Вы находитесь на странице: 1из 21

Abstract

An Experimental
Although much research has focused on factors
Investigation of driving the turnover behaviors of information
technology (IT) professionals once they are in an

Turnover Intentions employment relationship, little is known about their ex


ante intentions to stay when embarking on a new
employment relationship. The present study
Among New investigates the effects of individual and situational
factors on the turnover intentions of new entrants into
Entrants in IT the IT workforce. Arguably, the IT revolution of the
past decade, both at its peak as well as its decline,
has changed the expectations and values of this
population of workers considerably. Adopting an
Ritu Agarwal interactionist perspective, we examine the interaction
between individual preferences for organizational risk
University of Maryland and variety, and the level of entrepreneurial risk
inherent in the business model of a given employer
Thomas W. Ferratt together with the amount of variety that IT work with
University of Dayton that employer offers. The posited relationships are
tested in a policy capturing experiment with
graduating information systems majors as subjects.
Prabuddha De The results provide support for the interactionist
Purdue University perspective, while at the same time questioning its
basic premise.

ACM Categories: K.6, K.6.1, K.7, K.7.1, K.7.2

Keywords: Turnover Intentions, IT Professionals,


Situational Risk, Situational Variety, Experimental
Approach, New Entrants

Introduction
Scholars have argued that IT human capital
represents a strategic resource for firms, and has the
ability to bestow competitive advantage (Wade &
Hulland, 2004; Bharadwaj, 2000). Thus, the turnover
and retention of IT professionals persists as an
important managerial concern (Luftman, 2005;
Luftman & McLean, 2004; Hsu et al., 2003). To this
end, IS and organizational behavior researchers have
examined a range of individual level factors as
explanatory variables for intention to stay and
turnover among IS and other populations of
employees, including career anchors (Hsu et al.,
2003), perceived ease of movement, job
dissatisfaction, and organizational commitment (e.g.,
Igbaria & Greenhaus, 1992; March & Simon, 1958;
Mobley et al., 1979; Hom & Griffeth, 1995).
Researchers have also suggested that deliberate
managerial actions, including IT organizations’ human
resource management (HRM) practices, such as
compensation and training (Agarwal & Ferratt, 1999;
2001; 2002; Slaughter & Ang, 2002), and internal
work characteristics such as job design (Thatcher et
al., 2002), are important determinants of intention to

The DATA BASE for Advances in Information Systems 8 Volume 38, Number 1, February 2007
stay and turnover. Finally, IS scholars have examined nature and context of the IT profession and the
the effects of discrepancies between individual needs upheaval it has witnessed, in this paper we posit that
and what the organization supplies as determinants of risk and variety are two important explanatory
turnover intention (Jiang & Klein, 2002). variables for the length of time an IT professional
would choose to stay with a new employer. Arguably,
In spite of the significant body of research focused on
the IT revolution of the past decade, both at its peak
understanding turnover among IT professionals,
as well as its decline, has changed the expectations
however, there is a notable gap in this literature.
and values of this population of workers considerably.
Extant research examines the factors that influence IT
During the mid 90s, Y2K compliance concerns
professionals’ staying or leaving behavior after they
created a strong demand for IT workers with specific
are already in the employment relationship. This
skills that declined after the year 2000. The years
research, thus, provides limited insight into IT
1996-1999, riding upon the Internet revolution, were
professionals’ ex ante intentions to stay when they
characterized by a severe shortage of IT workers with
are entering a new employment relationship. High
skills in the new technologies and a supply-demand
costs associated with recruitment and socialization
asymmetry that was unprecedented. This resulted in
suggest that it is important for organizations to retain
a “careerist” mindset among IT workers (Agarwal &
new hires for a reasonable length of time so as to
Ferratt, 1999; Feldman & Weitz, 1991), where
recover their investment (Griffeth et al., 2000). Thus, it
opportunities for financial gain caused individuals to
is critical to be able to isolate factors that predict how
change employers frequently. Since the turn of the
long the individual will stay with a particular
century and the slowdown in economic activity, there
organization at the stage of organizational entry
is an ostensible oversupply of high-tech workers,
(Cable & Judge, 1996).
although projections of occupational growth suggest
Human resource researchers and practitioners have that any oversupply is likely to be temporary (United
long been intrigued by the question of how individuals States Department of Labor, 2004). Thus, this is an
embarking upon a career make employment choices, interesting occupational group to study because of
also referred to as the “attraction process” (Schneider, the turmoil it is experiencing.
1987). The decision about where to accept the first
We pose the broad question: “How do individual and
employment offer is a significant one: early work
situational factors interact in explaining a new IT
experiences and the socialization that occurs often
workforce entrant’s turnover intention?” We examine
have a profound impact on later stages of an
two “matched” individual and situational factors that
individual’s work life and have been shown to be a
are likely to be important for this set of high-tech
proximal cause of multiple significant outcomes such
workers: for the organization, the two factors are the
as turnover, satisfaction, and performance (Holton &
level of entrepreneurial risk inherent in the business
Russel, 1999). In general, the choice of the first job is
model of the employer and the amount of variety that
fraught with uncertainty since individuals typically
an IT job with the organization offers. From the
have little or no work experience to help shape their
individual perspective, these two factors are reflected
expectations about employers and workplace
in a new IT professional’s preferences for risk and
conditions in a concrete fashion. They are inundated
variety. These matched individual and situational
with information and advice about prospective
factors are consistent with the needs-supplies
employers from a variety of sources. Friends, family,
perspective in models of person-organization fit
and others in the social circle exert overt and covert
(Kristof, 1996). Risk and variety preferences
pressure to accept employment with organizations
represent individual needs, while organizational risk
that provide certain inducements, such as rewards,
and variety represent what the organization supplies
status, or security. Career placement centers and
to meet those needs.
counselors at academic institutions provide advice on
appropriate places of employment, and indeed, on We report the results of an experiment with new
careers to pursue. And recruiters compete for the entrants into the IT workforce where the level of risk
applicants’ attention, particularly in a tight labor and variety in a given employment context were
market, by painting a vivid picture of the benefits and manipulated. Such a research design allows us to
rewards of employment in their organization. Against gain further insights into an important organizational
this backdrop, early career entrants are challenged concern, the effect that its practices are likely to have
with synthesizing all this information and making an on turnover. We use four hypothetical work settings
informed choice about whom to work for in their first characterized by a parsimonious set of factors and
job. investigate their effects on future employees while
controlling for all other situational factors. As such,
Research in organizational behavior suggests that
because all other organizational factors are held
both individual and situational factors play a role in
constant for each individual while risk and variety are
organizational entry decisions. Drawing upon the

The DATA BASE for Advances in Information Systems 9 Volume 38, Number 1, February 2007
varied, the design of this study provides a powerful stability of job attitudes and affect among young
basis for examining the ability of the organization to professional engineers led them to conclude that
affect turnover intentions via its positioning related to individual and situational influences both matter in
risk and variety. Inspection of the moderating effects work reactions. Several other studies also support the
of individual preferences provides an understanding simultaneous existence of situational and individual
of the role of selection (of IT professionals with high or effects (Griffin, 2001; Glynn, 1998; Steel & Rentsch,
low preferences for risk and variety) and how 1997; Banks & Henry, 1993). For example, Steel &
organizations can expect their selection processes to Rentsch examined job satisfaction across a ten-year
interact with their “retention” practices (related to time period and showed that after accounting for the
organizational risk and variety). From the perspective effects of attitudinal stability, job characteristics
of the IT professional, the study provides a basis for explained significant variance in outcomes. In their
understanding the effects of their personal study, the success of job design interventions
preferences (for risk and variety) on their intentions to interacted with the personal characteristics of
stay with an employer, potentially offering guidance individuals.
for the job search and choice process.
Today, one perspective on the debate whether it is
the person or the situation that determines behavior is
Theoretical Background and Research the interactionist view of organizational behavior, as
Hypotheses articulated in Schneider’s (1987) influential attraction-
selection-attrition (ASA) framework. In this
Individual and Situational Influences in the perspective, organizational behavior is viewed as
Workplace being influenced by the characteristics of both the
individual and the organization. Scholars in the
The influence of individual and situational factors on interactionist tradition essentially reject the idea of
individuals’ work attitudes and behaviors continues to using either individual or organizational factors in
be the subject of much discussion among scholars in isolation to predict work attitudes and behavior and
organizational behavior. Among the individual factors exhort researchers to pay attention to a richer set of
examined in prior research, both dispositions such as factors that can influence these outcomes. Examples
personality variables and other more malleable of interactionist research may be found in the
characteristics such as job attitudes have been extensive literature on person-organization (P-O) fit,
studied. While dispositions are relatively stable which argues and presents empirical evidence for the
characteristics of individuals (Fridhandler, 1986), fact that employment attitudes and behaviors are
situational factors are the wide range of influences influenced by the congruence between individual
that characterize the individual’s work context and characteristics and organizational environments (e.g.,
environment. Early influential work by Staw and Ross Kristof, 1996; Kristof-Brown et al., 2005).
(1985) demonstrating that job satisfaction remained
considerably stable across time and across work In the IT literature, Agarwal and Ferratt (2000)
situations provided initial support for the relative proposed an extension to Rousseau’s (1995) theory
primacy of dispositional influences. Based on this of psychological contracts and argued that
work, scholars questioned the efficacy of managerial staying/leaving behavior would be influenced by an IT
interventions such as supervision and training or professional’s career motives, conceptualized as
situational factors inherent in the work environment in consisting of a preferred employment duration, career
promoting job satisfaction and other affective and anchor, and career stage. Implicit in their arguments
behavioral work outcomes. was the assertion that these three individual
characteristics would change over the life time of an
Since the publication of Staw and Ross’ work, IT professional. In later work, Agarwal et al. (2001;
researchers have offered some criticisms of the 2002) suggested that preferred employment duration
research and continued to examine the individual- with a specific organization is jointly determined by
situational issue from multiple perspectives. For career anchor, life stage, and competencies. They
instance, Gerhart (1987) raised questions about a further argued that the type of organization would
potential age bias in Staw and Ross’ study, and moderate the effects of these individual
observed that their assertions about the lack of characteristics on preferred employment duration with
influence of job-redesign programs were not that organization. Thus, in their conceptualization,
grounded in empirical evidence. Newton and Keenan individual preferences for length of stay with an
(1991) noted the existence of a methodological employer are driven by both individual and situational
problem with Staw and Ross’ work in that they used factors.
correlational analysis to test for stability in job
satisfaction. Newton and Keenan’s analysis of the Although there is now a growing corpus of work
devoted to understanding the effects of individual and

The DATA BASE for Advances in Information Systems 10 Volume 38, Number 1, February 2007
situational variables on job-related attitudes and Similarly, the effects of the variety offered by the
behaviors, there are also many unanswered employer (or the situation) are moderated by
questions (Kristof-Brown et al., 2005) and individual preferences for variety. The existence and
opportunities for further work. First is the crucial issue importance of preferences in work-related attitudes
of the number of characteristics of persons and and behaviors is underscored by Dawis (1991): he
situations that serve as explanations for behavior. notes that traits influence values that, in turn,
From a theoretical perspective, fewer characteristics influence preferences. Preferences are viewed as
are useful because they provide a more parsimonious being more proximal to choices among alternatives
explanation. From a managerial perspective, fewer than values (Dawis, 1991; Judge & Cable, 1997) and
characteristics should be easier and less costly to represent the needs and wants that individuals are
both evaluate when screening potential recruits for trying to satisfy via an employment relationship. To
employment and to influence in the organizational illustrate, an example of a personality trait is
environment. Much prior work has focused on profiles openness to experience, that would be reflected in
of individuals and organizations that involve a large work values that assign importance to broad-ranging
number of variables. For example, the Organizational work experiences. These work values then yield the
Culture Profile (OCP) developed by O’Reilly et al. individual characteristic of variety preference. We
(1991) has 54 value statements that yield eight further note that in examining these effects, it is
factors for individual values and seven factors for important to control for an individual’s propensity to
organizational values. With a goal of parsimony, one stay. Each of the constructs and relationships is
unanswered question we address is whether a elaborated upon below.
smaller number of important individual and
organizational characteristics explain an individual IT Workers and Situational Influences
outcome. Specifically, we focus on two such fit
characteristics: risk and variety. Considerable prior research has examined the
The second issue concerns itself with the motivations and management of high-tech workers in
characteristics of specific occupational groups and the information technology field (e.g., Cougar &
the work environments that typify the organizations Zawacki, 1980; Igbaria et al., 1991; Saxenian, 1994;
they are employed in. Holland’s (1985) influential Agarwal & Ferratt, 1999). Implicit in this research is
typology of occupations was developed for the the recognition of IT organizations and workers as
purpose of classifying vocational interests and important economic forces and interesting
preferences. In this typology, both people and occupational groups worthy of investigation. Of
environments are characterized as Realistic, particular note for this study are findings that IT
Investigative, Artistic, Social, Enterprising, and workers possess a high growth-need strength
Conventional (RIASEC). In performing a meta- (Cougar & Zawacki, 1980) and seek variety and
analysis of this theory, Barrick et al. (2003) note that technical challenge in job assignments (Hamblen,
the essence of RIASEC is that an “employee’s 2002). Further evidence for the importance of
satisfaction with a job, as well as propensity to leave studying this group of workers separately can be
that job, depend on the degree to which the found in surveys of the key issues confronting IT
individual’s personality matches his or her management, where attracting, developing, and
occupational environment” (p. 46). However, the retaining IT professionals consistently emerges as a
RIASEC model specifies occupational types (and critical management concern (Luftman, 2005;
associated work environments) at a high level of Luftman & McLean, 2004; Brancheau et al., 1996).
generality, representing distal variables in the In the past decade, this occupational group has
prediction of actual work behaviors and attitudes. witnessed a significant upheaval in at least two
Moreover, the typology was constructed to be related areas relevant to employment. First,
applicable to workers of all types, independent of their increasing organizational dependence on information
specific occupational status. Thus, a second issue we technology fueled by the Internet revolution resulted
address here is the identification of factors (both in a major supply-demand asymmetry during the
individual and organizational) that are likely to be second half of the previous decade that still persists
more salient to the specific population of IT workers. for certain technical competencies. Second, the
business climate has been characterized by
Conceptual Model considerable IT-based entrepreneurial activity and
innovation, driven largely by the capabilities offered
The conceptual model underlying this research is by new information technologies.
shown in Figure 1. As indicated there, we argue that
the effects of situational risk on turnover intention are
moderated by individual preferences for risk.

The DATA BASE for Advances in Information Systems 11 Volume 38, Number 1, February 2007
Preferred Risk

Situational
Risk Turnover
Intention
Situational
Variety

Preferred Propensity
Variety To
Stay

Figure 1. Conceptual Model of Individual and Situational Influences

With burgeoning demand, organizations struggled to individuals possess distinct propensities to stay with
recruit and retain IT workers, and IT wages were an employer is also reflected in Rousseau’s (1995)
driven up as a result. IT turnover rates as high as characterization of “careerists.” Careerists regard
20% were reported in some industries, further their current employment situation as merely a
exacerbating organizational costs associated with stepping stone to better opportunities. In a study of
maintaining the IT human capital base (Agarwal & the orientations of recent graduates toward their first
Ferratt, 1999). Given the challenges associated with job, Rousseau (1990) found that certain individuals
managing this occupational group, it is useful to were seeking to move quickly through employment
attempt to characterize organizations in terms of situations in search of advancement. These careerists
aspects that are likely to influence the decision of IT expected to stay with an organization for less than
workers to remain in the employment relationship for three years on average. At the other extreme,
a greater length of time. As explained below, we individuals low on careerism expected to stay for five
argue that risk and variety represent two important or more years.
situational characteristics with matching individual
The difference between propensity to stay found in
preferences that are relevant to IT workers’ turnover
the economics and organizational behavior literatures
intentions. However, as noted by researchers in
is informative. This construct from the economics
human capital theory, labor economics, and
literature is an individual difference inferred from
organizational behavior (e.g., Lazear, 1995; Topel &
behavioral observation. By contrast, it is based on
Ward, 1992; Rousseau, 1990), there may be
cognitions in the organizational behavior literature.
differences in individuals’ propensities to stay in a job.
Behavioral observation indicates the validity of the
We, therefore, include such a propensity as a control
construct.
variable for turnover intention.
Since we seek to control for this variable prior to
Controlling for Propensity to Stay behavior, we use a cognitively-based definition of the
construct. Individuals asked to think generally about
One important individual difference variable that how long they would stay with an employer need an
serves as a predictor of job mobility is the inherent anchor to describe that employer. We make the
propensity to be a “stayer” or a “mover” (Topel & assumption that this cognitive expectation or
Ward, 1992). While movers exhibit considerable job propensity to stay is anchored in an employer that the
mobility, stayers are characterized by relatively long individual would prefer. Another anchor that could be
tenure with specific employers. Thus, in a similar used is any employer; however, the uncertainties
organizational situation, stayers are likely to exhibit associated with this abstract employer would most
lower turnover intentions than movers. The notion that likely be much greater than those associated with one

The DATA BASE for Advances in Information Systems 12 Volume 38, Number 1, February 2007
for which the individual would prefer to work, leading which may come with considerable risk (Cher, 2000).
to a construct whose measurement would be Indeed, the much-publicized failures of a number of
inherently less reliable. In our model, propensity to entrepreneurial ventures exposed the risk inherent in
stay refers to the individual’s expected employment their business models.
duration with an employer for which the individual
At the same time, traditional brick and mortar
would prefer to work, i.e., an ideal employer.
organizations are also transforming themselves by
It is important to note that our conceptualization of focusing on core competencies and shedding
propensity to stay clarifies the distinction between functions that are peripheral to their main mission.
preferred employment duration as discussed in The ostensible reason for this shift is that a focus on
Agarwal and Ferratt (2000) and preferred core competencies should improve the efficiency and
employment duration with a specific organization in effectiveness of the organization, thereby increasing
Agarwal et al. (2001; 2002). Propensity to stay is a returns to key stakeholders. Many of these
relatively stable individual difference variable that transformations are IT-enabled. Increasing reliance
reflects inherent staying tendencies with an ideal on IT increases the vulnerability of the organization to
employer. We recognize that an individual’s ideal a scarce supply of IT professionals or short-staying IT
employer may change over time, but we would expect professionals when the organization requires firm-
that change to evolve gradually over time, e.g., as the specific competencies, thereby amplifying business
individual moves through life and career stages. In risk. Thus, while IT-enabled business innovation
effect, then, propensity to stay is an individual’s offers an opportunity for reward, these rewards are
preferred employment duration within the context of not without risk. Theories of choice assert individuals
an ideal employer. This conceptualization, which is will evaluate risk/reward tradeoffs in their decision-
similar to that in Agarwal and Ferratt (2000), is distinct making processes (Kahneman et al., 1982). These
from preferred employment duration with a specific theories also argue that the level at which risk is
organization, which is based on the interaction traded off with reward differs across individuals.
between the individual and specific situational
All other things being equal, situational risk is likely to
variables (see Agarwal et al., 2001; 2002). If a
be positively related to turnover intention. For
specific organization would be identical to an
example, between two individuals in situations with
individual’s ideal employer, then preferred
the same reward, the one in the situation with the
employment duration with that specific organization
higher risk is likely to have the higher turnover
would be identical to the individual’s propensity to
intention. However, there is considerable evidence to
stay.
suggest that individuals vary in their risk-taking
propensities (e.g., Dulebohn, 2002). To the extent
Situational Risk that an IT worker’s preference for risk and reward is
consonant with that offered by the organization – as
Organizations have different levels of employment suggested by considerable prior research (e.g.,
risk and commensurate returns associated with them. O’Reilly et al., 1991; Cable & Judge, 1996) – this
Prior research has tended to focus more on the should result in a lower turnover intention. The
“returns” aspect of organizations, by considering underlying causal mechanism is one of greater job
attributes such as pay and reward systems. However, satisfaction and organizational commitment (Hom &
we suggest that the returns associated with a Griffeth, 1995): when workers view their preferences
particular employment situation are likely to be as congruent with the situation, they develop stronger
evaluated in the context of the risks that the employer affective commitment (Meyer et al., 2002) and are
embodies. It is a truism to assert that advances in less likely to exhibit withdrawal behaviors such as
technology are changing organizations. In particular, turnover. These arguments suggest the following
web-based technologies led to the emergence of hypothesis:
young entrepreneurial ventures in the shape of
dotcom organizations during the latter half of the H1a: Situational risk has a positive relationship
1990s. This entrepreneurial activity was with turnover intention.
unprecedented in volume: in 1999 alone, 96 Internet H1b: Preferred risk moderates the relationship
companies went public, accounting for a total funding between situational risk and turnover intention;
of $6.7 billion (Tice, 2000). Some of the attractions of the relationship becomes less positive as
dotcom organizations for IT professionals are the preferred risk increases.
excitement and the potentially large financial gains,

The DATA BASE for Advances in Information Systems 13 Volume 38, Number 1, February 2007
Situational Variety human capital constraints the organization faces.
Thus, “situations” encountered by IT workers exhibit
While evaluating employment opportunities, a second considerable variance in the level of variety offered.
dimension that IT professionals are likely to consider We would expect variety to be negatively associated
is the range of technologies and experiences that are with turnover intention in that higher levels of variety
potentially available in an organization, i.e., the variety will likely imbue the IT job with greater significance
of opportunities available with a given employer. and challenge; the associated intrinsic rewards
Agarwal and Ferratt’s (1999) field research across a should lead to lower turnover intention. However, the
large number of organizations underscored the notion that individuals vary in their variety-seeking
importance of interesting work as an important behaviors is echoed in a significant body of consumer
predictor of intentions to stay. Indeed, surveys of behavior research (e.g., Ratner & Kahn, 2002; Inman,
“Best places to work” in information systems (Brandel, 2001; Trivedi, 1999). Given that innate preferences
2006) consistently point to the nature of the work for variety vary, these preferences are likely to
(e.g., leading-edge technologies, interesting projects) moderate the effects of variety on turnover intention.
as a significant driver of IT professionals’ attitudes Theoretically, the mechanism causing this interaction
and work behaviors. is similar to that discussed above for risk, and
Organizations exhibit significant diversity in the types reflected in the notion of P-O fit (e.g., Kristof, 1996).
of information technologies they utilize and in the More specifically, a match between preferences for
complexity and variety of their business processes. variety and opportunities for variety should result in
Some firms choose to standardize technology more positive outcomes. Thus, we propose the
platforms and use a small number of technologies to following hypotheses:
support all activities, while others run applications on H2a: Situational variety has a negative
multiple platforms. Some organizations are relationship with turnover intention.
characterized as laggards or traditional IT shops in
that their information technology has not been H2b: Preferred variety moderates the relationship
refreshed for several years, whereas others prefer to between situational variety and turnover intention;
stay closer to the leading edge and are constantly the relationship becomes more negative as
updating their technology base. Likewise, the extent preferred variety increases.
of infusion of information technology across different
processes within a business can vary considerably, Methodology
as can the complexity of a firm’s business processes.
Some organizations may choose to utilize IT purely in Sample
a utility mode, e.g., for automating back-office
operations, while for others IT is a source of Information systems majors about to graduate and
competitive advantage and permeates every accept employment constitute the subject pool. This
business process within the company (Applegate et choice of subjects limits the population to new college
al., 2003). Organizations can have a single strategic graduate entrants to the IT workforce. This
mission and offer a limited range of products and population, however, is important for those
services, or they may be diversified with a broad organizations that renew their IT workforce through
range of market offerings. Finally, there is college recruiting. In addition, this population makes
considerable variance in the types of job rotation and this study relevant for the job search and choice
other development practices adopted by IT decisions made by a large number of individuals
organizations (Agarwal & Ferratt, 1999). about to enter the IT profession. Voluntary
participants consist of 63 senior management
Each of these environments represents a very
information systems (MIS) majors at two U.S.
different set of work opportunities and experiences for
universities (35 at a private midwestern university and
an IT worker. For instance, in a large, global
28 at a public eastern university). Characteristics of
organization with varied IT-related activities, the
the sample, including demographics and the various
breadth and depth of experiences available to an IT
measures described below, are shown in Table 1. Of
professional are likely to be considerably greater than
the 63 subjects, 43 (68%) report some work
those available in a smaller, narrowly focused
experience in IT, with 39 indicating more than one
organization. Staying in the former type of
year. We conclude that the subjects are not naïve
organization provides potential opportunities to
about work environments and can make well-formed
acquire a diversity of IT skills through a broad range
judgments about different employment situations.
of experiences. Or it could be argued that IT
professionals in smaller organizations are required to
perform a variety of different tasks because of the IT

The DATA BASE for Advances in Information Systems 14 Volume 38, Number 1, February 2007
Private Midwestern Public Eastern Combined
Variables University Sample University Sample Samples
Mean S.D. N Mean S.D. N Mean S.D. N
Demographics++
Gender (0=woman, 1=man)* .74 .44 35 .44 .51 27 .61 .49 62
+
Ethnicity See below 35 See below# 28 See below +#
63
Years until Retirement 40.2 11.5 35 39.9 7.2 21 40.1 10.0 56
Individual Measures
Propensity to Stay 4.1 1.2 35 4.5 1.2 28 4.3 1.2 63
(Number of years to stay) 5.9 5.6 35 6.3 5.5 27 6.1 5.5 62
Preferred Risk 2.4 1.1 35 2.7 1.5 28 2.5 1.3 63
Preferred Variety 5.5 1.3 35 5.3 1.5 28 5.4 1.4 63
Situational Measures by Scenario
Scenario: Low Risk, Low Variety
Turnover Intention 4.4 1.0 34 4.4 1.3 28 4.4 1.1 62
(Number of years to stay) 3.5 1.8 34 4.1 3.2 28 3.8 2.5 62
Scenario: Low Risk, High Variety
Turnover Intention 3.4 0.9 35 3.0 1.3 27 3.2 1.1 62
(Number of years to stay) 6.2 4.1 35 9.3 11.1 26 7.6 7.9 61
Scenario: High Risk, Low Variety
Turnover Intention 5.4 1.0 35 5.1 1.3 28 5.3 1.1 63
(Number of years to stay) 1.8 1.0 35 2.6 2.4 28 2.2 1.8 63
Scenario: High Risk, High Variety
Turnover Intention 5.0 1.1 34 4.8 1.1 28 4.9 1.1 62
(Number of years to stay) 2.8 1.9 34 3.5 2.3 28 3.1 2.1 62
* p < .05; (Means of the two samples are not significantly different except as noted with *)
++
43 subjects (68%) report some prior work experience in IT. Of these, 39 report at least one year.
+
86% (30) White, 6% (2) Asian, 3% (1) Hispanic, 3% (1) Black/African American, 3% (1) Other
#
32% (9) White, 54% (15) Asian, 7% (2) Hispanic, 4% (1) Black/African American, 4% (1) Other

Table 1. Sample Characteristics

Measures • What is the length of time you would prefer to


stay with this organization at this stage of your
Propensity to Stay. The measure for propensity to
career? _________ years
stay is taken from Agarwal et al. (2002), but is
anchored to the context of an ideal employer. • This question is used with the following
Participants are instructed: “Think about the type of demographic question to obtain the fourth item
organization for which you would prefer to work.” in the measure:
They answer four questions related to how long they
are likely to stay with this organization. Three • Approximately how many years do you have
questions are on a seven-point Likert scale, where 1 until retirement? _________ years
is strongly disagree and 7 is strongly agree: The fourth item is the ratio of these last two
• My preference would be to stay with this questions (constrained to a maximum of 1),
organization for a long period of time. converted to a seven-point scale (by multiplying by
six and adding 1). Propensity to stay is the mean of
• I would most likely quit this organization within a these four items.
year.
Turnover Intention. For each of the scenarios in the
• I would most likely quit this organization within experiment (see Procedures below), participants are
three years. instructed in Section 1 of the scenario to assume that
they have just been hired by an organization with the
The second and third questions are reverse scored.
risk and variety profile described in the scenario (see
The fourth question is the following:
Appendix A for a sample scenario). They are then
instructed as follows: “Think about the organization
that has just hired you (see Section 1 above) in

The DATA BASE for Advances in Information Systems 15 Volume 38, Number 1, February 2007
comparison with the type of organization for which guaranteed work, and overall financial and
you would prefer to work. Assume that this employment risk – represent risk items.
organization (i.e., the one that has just hired you) is Organizational risk has generally been measured as
similar to your preferred type of organization on all some type of variance in return on assets and return
characteristics not included in the profile described on equity (e.g., Miller & Reuer, 1996) and, to the best
above. Given all those characteristics and the ones of our knowledge, there are no scale-based measures
described in the above profile for the organization that of organizational risk in the extant literature. Therefore
has just hired you, respond to each of the following we based the operational measure of risk on
statements by circling a number that best represents considerable work in labor economics that identifies
your degree of agreement or disagreement with each compensation and employment as the two major
statement.” The same four questions used to sources of risk that employees face (Turner, 2001;
measure propensity to stay are used to ask Berloffa & Simmons, 2003).
respondents how long they are likely to stay with the
The last five items – technologies involved in work,
organization described in the profile. The
skills developed, opportunity to move among different
demographic on years until retirement is used
IT jobs internally, experience gained by working on
similarly to obtain the fourth item. The mean of the
projects, and organization size – represent variety.
four items, scored in the same manner as propensity
Given the lack of existing scales for measuring
to stay, is reverse scored to obtain the measure of
organizational variety, we constructed these items
turnover intention. Each participant has a measure of
(also used in Agarwal et al., 2002) based on the
turnover intention for each of the four specific
conceptual definition of variety as the opportunities
organizational risk and variety profiles.
available in an organization for a range of different
Preferred Risk and Variety. The individual measures work experiences and associated skill development.
of preferred risk and variety use the same instructions A low risk (or variety) profile has an X placed on the
as those used for propensity to stay: “Think about the 7-point scale for each risk item at a value of 2. A high
type of organization for which you would prefer to risk (or variety) profile has an X placed at 6.
work.” Preferred risk is measured by the mean of the
following three items (reverse scored) on a seven- Procedures
point Likert scale, where 1 is strongly disagree and 7
This study is based on an experiment with
is strongly agree:
individuals preparing for entry into an IT career. The
• This organization provides stable, guaranteed specific employment context was controlled by
work and income. experimental manipulation of the two situational
factors in Figure 1. Specifically, situational (i.e.,
• A job with this organization would be fairly organizational) risk and variety were manipulated by
secure. randomly presenting each participant with scenarios
• The risks associated with this organization are describing organizations with specific risk and
low. variety profiles as described below. Participants
were instructed to assume that they had just been
Preferred variety is measured by the mean of the hired by the organization in a scenario. Participants
following three items on a seven-point Likert scale, answered turnover intention questions focused on
where 1 is strongly disagree and 7 is strongly agree: how long they would expect to be employed with an
• I could obtain a wide variety of experience with organization having the profile of characteristics
IT in this organization. identified in a scenario. This experimental design
allows us to have various combinations of high and
• I could move among many different types of jobs low organizational characteristics for hypothesis
in IT in this organization. testing. Such a within-subjects design that permits
the assessment of the relative attractiveness of
• I could develop many different skills related to a
different organizational environments is typically
career in IT in this organization.
referred to as a policy capturing study (Judge &
Situational (i.e., Organizational) Risk and Variety. Bretz, 1992; Bretz & Judge, 1994).
Each scenario of the experiment consists of an
The data for this study are obtained at two points in
organization with a specified risk and variety profile.
time. In the initial time period (called the baseline),
The profile is presented via semantic differential type
participants answer questions on individual factors
scales, as in Agarwal et al. (2002). For an example,
and demographics. Relative to Figure 1, they answer
see Appendix A which presents a low risk, high
questions on their propensity to stay with an
variety profile. The first five items in the profile –
organization for which they would prefer to work; they
financial stability, guaranteed income, job security,

The DATA BASE for Advances in Information Systems 16 Volume 38, Number 1, February 2007
also answer questions on their preferred risk and psychometric properties. Cronbach alpha reliability
variety. coefficients for the scales used in this study are all
greater than .70 as shown in Table 2. These
At a later time (approximately 3 weeks after the
coefficients indicate that all scales have adequate
baseline measurement), they are given four scenarios
internal consistency. Moreover, the situational risk
corresponding to four types of organizations that have
and variety scales are significantly correlated with
one of the following profiles for situational risk and
turnover intentions in a manner consistent with
variety: (1) low risk, low variety, (2) low risk, high
theoretical expectations (negative for situational
variety, (3) high risk, low variety, and (4) high risk,
variety and positive for risk), attesting to the predictive
high variety. In all other respects, the organizations in
validity of the measures. Factor analysis of the items
each scenario are similar to their preferred employer.
comprising the measures completed by participants –
The scenarios and related questions are presented in
propensity to stay, preferred risk, preferred variety,
random sequence, one right after another, in a single
and turnover intention – shows that these expected
sitting.
factors cleanly emerge with eigenvalues greater than
For the last of the four scenarios, participants answer 1 and appropriately high item loadings on the a priori
questions about the risk and variety that they perceive factors for each measure (all greater than .60) and no
as being associated with the organization presented high cross loadings (none greater than .40). Even
in that scenario to allow a manipulation check for risk though the same items are used to measure
and variety. If the manipulation is working as propensity to stay and turnover intention, the referents
expected, subject responses to the risk and variety (i.e., preferred organization vs. organization with a
they perceive in the scenario should match the risk specific risk and variety profile) are different. Indeed,
and variety assigned in the scenario description. The the factor analyses demonstrate that these and the
questions used are the same as those used for other scales have discriminant and convergent
measuring preferred risk and variety, but the referent validity (See Appendix B).
organization is the one in the scenario rather than the
A series of linear mixed effects models are used to
ideal employer. These questions were not used after
test the hypotheses. Such models are particularly
each scenario to limit the amount of time required for
suited to the analysis of data that involves repeated
the experiment and to limit participant fatigue and
measurements for different treatments on the same
monotony. Since the scenarios are in random
subject, as is the case in this experiment (McCulloch
sequence, the manipulation check covers the range
& Searle, 2001). This modeling technique specifically
of situations.
accounts for correlations between the repeated
measures. In all models, individuals with nested risk
Statistical Analysis
by variety treatment sequences are specified as
Prior to the analyses for testing hypotheses, various random effects and all other variables included in the
scale analyses are conducted to ascertain model are specified as fixed effects.

Correlations
Std.
Mean Deviation Turnover Propensity Preferred Preferred Situational Situational
Intention to Stay Risk Variety Risk Variety
Turnover Intention2 4.5 1.5 0.853
Propensity to Stay4 4.3 1.2 -.16 0.73
Preferred Risk4 2.5 1.3 .11 -.36** 0.84
Preferred Variety4 5.4 1.4 -.04 .43** -.47** 0.87
Situational Risk5 4.0 2.0 .47** .06 .14 -.22 --
Situational Variety5 4.0 2.0 -.25* .20 -.16 .06 -.02 --
Notes: * p ≤ .05; ** p ≤ .01; *** p ≤ .001
1. N = 63 for all means, standard deviations, and correlations except for those involving turnover intention, where N = 62.
2. Measured for each scenario. The data reported here are based on the second scenario presented to each participant. Since
scenarios were presented in a random sequence, all situational risk and variety conditions are represented in the data presented
here.
3. Coefficient of reliability (Cronbach’s alpha) reported on the diagonal.
4. Measured at the baseline.
5. Experimentally assigned.
Table 2. Scale Analyses1

The DATA BASE for Advances in Information Systems 17 Volume 38, Number 1, February 2007
In the first stage of analysis, propensity to stay is used positive to be consistent with the hypothesis, i.e., to
in the model as a control variable. For hypothesis 1, show that for lower risk situations turnover intention
situational variety, preferred variety, and their increases as preferred risk increases. A similar
interaction are also used as control variables in the analytic approach is used to test hypotheses 2a and
first stage. Situational risk is added to the model in the 2b.
second stage of analysis to assess its relationship
with turnover intention after controlling for the non-risk Results
related variables in Figure 1. Hypothesis 1a is tested
by examining the change in information criteria Manipulation Check
between the estimated models from the first and
For the experimental scenario presented fourth to
second stages. If the model with situational risk fits
each participant (see Procedures), the absolute
the data better than the model without it, the
difference between the assigned and perceived
information criteria should decrease in value from
values of both risk and variety (which are both on 7-
stage one to stage two. In addition, the sign of the
point scales) is not significantly different from 1.0.
situational risk coefficient is examined to determine if
Since the assigned values are either 2 (to represent
it is consistent with the hypothesis. For example,
low risk or low variety) or 6 (to represent high risk or
given that situational risk is a categorical variable (i.e.,
high variety), perceived values within 1.0 of the
2 = low risk and 6 = high risk), one category will serve
assigned values indicate that the perceived values of
as the base condition. If the high risk situation is the
risk and variety are consistent with the assigned
base condition against which low risk is compared in
values. Thus, the experimental procedure
the statistical analysis, the low risk situation
demonstrates internal validity.
coefficient must be significant and negative to be
consistent with the hypothesis, i.e., to show that lower
situational risk is associated with lower turnover Hypothesis Tests
intention.
H1a asserts that situational risk has a positive
Preferred risk and the interaction of preferred risk and relationship with turnover intention. As shown in Table
situational risk are added to the model in the third 3, after controlling for the effects of propensity to stay
stage of analysis. Hypothesis 1b is tested by (plus the other variables in the model not related to
examining the change in information criteria between hypothesis 1), the introduction of situational risk yields
the estimated models from the second and third a decrease in the information criteria (e.g., AIC
stages. If the model with the moderating effects of decreases from 828.1 to 756.5). The F statistic for
individual characteristics fit the data better than the situational risk is 89.1 with a significance of .000.
model without moderation, the information criteria Given that the high risk situation is the base condition
should decrease in value from stage two to stage against which low risk is compared in the statistical
three. In addition, the sign of the coefficient for the analysis, the low risk situation coefficient must be
interaction of preferred and situational risk is significant and negative to be consistent with the
examined to determine if it is consistent with the hypothesis.
hypothesis. Continuing with the example from above,
the coefficient in that case must be significant and

Hypothesis 1
Controls + Controls + Situational and Preferred
ModelÎ Controls
Situational Risk Risk and their Interaction
-2 Restricted Log Likelihood 824.1 752.5 750.7
Akaike’s Information Criterion (AIC) 828.1 756.5 754.7
Bozdogan’s Criterion (CAIC) 837.0 765.4 763.6
Hypothesis 2
Controls + Controls + Situational and Preferred
ModelÎ Controls
Situational Variety Variety and their Interaction
-2 Restricted Log Likelihood 774.1 747.1 750.7
Akaike’s Information Criterion (AIC) 778.1 751.1 754.7
Bozdogan’s Criterion (CAIC) 787.1 760.0 763.6
Table 3. Information Criteria for Hypothesis Tests

The DATA BASE for Advances in Information Systems 18 Volume 38, Number 1, February 2007
Its value is -1.315 with a standard error of .139, a t- situational variety is 31.5 with a significance of .000.
value of -9.441, degrees of freedom of approximately Given that the high variety situation is the base
180, and significance of .000 (see Appendix C). Thus, condition against which low variety is compared in the
H1a is supported, i.e., individuals who perceive the statistical analysis, the low variety situation coefficient
organization to be high risk have greater turnover must be significant and positive to be consistent with
intentions. the hypothesis, i.e., to show that lower situational
variety is associated with higher turnover intention. Its
H1b asserts that the positive relationship between
value is .768 with a standard error of .137, a t-value of
situational risk and turnover intention becomes less
5.611, and significance of .000 (see Appendix C).
positive as preferred risk increases. In other words,
Thus, H2a is supported.
for higher risk situations individuals with higher
preferred risk should have lower turnover intention H2b posits that the negative relationship between
than individuals with lower preferred risk; for lower situational variety and turnover intention becomes
risk situations, individuals with higher preferred risk more negative as preferred variety increases. In other
should have higher turnover intentions. The addition words, for higher variety situations, individuals with
of preferred risk as a moderator yields a further higher preferred variety should have lower turnover
decrease in the information criteria (e.g., AIC intention than individuals with lower preferred variety;
decreases from 756.5 to 754.7 in Table 3). The F for lower variety situations, individuals with higher
statistic for the interaction of preferred and situational preferred variety should have higher turnover
risk is 8.1 with a significance of .005. As above, the intentions. The addition of preferred variety as a
high risk situation is the base condition against which moderator does not yield a further decrease in the
low risk is compared in the statistical analysis; thus, information criteria (e.g., AIC increases from 751.1 to
the coefficient for the interaction between preferred 754.7 in Table 3). The F statistic for the interaction of
and situational risk must be significant and positive to preferred and situational variety is 1.822 with a
be consistent with the hypothesis. The interaction significance of .179. Thus, H2b is not supported;
coefficient is .296 with a standard error of .104, a t- however, the sign of the coefficient is in the direction
value of 2.853, and significance of .005 (see hypothesized. As above, the high variety situation is
Appendix C). These findings indicate that H1b is the base condition against which low variety is
supported. The interaction between preferred and compared in the statistical analysis; thus, the
situational risk is graphically illustrated in Figure 2. coefficient for the interaction between preferred and
situational variety must be significant and positive to
Moderating Effect of Preferred Risk be consistent with the hypothesis, i.e., to show that for
lower variety situations turnover intention increases
as preferred variety increases. The interaction
7.0
coefficient is .135 (see Appendix C). Figure 3
provides a visual plot of the interaction of situational
Turnover Intention

6.0 High Preferred and preferred variety.


Risk
Moderate Moderating Effect of Preferred
5.0
Preferred Risk
Variety
Low Preferred
4.0 Risk
6.0
High Preferred
Variety
Turnover Intention

3.0
Lo Hi 5.0
Moderate
Organizational Risk Preferred
Variety
4.0
Low Preferred
Figure 2. Moderating Effect of Preferred Risk Variety
3.0
H2a asserts that situational variety has a negative
Lo Hi
relationship with turnover intention. As shown in Table
3, after controlling for the effects of propensity to stay Organizational Variety
(plus the other variables in the model not related to
hypothesis 2), the introduction of situational variety Figure 3. Moderating Effect of Preferred Variety
yields a decrease in the information criteria (e.g., AIC
decreases from 778.1 to 751.1). The F statistic for

The DATA BASE for Advances in Information Systems 19 Volume 38, Number 1, February 2007
Limitations For IT managers, our conceptual model (Figure 1)
helps direct their attention toward practices affecting
The results from this study must be interpreted in perceptions of the situation (risk and variety) and our
light of four limitations that future studies could be empirical findings reinforce the importance of
designed to address. First, the sample is not a large individual and situational factors. It also helps direct
random sample of new entrants to the IT workforce. them to think of practices affecting the types of
Nevertheless, the results are strong enough with this individuals their organization recruits, selects, and
sample to suggest that studies with additional retains since individual characteristics moderate the
samples would be appropriate to assess how effect of the situation on turnover intention,
generalizable the results are. Second, the scenarios particularly when considering risk preferences (H1b).
represent a broad range of potential employers, but Finally, our results suggest that the effects of
the distribution of actual employers with the range of organizational practices, whether directed at affecting
risk and variety represented in the experimental perceptions of the situation or affecting the types of IT
situations is unknown. To understand how professionals the organization employs, will affect
generalizable our results are to actual employers of turnover intention through the interaction of the
IT professionals, future studies could include situation and individual characteristics, particularly
assessments of the risk and variety of actual when considering risk.
employers. Of course, as with any study based on
For prospective employers of IT professionals, our
simulated situations, field studies are needed to
results have important implications for how they
support the generalizability of the results to real
position themselves to potential new hires. Individuals
situations. Third, although we experimentally
form assessments of organizational characteristics at
controlled the situational data, the data for the
the stage of organizational entry by processing and
individual variables – propensity stay, preferred risk,
synthesizing information from multiple sources.
preferred variety, and turnover intention – were
Organizations reveal information about themselves
collected from a single source. We did collect data
through a variety of mechanisms, both purposely and
for propensity to stay, preferred risk, and preferred
inadvertently. For instance, they may develop
variety after the survey and with alternative
promotional materials that are distributed to potential
measures. Similar results were obtained, indicating
applicants. For college recruiting, organizational
that the results are fairly robust with regard to
representatives often host information sessions on
instruments. Finally, this study focused on
campus. As discussed extensively in the realistic job
understanding turnover intentions. Turnover
preview literature (e.g., Phillips, 1998), the more
intentions do not necessarily lead to actual turnover.
accurate this information is, the greater is the
It would be informative to study how the variables
likelihood of achieving organizational effectiveness
studied in this research on new IT workforce
through outcomes such as job satisfaction,
entrants are related to actual turnover.
organizational commitment, and job performance. A
wide variety of organizational data is also accessible
Discussion and Conclusion to the discerning job seeker through, for example,
publicly available corporate documents such as
Our goal in this paper is to understand the role of a annual reports and press coverage of significant
parsimonious set of two individual and situational organizational events. Another important source of
characteristics expected to be important to IT information over which organizations have less
professionals, viz., risk and variety, on the turnover control is current and former employees; often the
intentions of IT workers who are entering the IT testimonials and experiences of these workers offer
workforce. By using an experimental research design, new applicants a richer and more contextualized view
we are able to examine the interactions between of organizational life. Together, these multiple sources
these factors across a range of organizational of information help shape the applicant’s opinion
situations. We find that organizational risk interacts about specific characteristics of the organization. The
with individual preferences for risk in explaining implication of this study for prospective employers of
turnover intentions, while organizational variety IT professionals then is that risk and variety are
exhibits a main effect on turnover intentions with no powerful, parsimonious characteristics of the
interaction with individual preferences for variety. employment context, and realistic pictures of these
Thus, the former result supports the interactionist situational characteristics should be emphasized
perspective, while the latter questions its basic through such channels.
premise. The findings yield several implications for
practice and research that are elaborated upon Several areas for future research arise from this
below. study. Six such areas are discussed next. First, even
though risk and variety are important predictors of

The DATA BASE for Advances in Information Systems 20 Volume 38, Number 1, February 2007
turnover intentions, additional potentially powerful fit The need to include competencies in a model of IT
characteristics could be identified and investigated. employment behavior is evident when one considers
Some of the characteristics in O’Reilly et al.’s (1991) the rapid pace of technological change in IT, along
OCP might arguably be correlated with risk and with variations in demand, supply, and market value
variety. For instance, items related to innovation, of various competencies. On the one hand, technical
experimenting, risk-taking, security, and stability all skills decay rapidly as technologies change, and the
load on the individual factor of “innovation” and are new competence-destroying technologies often have
similar to the notion of risk described here. Likewise, limited lives, suggesting that the competencies of IT
items tapping into opportunities and professional professionals erode rather than increase over time.
growth, which loaded on different factors in the OCP On the other hand, IT organizations often need IT
(aggressiveness and emphasis on rewards, professionals with managerial and business skills that
respectively), are reflective of our notion of variety. arguably increase in value with experience, in
Although the explanatory power of the OCP is limited, general, and perhaps even more with firm-specific
the other values measured in the OCP may suggest experience. In addition, the organization’s need for IT
additional potentially relevant fit characteristics. professionals with specific competencies may be high
for a limited period of time, which may coincide with a
Second, there are several ways in which the
labor market that has a particularly acute undersupply
experiment described here can be extended and
of those competencies. Furthermore, IT workers may
enriched. For instance, we assumed that risk and
choose to develop a variety of different competencies.
variety are important individual and situational
For instance, they may focus on generic skills or on
characteristics and, with a goal of parsimony,
firm-specific skills (Becker, 1983). Generic skills are
included only these two factors. An alternative
those that can be traded in the open labor market and
approach, in the spirit of the P-O fit literature (Kristof,
that are equally valuable to all employers. By contrast,
1996), would be to elicit the importance of different
firm-specific skills, which include deep knowledge of
job and organization characteristics from individuals
the firm’s context, social networks and relationships
prior to presenting the organizational scenarios. Such
developed within the firm, and an understanding of
a strategy could help alleviate the potential for social
firm culture, may be of limited value to competing
desirability bias in the responses (e.g., a preference
employers.
for variety is generally considered good) and also
provide insights into additional factors that should be Fourth, our study was conducted in the 2001-2002
included in the model. Further, we chose risk and time frame after the dot-com bubble was already in
variety as two factors that drive the turnover intentions decline and there was a purported excess supply of
of new entrants. It may well be the case that for IT labor in the market. We anchored our
individuals who are already in an employment measurement of preferred risk and variety on the
situation, factors other than risk and variety are more characteristics of an “ideal” employer. It is possible
salient to determining their turnover intentions. This that the profile of such an ideal employer may be
possibility can be investigated in future research. affected by the labor market situation and other
environmental factors. Future research could theorize
A third avenue for future research is to examine the
about and test the effects of such factors on IT
antecedents of individual preferences. Much research
professionals’ ideal employers.
asserts that individual preferences for situational
factors are driven by their personal characteristics It would also be interesting to examine if risk and
(Judge & Cable, 1997; Kristof, 1996; Kristof-Brown, variety are important in explaining turnover intention
2000). In other words, what an individual seeks in an in IT professionals other than new entrants or in
employment relationship is a function of some occupational groups that are relatively more stable
attributes that characterize the job seeker. Therefore, and less entrepreneurial than the IT profession, such
in light of our findings regarding risk, it would be as retail banking. Our sample is limited to new
fruitful to examine what the drivers of risk preferences entrants to the IT profession, and our hypothesis that
are. Given our focus on IT workers, one important the person-situation interaction on the dimensions of
driver of risk preferences may be the competency risk and variety would be significantly associated with
bundle the IT worker perceives himself or herself as turnover intention is predicated on idiosyncratic
possessing, characterized in terms of the value of the characteristics of the IT profession. Confirming these
knowledge, skills, and abilities (KSA) possessed and relationships with established IT professionals, e.g.,
the performance level at which they are exhibited. those in the later career stages studied by Hsu et al.
Competencies are one instantiation of the general (2003), and examining them with other professions
KSA construct prevalent in the job choice and would be interesting directions for future research.
employment behavior literature.

The DATA BASE for Advances in Information Systems 21 Volume 38, Number 1, February 2007
Finally, the finding that the individual characteristic of Conference (SIG CPR), Chicago, IL, pp. 158-
preferred risk moderates the effects of situational risk 166.
but that the individual characteristic of preferred Agarwal, R. and Ferratt, T.W. (2001). “Crafting an
variety does not moderate the effects of situational HR Strategy to Meet the Need for IT Workers,”
variety suggests that some situational characteristics Communications of the ACM, Vol.44, No.7, pp.
dominate individual characteristics while others do 58-64.
not. Since these findings both confirm and question Agarwal, R. and Ferratt, T.W. (2002). “Enduring
the basic premise of the interactionist perspective, it Practices for Managing Information Technology
would be helpful to understand better the situational Professionals,” Communications of the ACM,
or individual characteristics that yield such results. Vol. 5, No.9, pp. 73-79.
Future research in this area could extend prior work Applegate, L.M., Austin, R.D., and McFarlan, F.W.
on strong situations and individual differences that (2003). Corporate Information Strategy and
matter (Bowen & Ostroff, 2004; Mischel, 1977; 1997). Management, 6th edition, Boston: McGraw-Hill.
Banks, M.H. and Henry, P. (1993). “Change and
In conclusion, this study makes substantive and
Stability in Employment Commitment,” Journal of
methodological contributions to the interactionist
Occupational & Organizational Psychology,
perspective on understanding turnover intentions of IT
Vol.66, No.2, pp. 177-184.
professionals, particularly new entrants who have
Barrick, M.R., Mount, M.K., and Gupta, R. (2003).
been neglected in prior research on turnover
“Meta-analysis of the Relationship Between the
intentions. We have argued that risk and variety are
Five-factor Model of Personality and Holland’s
parsimonious, important situational and individual
Occupational Types,” Personnel Psychology,
characteristics predicting turnover intentions among
Vol.56, pp. 45-74.
these new entrants to the high-tech profession and
Becker, G. (1983). Human Capital: A Theoretical and
found that these two characteristics have strong
Empirical Analysis with Special Reference to
explanatory power in a simulated environment. In
Education, Chicago: University of Chicago Press.
addition, our experimental scenarios provide an
Berloffa, G. and Simmons, P. (2003).
example of a methodological approach for examining
“Unemployment Risk, Labor Force Participation,
individual-situational interactions across a broad
and Savings,” Review of Economic Studies,
range of organizational settings. Finally, our scales for
Vol.70, pp. 521-539.
the fit characteristics of risk and variety provide a
Bharadwaj, A.S. (2000). “A Resource-Based
measurement foundation upon which future research
Perspective on Information Technology
can build.
Capability and Firm Performance: An Empirical
Investigation,” MIS Quarterly, Vol.24, No.1, pp.
References 169-196.
Bowen, D.E. and Ostroff, C. (2004). “Understanding
Agarwal, R., De, P., and Ferratt, T.W. (2001). “How HRM-firm Performance Linkages: The Role of
Long Will They Stay? Predicting an IT the “Strength” of the HRM System,” Academy of
Professional’s Preferred Employment Duration,” Management Review, Vol.29, No.2, pp. 203-
Proceedings of the 2001 ACM Special Interest 221.
Group on Computer Personnel Research Brandel, M. (2006). “Computerworld Special Report:
Conference (SIG CPR), San Diego, CA, pp. 2006 100 Best Places to Work in IT,” online
132-138. version
Agarwal, R., De, P., and Ferratt, T.W. (2002). http://www.computerworks.com/specila_report/0
“Explaining an IT Professional’s Preferred 09/000/000/special_report_009000080_primary_
Employment Duration: Empirical Tests of a article.jsp#toc
Causal Model of Antecedents,” Proceedings of Brancheau, J.C., Janz, B.D., and Wetherbe, J.C.
the 2002 ACM Special Interest Group on (1996). “Key Issues in Information Systems
Computer Personnel Research Conference (SIG Management: 1994-1995 SIM Delphi Results,”
CPR), Kristiansand, Norway, pp. 14-24. MIS Quarterly, Vol.20, No.2, pp. 225-242.
Agarwal, R. and Ferratt, T.W. (1999). “Coping with Bretz, R.D., Jr. and Judge, T.A. (1994). “The Role of
Labor Scarcity in Information Technology: Human Resource Systems in Job Applicant
Strategies and Practices for Effective Recruitment Decision Processes,” Journal of Management,
and Retention, Cincinnati, OH: Pinnaflex. Vol.20, No.3, pp. 531-551.
Agarwal, R. and Ferratt, T.W. (2000). “Retention and Cable, D.M. and Judge, T.A. (1996). Person-
the Career Motives of IT Professionals,” organization Fit, Job Choice Decisions and
Proceedings of the 2000 ACM Special Interest Organizational Entry,” Organizational Behavior
Group on Computer Personnel Research

The DATA BASE for Advances in Information Systems 22 Volume 38, Number 1, February 2007
and Human Decision Processes, Vol.67, pp. 294- Stages,” The DATA BASE for Advances in
311. Information Systems, Vol.34, No.4, pp. 64-72.
Cher, K.S. (2000). “Fatal Assumptions of Dotcoms,” Hsu, M.K., Jiang, J.J., Klein, and G., Tang, Z.
Business Times (Singapore), April 21. (2003). ’’Perceived Career Incentives and Intent
Cougar, J.D., and Zawacki, R.A. (1980). Motivating to Leave” Information and Management, Vol.B,
and Managing Computer Personnel, New York: No.5, pp. 361-169.
Wiley. Hom, P.W., and Griffeth, R.W. (1995). Employee
Dawis, R.V. (1991). “Vocational Interests, Values, and Turnover, Cincinnati: South-Western.
Preferences,” in Dunnette, M.D., and Hough, L.M. Igbaria, M. and Greenhaus, J.H. (1992).
(Eds.), Handbook of Industrial and Organizational “Determinants of MIS Employees’ Turnover
Psychology, Palo Alto, CA: Consulting Intentions: A Structural Equation Model,”
Psychologists Press, pp. 883-871. Communications of the ACM, Vol.35, No.2, pp.
Dulebohn, J.H. (2002). “An Investigation of the 35-49.
Determinants of Investment Risk Behavior in Igbaria, M., Greenhaus, J., and Parasuraman, S.
Employer-sponsored Retirement Plans,” Journal (1991). “Career Orientations of MIS Employees:
of Management, Vol.28, No.1, pp. 3-26. An Empirical Analysis,” MIS Quarterly, Vol.15,
Feldman, D.C. and Weitz, B.A. (1991). “From the No.2, pp. 151-169.
Invisible Hand to the Gladhand: Understanding a Inman, J. (2001). “The Role of Sensory-Specific
Careerist Orientation to Work,” Human Resource Satiety in Attribute-Level Variety Seeking,”
Management, Vol.30, No.2, pp. 237-257. Journal of Consumer Research, Vol.28, No.1,
Fridhandler, B.M. (1986). “Conceptual Note on State, Vol.105-120.
Trait, and the State-trait Distinction,” Journal of Jiang, J.J. and Klein, G.A. (2002). ‘‘A Discrepancy
Personality & Social Psychology, Vol.50, No.1, Model of Information System Personnel
pp. 169-174. Turnover,” Journal of Management Information
Gerhart, B. (1987). “How Important are Dispositional Systems, Vol.19, No.2, pp. 249-262.
Factors as Determinants of Job Satisfaction? Judge, T.A. and Bretz, R.D., Jr. (1992). “Effects of
Implications for Job Design and Other Personnel Work Values on Job Choice Decisions,” Journal
Programs,” Journal of Applied Psychology, of Applied Psychology, Vol.77, No.3, pp. 261-271.
Vol.72, No.3, pp. 366-373. Judge, T.A. and Cable, D.M. (1997). “Applicant
Glynn, M.A. (1998). “Situational and Dispositional Personality, Organizational Culture, and
Determinants of Managers’ Satisfaction,” Journal Organizational Attraction,” Personnel Psychology,
of Business and Psychology, Vol.13, No.2, pp. Vol.50, pp. 359-394.
193-209. Kahneman, D., Slovic, P., and Tversky, A. (1982).
Griffeth, R.W., Hom, P.W., and Gaertner, S. (2000). Judgment Under Uncertainty: Heuristics and
“A Meta-Analysis of Antecedents and Correlates Biases, Cambridge, UK: Cambridge University
of Employee Turnover: Update, Moderator Tests, Press.
and Research Implications for the Next Kristof, A.L. (1996). “Person-organization Fit: An
Millennium,” Journal of Management, Vol.26, Integrative Review of Its Conceptualizations,
No.3, pp. 463-488. Measurement, and Implications,” Personnel
Griffin, M. (2001). “Dispositions and Work Reactions: Psychology, Vol.49, pp. 1-49.
A Multilevel Approach,” Journal of Applied Kristof-Brown, A.L. (2000). “Perceived Applicant Fit:
Psychology; Vol.86, No.6, pp. 1142-1151. Distinguishing Between Recruiters’ Perceptions
Hamblen, M. (2002). “Offer Access to Hot of Person-job and Person-organization Fit,”
Technology,” Computerworld, Vol.36, No.19, pp. Personnel Psychology, Vol.53, pp. 643-671.
28-29. Kristof-Brown, A.L., Zimmerman, R.D., and Johnson,
Holland, J.L. (1985). Manual for the Self-directed E.C. (2005). “Consequences of Individuals’ Fit at
Search, Odessa, FL: Psychological Assessment Work: A Meta-analysis of Person-job, Person-
Resources. organization, Person-group, and Person-
Holton, E.F., III and Russel, C.J. (1999). supervisor Fit,” Personnel Psychology, Vol.58,
“Organizational Entry and Exit: An Exploratory pp. 281-342.
Longitudinal Examination of Early Careers,” Lazear, E.P. (1995). “A Jobs-Based Analysis of Labor
Human Performance, Vol.12, No.3/4, pp. 311- Markets,” American Economic Review, Vol.85,
341. No.2, pp. 260-265.
Hsu, M.K., Chen, H.G., Jiang, J.J., and Klein G. Luftman, J. (2005). “Key Issues for IT Executives
(2003). “Career Satisfaction for Managerial and 2004,” MIS Quarterly Executive , Vol.4, No.2, pp.
Technical Anchored IS Personnel in Later Career 269-285.

The DATA BASE for Advances in Information Systems 23 Volume 38, Number 1, February 2007
Luftman, J. and McLean, E.R. (2004). “Key Issues for Saxenian, A. (1994). “Lessons from Silicon Valley,”
IT Executives,” MIS Quarterly Executive, Vol.3, Technology Review, Vol.97, No.5, pp. 42-51.
No.2, pp. 89-104. Schneider, B. (1987). “The People Make the Place.
March, J.G., and Simon, H.A. (1958). Organizations, Personnel Psychology, Vol.40, pp. 437-453.
New York: Wiley. Schneider, B., Goldstein, H.W., and Smith, D.B.
McCulloch, C.E. and Searle, S.R. (2001). (1995). „The ASA Framework: An Update,”
Generalized, Linear, and Mixed Models, New Personnel Psychology, Vol.48, pp. 747-773.
York: John Wiley & Sons. Sheridan, J.E. (1992). “Organizational Culture and
Meyer, J.P., Stanley, D.J., Herscovitch, L., and Employee Retention,” Academy of Management
Topolnytsky, L. (2002). “Affective, Continuance, Journal, Vol.35, No.5, pp. 1036-1056.
and Normative Commitment to the Organization: Slaughter, S., and Ang, S. (2002). “Internal Labor
A Meta-analysis of Antecedents, Correlates, and Market Strategies and Turnover of Information
Consequences,” Journal of Vocational Behavior, Technology Professionals,” Proceedings of the
Vol.61, No.1, pp. 20-52. 2002 ACM SIGCPR Conference, M.
Miller, K.D. and Reuer, J.J. (1996). „Measuring Mandviwalla (Ed.), Kristiansand, Norway, pp.
Organizational Downside Risk,” Strategic 139-144.
Management Journal, Vol.17, No.9, pp. 671-691. Staw, B.M. and Ross, J. (1985). “Stability in the Midst
Mischel, W. (1997). “Personality Dispositions of Change: A Dispositional Approach to Job
Revisited and Revised: A View after Three Attitudes,” Journal of Applied Psychology, Vol.70,
Decades,” in Hogan, R., Johnson, J., and No.3, pp. 469-480.
Briggs, S. (Eds.), Handbook of Personality Steel, R.P. and Rentsch, J.R. (1997). “The
Psychology, New York: Academic Press, pp. Dispositional Model of Job Attitudes Revisited:
113-132. Findings of a 10-year Study,” Journal of Applied
Mischel, W. (1977). “The Iteraction of Person and Psychology, Vol.82, No.6, pp. 873-879.
Situation,” in Magnusson, D., and Endler, N.S. Thatcher, J.B., Stepina, L.P., and Boyle, R.J. (2002).
(Eds.), Personality at the crossroads: Current “Turnover of Information Technology Workers:
issues in interactional psychology, Hillsdale, NJ: Examining Empirically the Influence of Attitudes,
Lawrence Erlbaum Associates, Inc., pp. 333-352. Job Characteristics, and External Markets,”
Mobley, W.H., Griffeth, R.W., Hand, H.H., and Journal of Management Information Systems,
Meglino, B.M. (1979). “Review and Conceptual Vol.19, No.3, pp. 231-260.
Analysis of the Employee Turnover Process,” Tice, D. (2000). “Irrational Exuberance,” On Wall
Psychological Bulletin, Vol.86, pp. 493–522. Street, Vol.10, No.1, pp. 52-53.
Newton, T. and Keenan, T. (1991). Further Analyses Topel, R.H. and Ward, M.P. (1992). “Job Mobility and
of the Dispositional Argument in Organizational the Careers of Young Men,” Quarterly Journal of
Behavior.,” Journal of Applied Psychology, Economics, Vol.107, No.2, pp. 439-79.
Vol.76, No.6, pp. 781-787. Trivedi, M. (1999). “Using Variety-seeking-based
O’Reilly, C.A. III, Chatman, J., and Caldwell, D.F. Segmentation to Study Promotional Response,”
(1991). “People and Organizational Culture: A Journal of the Academy of Marketing Science,
Profile Comparison Approach to Assessing Vol.27, No.1, pp. 37-49.
Person-organization Fit,” Academy of Turner, J.A. (Ed.) (2001). Compensation and
Management Journal, Vol.34, No.3, pp. 487-516. Employment Risk in the United States and
Phillips, J.M. (1998). “Effects of Realistic Job Canada, Kalamazoo, MI: W.E. Upjohn Institute
Previews on Multiple Organizational Outcomes: A for Employment Research.
Meta-analysis,” Academy of Management United States Department of Labor, Bureau of Labor
Journal, Vol.41, No.6, pp. 673-690. Statistics, “BLS Releases 2002-12 Employment
Ratner, R.K. and Kahn, B.E. (2002). “The Impact of Projections,”
Private Versus Public Consumption on Variety- http://stats.bls.gov/news.release/pdf/ecopro.pdf
seeking Behavior,” Journal of Consumer (accessed February 2004)
Research, Vol.29, No.2, pp. 246-257. Wade, M. and Hulland, J. (2004). The Resource-
Rousseau, D.M. (1990). “New Hire Perceptions of the based View and Information Systems Research:
Their Own and Their Employer’s Obligations: A Review, Extension, and Suggestions for Future
Study of Psychological Contracts,” Journal of Research, MIS Quarterly, Vol.28, No.1, pp. 107-
Organizational Behavior, Vol.11, pp. 389-400. 142.
Rousseau, D.M. (1995). Psychological Contracts in
Organizations: Understanding Written and
Unwritten Agreements, Thousand Oaks, CA:
Sage.

The DATA BASE for Advances in Information Systems 24 Volume 38, Number 1, February 2007
About the Authors SE here, he is an AE at ISR. He served as Chair of
the Special Interest Group on Computer Personnel
Ritu Agarwal is Professor and Robert H. Smith Research (SIGCPR) before its merger with ACM
Dean’s Chair of Information Systems at the Smith SIGMIS.
School of Business, University of Maryland. She is
also the Director of the Center for Health Information Prabuddha De (Ph.D., Carnegie-Mellon University)
and Decision Systems. Professor Agarwal is is Accenture Professor of IT at Purdue University.
currently a Senior Editor at Information Systems He has published in Management Science,
Research. She has published in MISQ, ISR, Operations Research, ISR, MISQ, and other
Management Science, IEEE Transactions, JMIS, journals. He serves or earlier served on the Editorial
and other journals. Boards of Management Science, ISR, JMIS, and
DSS, among others. He was Doctoral Consortium
Thomas W. Ferratt is Sherman-Standard Register Co-Chair for ICIS’2006, Program Chair for
Professor of MIS at the University of Dayton. His ICIS’1999, and Workshop Co-Chair for WITS’1994.
work appears in ISR, MISQ, JMIS, Academy of
Management Journal, and others. Besides being an

Appendix A: Organization Profile


Section 1. Assume you have just taken a job with an organization that has the following profile. Think about this
organization in comparison with the type of organization for which you would prefer to work. Assume that the
organization described in the profile below is similar to your preferred type of organization on all characteristics
not included in the profile described below. Each X represents where the organization that has just hired you lies
on a 7-point linear scale. The two extremes of each 7-point scale are described, more or less, by the words at the
opposite ends of the scale. Thus, each X is closer to the words most descriptive of the organization that just hired
you.
Profile
Organization value may vary widely over
Organization is financially stable |------X------|-------|-------|-------|-------|
time
Personal financial gains or losses vary
Income, benefits, and retirement are
|------X------|-------|-------|-------|-------| considerably with organization’s
guaranteed
performance
Future employment depends on
Job security is high |------X------|-------|-------|-------|-------|
organization’s performance
Work is guaranteed |------X------|-------|-------|-------|-------| Work load is driven by market demand
Financial and employment risks
Financial and employment risks
associated with this organization are |------X------|-------|-------|-------|-------|
associated with this organization are high
low
Work involves a few selected Work involves a wide variety of
|-------|-------|-------|-------|------X------|
technologies technologies
High skill levels are developed in a
|-------|-------|-------|-------|------X------| Many different skills are developed
limited number of areas
Opportunity for IT job movement
Opportunity to move among many
internally is limited by the size of |-------|-------|-------|-------|------X------|
different IT jobs internally is high
organization
IT experience is gained by working on IT experience is gained by working on a
|-------|-------|-------|-------|------X------|
similar projects over time variety of different projects over time
Organization is small |-------|-------|-------|-------|------X------| Organization is large

* This appendix shows a portion of the instrument used during the experiment.

The DATA BASE for Advances in Information Systems 25 Volume 38, Number 1, February 2007
Appendix B: Rotated Component Matrix for Items Measuring Individual Characteristicsa

Component
1 2 3 4
b
Turnover intention-quit 3 years-specific-reverse .896
Turnover intention-quit 1 year-specific-reverseb .851
b
Turnover intention-stay long time-specific .855
b
Turnover intention-years/yrs_until_retire-specific-7points .712
Preferred variety-move among IT jobs-baseline .855
Preferred variety-experience with IT-baseline .834
Preferred variety-develop many different skills-baseline .817
Preferred risk-risks low-baseline-reverse .877
Preferred risk-secure job-baseline-reverse .850
Preferred risk-stable, guaranteed work, income-baseline-reverse .788
Propensity to stay-long time-baseline .768
Propensity to stay-quit 3 years-baseline-reverse .686
Propensity to stay-quit 1 year-baseline-reverse .636
Propensity to stay-years/yrs_until_retire-baseline-7points .630

a
Extraction Method: Principal Component Analysis. Rotation Method: Varimax with Kaiser Normalization.
Loadings < .40 not shown. These results were obtained with 54 observations for the second experimental
situation. Substantially similar results were obtained for separate factor analysis of the observations on
turnover intention in each experimental situation and the baseline observations on the other measures.
b
Turnover intention items are scored in the same direction as the propensity to stay items for this analysis.

The DATA BASE for Advances in Information Systems 26 Volume 38, Number 1, February 2007
Appendix C: Estimates of Fixed Effects a

H1a Test
Parameter Estimate Std. Error df t Sig.
Intercept 5.709 .451 147.056 12.666 .000
[Situational Variety=2.00] .039 .567 179.455 .069 .945
[Situational Variety=6.00] 0(a) 0 . . .
[Situational Risk=2.00] -1.315 .139 179.695 -9.441 .000
[Situational Risk=6.00] b
0 0 . . .
Preferred Variety -.008 .082 138.459 -.096 .923
Propensity to stay -.225 .071 58.110 -3.157 .003
Preferred Variety*([Situational Variety=2.00]) .136 .102 179.392 1.325 .187
Preferred Variety*([Situational Variety=6.00]) b
0 0 . . .

H1b Test
Parameter Estimate Std. Error df t Sig.
Intercept 6.024 .596 104.638 10.104 .000
[Situational Variety=2.00] .041 .556 178.432 .073 .942
[Situational Variety=6.00] b
0 0 . . .
[Situational Risk=2.00] -2.066 .296 178.558 -6.970 .000
[Situational Risk=6.00] b
0 0 . . .
Preferred Variety -.003 .086 121.301 -.038 .970
Preferred Risk -.137 .086 128.555 -1.603 .111
Propensity to stay -.222 .073 57.061 -3.063 .003
Preferred Variety*([Situational Variety=2.00]) .135 .100 178.371 1.350 .179
Preferred Variety*([Situational Variety=6.00]) b
0 0 . . .
Preferred Risk*([Situational Risk=2.00]) .296 .104 178.396 2.853 .005

Notes:
a
Dependent variable: Turnover intention
b
This parameter is set to zero because it is redundant.

The DATA BASE for Advances in Information Systems 27 Volume 38, Number 1, February 2007
H2a Test
Parameter Estimate Std. Error df t Sig.
Intercept 5.962 .419 79.641 14.220 .000
[Situational Variety=2.00] .768 .137 179.653 5.611 .000
[Situational Variety=6.00] 0b 0 . . .
[Situational Risk=2.00] -2.065 .297 179.547 -6.953 .000
b
[Situational Risk=6.00] 0 0 . . .
Preferred Risk -.162 .081 143.189 -1.983 .049
Propensity to stay -.198 .068 57.964 -2.932 .005
Preferred Risk*([Situational Risk=2.00]) .296 .104 179.382 2.846 .005
b
Preferred Risk*([Situational Risk=6.00]) 0 0 . . .

H2b Test
Parameter Estimate Std. Error df t Sig.

Intercept 6.024 .596 104.638 10.104 .000


[Situational Variety=2.00] .041 .556 178.432 .073 .942
b
[Situational Variety=6.00] 0 0 . . .
[Situational Risk=2.00] -2.066 .296 178.558 -6.970 .000
[Situational Risk=6.00] 0b 0 . . .
Preferred Variety -.003 .086 121.301 -.038 .970
Preferred Risk -.137 .086 128.555 -1.603 .111
Propensity to stay -.222 .073 57.061 -3.063 .003
Preferred Variety*([Situational Variety=2.00]) .135 .100 178.371 1.350 .179
b
Preferred Variety*([Situational Variety=6.00]) 0 0 . . .
Preferred Risk*([Situational Risk=2.00]) .296 .104 178.396 2.853 .005
b
Preferred Risk*([Situational Risk=6.00]) 0 0 . . .

Notes:
a
Dependent variable: Turnover intention
b
This parameter is set to zero because it is redundant.

The DATA BASE for Advances in Information Systems 28 Volume 38, Number 1, February 2007

Вам также может понравиться