Modeling the determinants of turnover intentions: a Bayesian approach

Published date03 April 2018
DOIhttps://doi.org/10.1108/EBHRM-10-2016-0025
Date03 April 2018
Pages2-24
AuthorAnup Menon Nandialath,Emily David,Diya Das,Ramesh Mohan
Subject MatterHR & organizational behaviour,Global HRM
Modeling the determinants
of turnover intentions:
a Bayesian approach
Anup Menon Nandialath
Department of Management, University of Wisconsin La Crosse,
La Crosse, Wisconsin, USA
Emily David
China Europe International Business School (CEIBS), Shanghai, China
Diya Das
Department of Management, Bryant University, Smithfield, Rhode Island,
USA, and
Ramesh Mohan
Department of Economics, Bryant University, Smithfield,
Rhode Island, USA
Abstract
Purpose Much of what we learn from emp irical research is base d on a specific empirical mo del(s)
presented in the liter ature. However, the range of pl ausible models given the data i s potentially larger, thus
creating an additiona l source of uncertainty t ermed: model uncertaint y. The purpose of this pap er is to
examine the effect of mode l uncertainty on empiri cal research in HRM and sug gest potential solutio ns
to deal with the same.
Design/methodology/approach Using a sample of call center employ ees from India, the authors test
the robustness of predi ctors of intention to leave based on the unfo lding model proposed by Harman et.al.
(2007). Methodologically, the authors use Bayesian Model Averaging (BMA) to identify the specific
variables within the un folding model that have a robust relationship with turnover intent ions after
accounting for model un certainty.
Findings The findings show that indeed model uncertainty can impact what we learn from empirical
studies. More specifically, in the context of the sample, using four plausible model specifications, the authors
show that the conclusions can vary depending on which model the authors choose to interpret. Furthermore,
using BMA, the authors find that only two variables, job satisfaction and perceived organizational support,
are model specification independent robust predictors of intention to leave.
Practical implications The research has specific implications for the development of HR analytics and
informs managers on which are the most robust elements affecting attrition.
Originality/value While empirical research typically acknowledges and corrects for the presence of
sampling uncertainty through p-values, rarely does it acknowledge the presence of model uncertainty
(which variables to include in a model). To the best of the authorsknowledge, it is the first study to show the
effect and offer a solution to studying total uncertainty (sampling uncertainty +model uncertainty)
on empirical research in HRM. The work should open more doors toward more studies evaluating the
robustness of key HRM constructs in explaining important work-related outcomes.
Keywords Employee turnover, Research methodology, Human resource management, Model averaging
Paper type Research paper
1. Introduction
A perennial probl em facing organizati ons is the retention of key talent. Volun tary
employee turnover is costly for organizations, with financial estimates ranging from one
years salary for each departing employee (Boroşand Curşeu, 2013) to millions of dollars
lost to training, recruitment, and lost productivity (Sagie et al.,2002).Perhapsmore
worrisomearetheintangiblecoststoemployee morale for remaining workers. These
employees are often expected to take on the extra work until a replacement is found and
Evidence-based HRM: a Global
Forum for Empirical Scholarship
Vol. 6 No. 1, 2018
pp. 2-24
© Emerald PublishingLimited
2049-3983
DOI 10.1108/EBHRM-10-2016-0025
Received 20 October 2016
Revised 15 March 2017
Accepted 22 March 2017
The current issue and full text archive of this journal is available on Emerald Insight at:
www.emeraldinsight.com/2049-3983.htm
2
EBHRM
6,1
may also experience feelings of being left behind in the caustic environment that drove
their former coworkers to leave. Given that turnover can have contagion effects
(Felps et al., 2009), high turnover rates may lead to even further loss of talent as employees
follow one another to better opportunities.
Turnover intentions (also referred to as employee intent to leave or quit) reflect a
workers desire to leave the organization, sentiments that typically precede the actual act of
quitting (Griffeth et al., 2000; Zimmerman and Darnold, 2009). Indeed, research has found
that this mental intention is the single biggest predictor of actual employee turnover
(Van Breukelen et al., 2004). Contemplating leaving ones job is often a complex process that
involves a myriad of considerations including ones attitudes, finances, family
circumstances, and job alternatives (Griffeth et al., 2000; Harman et al., 2007; Woo and
Allen, 2014). According to the unfolding model of voluntary turnover (Lee and Mitchell,
1994), employees face a number of decision points before ultimately making the choice to
leave. Sometimes quit intentions arise rather quickly following a shock event (e.g. a parent
becoming terminally ill or a company merger causing unpleasant policy changes). Other
times they form gradually over time due to disagreeable situations and mounting
dissatisfaction (Harman et al., 2007).
Several studies have focused on the latter situation and identified variables that affect an
employeesintention to leave over time such as labor market conditions (Mowday et al.,
1979), role attributes (Eagly and Chaiken, 1993), relational variables (Mossholder et al.,
2005), and employee attitudes (Van Breukelen et al., 2004). What is intriguing is that many of
these studies reported variable or mixed results. As an example, we highlight the existing
debate about the relationship between job satisfaction and intention to leave. Several
scholars contend that this relationship is often unclear due to moderating variables such as
personality, perception of equity, and employee well-being (Wright and Bonnet , 2007;
Judge, 1993; Berg, 1991). Similarly, there are debates about the relationship between
leadership and intention to leave such as whether the quality of leader-member exchange
(LMX) has a direct or indirect impact on an employees turnover intentions (Gerstner and
Day, 1997; DeConinck, 2009; Morrow et al., 2005). As a result of these mixed findings, there is
no way to be sure which of these predictors has the most consistent and powerful effects on
deterring employee intentions to quit.
Recent literature in management suggests that one potential reason for the lack of
consensus is model uncertainty (Arin et al., 2015). Model uncertainty arises from the choices
a researcher makes about which variables should be included in an empirical analysis. Such
choices can have a significant impact on observed empirical results (see Simmons et al.,
2011). Prior literature suggests that Bayesian Model Averaging (BMA) may serve as an
appropriate method for guarding against model uncertainty (Raftery, 1995; Fernandez et al.,
2001). More recently, there has been an increased call for using Bayesian inference in the
organizational sciences (Kruschke et al., 2012). While Bayesian inference can help answer
several interesting questions relevant to the managerial problems, a common critique
focuses on the role of prior distributions and how they influence outcomes. Within the
context of BMA, prior literature suggests several different specifications for prior
distributions such that they do not unduly influence the posterior distribution of effects
which is of focal interest (Ley and Steel, 2009). To further demonstrate that prior
distributions are not influencing results, we follow a strategy where results from using all
prior distributions are reported and compared to check if inference changes depending on
prior choice.
With the present study, we aimed to make several contributions to the literature.
Development of good theory requires sound empirical research to demonstrate the efficacy
of theory in explaining observed outcomes. Model uncertainty can be a serious impediment
to drawing durable conclusions. We apply BMA in the domain of attrition research,
3
Turnover
intentions

To continue reading

Request your trial

VLEX uses login cookies to provide you with a better browsing experience. If you click on 'Accept' or continue browsing this site we consider that you accept our cookie policy. ACCEPT