.

Naktiyok

(2005). Attitudes towards telecommuting: The Turkish case. Journal of Information Technology, 20, 52-63.

44.

Johnston

M. W.

Griffeth

R. W.

Burton

Carson

P. P.

(1993). An exploratory investigation into the relationships between promotion and turnover: A quasi-experimental longitudinal study. Journal of Management, 19, 33-49.

45.

Jung

C. S.

(2010). Predicting organizational actual turnover rates in the US federal government. International Public Management Journal, 13, 297-317.

46.

Kellough

J. E.

Osuna

(1995). Cross-agency comparisons of quit rates in the federal service another look at the evidence. Review of Public Personnel Administration, 15(4), 58-68.

47.

Kettl

D. F.

(2005). Reforming the executive branch of the US Government. In Aberbach

J. D.

Peterson

M. A.

(Eds.), The executive branch (pp. 344-378). Oxford, UK: Oxford University Press.

48.

Kim

(2005). Factors affecting state government information technology employee turnover intentions. The American Review of Public Administration, 35, 137-156.

49.

Kirschenbaum

Weisberg

(1990). Predicting worker turnover: An assessment of intent on actual separations. Human Relations, 43, 829-847.

50.

Kraut

A. I.

(1975). Predicting turnover of employees from measured job attitudes. Organizational Behavior and Human Performance, 13, 233-243.

51.

Lambert

E. G.

Hogan

N. L.

Barton

S. M.

(2001). The impact of job satisfaction on turnover intent: A test of a structural measurement model using a national sample of workers. The Social Science Journal, 38, 233-250.

52.

LeCompte

M. D.

Dworkin

A. G.

(1991). Giving up on school: Student dropouts and teacher burnouts. Newbury Park, CA: Corwin Press.

53.

Lee

Jimenez

B. S.

(2011). Does performance management affect job turnover intention in the Federal Government? The American Review of Public Administration, 41, 168-184.

54.

Lee

S. Y.

Hong

J. H.

(2011). Does family-friendly policy matter? Testing its impact on turnover and performance. Public Administration Review, 71, 870-879.

55.

Lee

S. Y.

Whitford

A. B.

(2007). Exit, voice, loyalty, and pay: Evidence from the public workforce. Journal of Public Administration Research and Theory, 18, 647-671.

56.

Lee

T. W.

Mowday

R. T.

(1987). Voluntarily leaving an organization: An empirical investigation of Steers and Mowday’s model of turnover. Academy of Management Journal, 30, 721-743.

57.

Lewis

G. B.

(1991). Turnover and the quiet crisis in the federal civil service. Public Administration Review, 51, 145-155.

58.

Lewis

G. B.

Park

(1989). Turnover rates in federal white-collar employment: Are women more likely to quit than men? The American Review of Public Administration, 19, 13-28.

59.

Lewis

(2007). Stepwise versus hierarchical regression: Pros and cons. Retrieved from http://www.academia.edu/1860655/Stepwise_versus_hierarchical_regression_Pros_and_cons

60.

Martin

T. N.

(1979). A contextual model of employee turnover intentions. Academy of Management Journal, 22, 313-324.

61.

Meier

K. J.

Hicklin

(2008). Employee turnover and organizational performance: Testing a hypothesis from classical public administration. Journal of Public Administration Research and Theory, 18, 573-590.

62.

Mobley

W. H.

(1977). Intermediate linkages in the relationship between job satisfaction and employee turnover. Journal of Applied Psychology, 62, 237-240.

63.

Mobley

W. H.

Griffeth

R. W.

Hand

H. H.

Meglino

B. M.

(1979). Review and conceptual analysis of the employee turnover process. Psychological Bulletin, 86, 493-522.

64.

Mobley

W. H.

Horner

S. O.

Hollingsworth

A. T.

(1978). An evaluation of precursors of hospital employee turnover. Journal of Applied Psychology, 63, 408-414.

65.

Moynihan

D. P.

Landuyt

(2008). Explaining turnover intention in state government: Examining the roles of gender, life cycle, and loyalty. Review of Public Personnel Administration, 28, 120-143.

66.

Office of Personnel Management. (2005). Federal civilian workforce statistics: Employment and trends as of January 2005. Retrieved from http://www.opm.gov/feddata

67.

Office of Personnel Management. (2011). Federal employee viewpoint survey. Empowering–inspiring change. Retrieved from www.fedscope.opm.gov

68.

O’Reilly

C. A.

Chatman

Caldwell

D. F.

(1991). People and organizational culture: A profile comparison approach to assessing person-organization fit. Academy of Management Journal, 34, 487-516.

69.

Park

H. Y.

Ofori-Dankwa

Bishop

D. R.

(1994). Organizational and environmental determinants of functional and dysfunctional turnover: Practical and research implications. Human Relations, 47, 353-366.

70.

Perry

J. L.

Hondeghem

Wise

L. R.

(2010). Revisiting the motivational bases of public service: Twenty years of research and an agenda for the future. Public Administration Review, 70, 681-690.

71.

Pfeffer

(1985). Organizational demography: Implications for management. California Management Review, 28, 67-81.

72.

Pitts

Marvel

Fernandez

(2011). So hard to say goodbye? Turnover intention among US federal employees. Public Administration Review, 71, 751-760.

73.

Porter

L. W.

Steers

R. M.

(1973). Organizational, work, and personal factors in employee turnover and absenteeism. Psychological Bulletin, 80, 151-176.

74.

Price

J. L.

(2004). The development of a causal model of voluntary turnover. In Griffeth

R. W.

Hom

P. W.

(Eds.), Innovative theory and empirical research on employee (pp. 3-32). Greenwich, CT: Information Age Publishing.

75.

Price

J. L.

Mueller

C. W.

(1981). A causal model of turnover for nurses. Academy of Management Journal, 24, 543-565.

76.

Sackett

P. R.

DuBois

C. L.

Noe

A. W.

(1991). Tokenism in performance evaluation: The effects of work group representation on male–female and white–black differences in performance ratings. Journal of Applied Psychology, 76, 263-267.

77.

Selden

S. C.

Moynihan

D. P.

(2000). A model of voluntary turnover in state government. Review of Public Personnel Administration, 20(2), 63-74.

78.

Shaw

J. D.

Delery

J. E.

Jenkins

G. D.

Gupta

(1998). An organization-level analysis of voluntary and involuntary turnover. Academy of Management Journal, 41, 511-525.

79.

Spector

P. E.

(1985). Measurement of human service staff satisfaction: Development of the job satisfaction survey. American Journal of Community Psychology, 13, 693-713.

80.

Spell

C. S.

Blum

T. C.

(2005). Adoption of workplace substance abuse prevention programs: Strategic choice and institutional perspectives. Academy of Management Journal, 48, 1125-1142.

81.

Steel

R. P.

Ovalle

N. K.

(1984). A review and meta-analysis of research on the relationship between behavioral intentions and employee turnover. Journal of Applied Psychology, 69, 673-686.

82.

Terborg

J. R.

Lee

T. W.

(1984). A predictive study of organizational turnover rates. Academy of Management Journal, 27, 793-810.

83.

Tett

R. P.

Meyer

J. P.

(1993). Job satisfaction, organizational commitment, turnover intention, and turnover: Path analyses based on meta-analytic findings. Personnel Psychology, 46, 259-293.

84.

Trevor

C. O.

Nyberg

A. J.

(2008). Keeping your headcount when all about you are losing theirs: Downsizing, voluntary turnover rates, and the moderating role of HR practices. Academy of Management Journal, 51, 259-276.

85.

Tsui

A. S.

Egan

T. D.

O’Reilly

C. A.

III (1992). Being different: Relational demography and organizational attachment. Administrative Science Quarterly, 37, 549-579.

86.

U.S. Congress House. (1999). Department of Transportation and Related Agencies Appropriations Act of 2000. Public Law 106–069 106th Cong., 2nd sess. Oct 9.

87.

Vandenberg

R. J.

Nelson

J. B.

(1999). Disaggregating the motives underlying turnover intentions: When do intentions predict turnover behavior? Human Relations, 52, 1313-1336.

88.

Van Marrewijk

Timmers

(2003). Human capital management: New possibilities in people management. Journal of Business Ethics, 44, 171-184.

89.

Wiersema

M. F.

Bird

(1993). Organizational demography in Japanese firms: Group heterogeneity, individual dissimilarity, and top management team turnover. Academy of Management Journal, 36, 996-1025.

90.

Wright

P. M.

Boswell

W. R.

(2002). Desegregating HRM: A review and synthesis of micro and macro human resource management research. Journal of Management, 28, 247-276.

91.

Yanadori

Kato

(2007). Average employee tenure, voluntary turnover ratio, and labour productivity: Evidence from Japanese firms. The International Journal of Human Resource Management, 18, 1841-1857.