Synthesizing content models of employee turnover

• Date: 01 January 2019
• Author: Brian W. Swider,Ryan D. Zimmerman,Wendy R. Boswell
• Published date: 01 January 2019
• DOI: http://doi.org/10.1002/hrm.21938

HR SCIENCE FORUM

Synthesizing content models of employee turnover

Ryan D. Zimmerman

1

| Brian W. Swider

2

| Wendy R. Boswell

3

1

Department of Management, Virginia Tech

University, Blacksburg, Virginia

2

Department of Management, University of

Florida, Gainesville, Florida

3

Department of Management, Texas A&M

University, College Station, Texas

Correspondence

Ryan D. Zimmerman, Department of

Management, Pamplin College of Business,

Virginia Tech University, Blacksburg, VA

24061.

Email: rdzimmer@vt.edu

To identify a set of broad factors that reflect the constructs measured in three content models

of employee turnover, we hypothesized 19 scales would reduce to five factors related to

employee job-search behavior and actual turnover decisions: one's affect toward the organiza-

tion, work environment, instrumental attachment, extraorganizational ties, and sense of obliga-

tion. Using a sample of 888 staff members from a large university, the factor structure was

confirmed. Multivariate regression results also indicated one's work environment, instrumental

attachment, and sense of obligation were significantly and negatively related to both job search

and turnover, with work environment and instrumental attachment exhibiting the strongest

effects. Extraorganizational ties were only significantly and negatively related to job search.

Interestingly, affect toward the organization was not significantly related to either job search or

turnover. Path analyses indicated the effects of four of the factors on turnover were fully medi-

ated by job search, with instrumental attachment the sole factor that was only partially medi-

ated. Our model provides a foundation for future researchers to test the uniqueness of new

predictors of turnover, as well as guidance to practitioners regarding where resources might be

best utilized in curbing turnover.

KEYWORDS

embeddedness, job satisfaction, job search, organizational commitment, turnover

As an outcome of importance to all organizations and a complex pro-

cess for individuals, employee turnover has elicited significant

research endeavors as well as a multitude of theoretical models. His-

torically, March and Simon's (1958) model of turnover has received

the most research attention, with job satisfaction and affective organi-

zational commitment as job attitudes most often representing one's

desire to leave and job alternatives representing one's ease of leaving

(Mitchell, Holtom, Lee, Sablynski, & Erez, 2001). While this seminal

work has received widespread empirical support, more recent

research has focused on identifying other factors affecting employee

turnover. These newer models include Mitchell and Lee's (Lee, Mitch-

ell, Sablynski, Burton, & Holtom, 2004; Mitchell et al., 2001) theory of

job embeddedness as well as Maertz et al.’s (Maertz & Campion,

2004; Maertz & Griffeth, 2004) model outlining motivational forces

that influence employee-turnover decisions. Although these authors

frequently (and rightfully so) highlight the theoretical advancements in

their models, far less discussion has been given to commonalities with

extant models. That is, despite the new perspectives offered by these

more recent models, there is likely to be overlap with both the tradi-

tional predictors of turnover as well as other recent models.

Therefore, it is critical to take a holistic view of existing employee-

turnover models to better understand the core of why employees

leave organizations.

As noted by Hambrick (2007) and Leavitt, Mitchell, and Peterson

(2010), the academic community tends to prioritize the creation of

new theories over the refinement of current ones. Yet, because of this

practice, oftentimes different constructs in different models are, in

reality, the same. This redundancy can manifest in two ways: either

theoretical redundancy, when the constructs are conceptually the

same, and/or empirical redundancy, when the constructs cannot be

distinguished from each other based on the magnitude of their rela-

tionships (Schwab, 1980). At best, this construct overlap leads

researchers to spend time reinventing (and retesting) the wheel, and

at worse it provides inconsistent empirical results due to multicollin-

earity issues when including empirically redundant constructs in the

same analyses.

Therefore, the purpose of this article is to advance a more concise

turnover model by identifying a parsimonious group of factors from

prevalent theoretical models of turnover, confirm their empirical

uniqueness (or redundancy) to one another, and evaluate their relative

DOI: 10.1002/hrm.21938

Hum Resour Manage. 2019;58:99–114. wileyonlinelibrary.com/journal/hrm © 2018 Wiley Periodicals, Inc. 99

ability to predict employee job search and turnover. We note while

technically dozens of turnover-related variables could be included in

the models that we test, we elected to include those of recent and

prevalent turnover models as well as scales representing the well-

established constructs of job satisfaction and affective organizational

commitment (Hom, Lee, Shaw, & Hausknecht, 2017). Our findings will

contribute to the turnover literature by establishing a parsimonious

content model of employee turnover that succinctly captures the

important, unique aspects of multiple existing models. Such a model

will also provide guidance to practitioners by identifying the core rea-

sons for employee job search and turnover, thereby allowing organiza-

tions to focus employee retention initiatives on factors that are most

effective in promoting employee retention.

1|MODELS OF EMPLOYEE TURNOVER

Models of employee turnover tend to fall into one of the two catego-

ries: process models or content models (Maertz & Campion, 2004).

Process models focus on the sequence of steps employees go through

during the process of quitting, such as developing feelings of dissatis-

faction, thinking about quitting, searching for alternative employment,

and then quitting their current jobs. Content models, on the other

hand, focus on factors that cause employees to quit, incorporating

constructs such as attributes of the job, organization, and individual as

well as alternative opportunities. Finally, some models include both

the process and content (Hom, Mitchell, Lee, & Griffeth, 2012).

While there are some differences among process models, a

review of the turnover literature indicates content models tend to uti-

lize far more diverse sets of variables when predicting employee job

search and turnover. Although other unified models of turnover have

been created, their commonality is that they tend to lack parsimony. It

is not uncommon for content models to present anywhere from 8

(e.g., Maertz & Griffeth, 2004), to about 16 (Bluedorn, 1982; Price &

Mueller, 1986), to over 24 (e.g., Mobley, Griffeth, Hand, & Meglino,

1979) distinct factors that influence employees' turnover decisions.

Griffeth, Hom, and Gaertner's (2000) meta-analysis included over

30 effect size estimates for the relationships between content vari-

ables and turnover, with this meta-analysis predating the development

of turnover predictors developed in the more recent content models

of turnover (e.g., Maertz & Campion, 2004; Mitchell et al., 2001). As

there appears to be an overabundance of content models in need of

“pruning,”we have chosen to focus on content models of turnover.

1.1 |Early content model of turnover

March and Simon's (1958) turnover model was quite parsimonious,

focusing on two general reasons why employees left organizations:

the perceived desirability of movement and perceived ease of move-

ment. As noted by previous authors (e.g., Hom & Griffeth, 1995;

Mitchell et al., 2001), in the 60 years since March and Simon's model

was published, the most common ways of measuring employees' per-

ceived desirability of movement is through their satisfaction with vari-

ous aspects of their jobs (typically including pay, promotion

opportunities, coworkers, supervisors, and the work itself ) or their

affective organizational commitment; meta-analytic results indicate

both job satisfaction and affective organizational commitment do pre-

dict turnover across all jobs (Griffeth et al., 2000). One's perceived

ease of movement has historically been measured by asking

employees about their perceptions of alternative job opportunities

(Hom & Griffeth, 1995). Again, published studies provided support for

March and Simon's (1958) proposition, as perceived job alternatives

were generally found to positively predict turnover.

1.2 |More recent content models of turnover

One of the most popular recent content models of turnover, intro-

duced by Mitchell et al. (2001), is the theory of job embeddedness.

Job embeddedness represents an employee's investments made on-

and off-the-job. Specifically, the employee's links to others on- and

off-the-job, fit in the organization and community, and potential sacri-

fices if the employee were to leave the job or community were all pro-

posed and have been shown to predict employee turnover (Jiang, Liu,

McKay, Lee, & Mitchell, 2012). Although aspects of job embedded-

ness do reflect attitudes toward the job and organization, as well as

consideration of alternative employment opportunities, it also intro-

duces two important elements, which augment March and Simon's

model of turnover.

One significant theoretical contribution by Mitchell et al. (2001) is

the emphasis on nonaffective factors that may prevent people from

leaving their position of employment and ultimately help keep them

on the job. The basic argument (and finding) is that employees who

are more embedded (i.e., have a more intricate or complex web in

which they can become stuck) are not as likely to leave an organiza-

tion as those employees who are less embedded. The second key con-

tribution of job embeddedness theory is the idea that off-the-job

factors can have an influence on employees' decisions to leave

(or stay at) their organizations (Hom et al., 2017). That is, it is not just

what happens in the work environment that affects employees' turn-

over decisions, but also what is going on in the community in which

the employees live. These off-the-job factors can include relationship

dynamics with family, friends, and acquaintances in employees'

communities.

Another modern model of turnover introduced around the same

time as Mitchell et al.’s is Maertz et al.’s (Maertz & Campion, 2004;

Maertz & Griffeth, 2004) model of motivational forces. Maertz

et al. created this theoretical model to conceptualize and distinguish

the various factors that may influence employees' turnover decisions.

Several of the motivation forces are similar in content to March and

Simon's (1958) model, including number and equivalence of extraorga-

nizational opportunities perceived (termed alternative forces and

behavioral forces, respectively), attitudes toward the organization

(affective forces), as well as job-related factors such as coworkers (con-

stituent forces) and promotion/development opportunities (calculative

forces). However, as was the case with job embeddedness theory,

Maertz et al. developed a new theory in the form of three additional

forces, which subsequently extended March and Simon's (1958) semi-

nal model.

Similar to the aspects of Mitchell et al.’s (2001) theory of job

embeddedness, Maertz et al. consider the effects of external

100 ZIMMERMAN ET AL.