Beyond skill development: The effects of training and development on the attitudes and retention of employees

PDF

Download

Open document

Flip pages

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content BEYOND SKILL DEVELOPMENT:
THE EFFECTS OF TRAINING AND DEVELOPMENT ON
THE ATTITUDES AND RETENTION OF EMPLOYEES
by
George S. Benson
A Dissertation Presented to the
FACULTY OF THE GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
in Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(BUSINESS ADMINISTRATION)
\ May 2002
Copyright 2002 George S. Benson
UM I Number: DP22713
All rights reserved
INFORMATION TO ALL U SERS
The quality of this reproduction is d ep en d en t upon the quality of the copy subm itted.
In th e unlikely event that the author did not sen d a com plete m anuscript
and there a re missing pages, th ese will be noted. Also, if material had to be rem oved,
a note will indicate the deletion.
U M T
Dissertation PubJishing
UMI D P22713
Published by ProQ uest LLC (2014). Copyright in the Dissertation held by the Author.
Microform Edition © P roQ uest LLC.
All rights reserved. This work is protected against
unauthorized copying under Title 17, United S tates C ode
ProQ uest LLC.
789 E ast Eisenhow er Parkway
P.O. Box 1346
Ann Arbor, Ml 4 8 1 0 6 - 1346
UNIVERSITY OF SOUTHERN CALIFORNIA
The Graduate School
University Park
LOS ANGELES, CALIFORNIA 900894695
This dissertation , w ritte n b y
George Sinclair Benson
U nder th e direction o f Ai.s.. D issertation
C om m ittee, an d a p p ro ved b y a ll its m em bers,
has been p re sen ted to a n d a ccep ted b y The
G raduate School, in p a rtia l fu lfillm en t o f
requirem ents fo r th e degree o f
DOCTOR OF PHILOSOPHY
DISSER TA TION COM M ITTEE
ii
ACKNOWLEDGEMENTS
This dissertation would not have been possible without help from a number
of people. Most important has been the support of my wife, Jennifer. This
dissertation has occupied many hours that would have otherwise been spent with her.
I would also like to thank my parents and sisters for their constant encouragement.
This dissertation represents a great deal of time and effort from my committee
members: Ed Lawler, Sue Mohrman, David Finegold and Bill Maxwell. In
particular, Ed, Sue, and David have been excellent mentors and instrumental in
completing this dissertation. I owe each of them a great deal. Special thanks are
also due to the staff of the Center for Effective Organizations who helped in a
number of ways. They include: Alice Yee Mark, Beth Neilson, and Nora Osganian
for data processing; Arienne McCracken, and Dan Canning for proofreading; and
Annette Yakushi, Lydia Arakaki, and Patty Trinidad for project management. Alec
Levenson, Murat Alpaslan, and Cecily Cooper read multiple drafts and provided
insightful comments. Tom Lee and Terri Kang provided statistical consultation.
Finally, I would like to thank the individuals who made this dissertation possible by
sponsoring the research and providing the data. I benefited from the help of Jack
Leary, Susan Rennie, Raquel Marero, Gail Heiss, Edward Duffy, Sandra Martinez,
Jonathan Brink, Kathy Biederman, Nancy Coleman, Henry Mercolino, and many
others.
iii
TABLE OF CONTENTS
ACKNOWLEDGEMENTS ,.........................................................................ii
LIST OF TABLES........................................................ .........................................v
LIST OF FIGURES................ vi
ABSTRACT............................................................................................................ vii
I. THE EFFECTS OF TRAINING AND DEVELOPMENT ON THE
COMMITMENT AND RETENTION OF EMPLOYEES........................... 1
Literature and Previous Research...........................................................................4
Theory and Hypotheses.......................................................................................... 9
II. METHODS........................................................................................................ 30
Sample and Data Sources.....................................................................................31
Measures ......................................................................................................33
Analyses................................................................................................................49
III. DEVELOPMENT AND EMPLOYEE ATTITUDES................................ 57
Results................................................................................................................... 57
IV. DEVELOPMENT AND EMPLOYEE RETENTION............................... 72
Results...................................................................................................................74
V. DISCUSSION........................................................................................? ...........91
Theoretical Contribution......................................................................................97
Limitations of the Research............................................................................... 103
Questions for Future Research........................................................................... 107
Conclusion..........................................................................................................109
BIBLIOGRAPHY................................................................................................ 112
APPENDIX 1 .........................................................................................................124
APPENDIX 2 .........................................................................................................125
APPENDIX 3 .........................................................................................................126
APPENDIX 4 .........................................................................................................127
APPENDIX 5 .........................................................................................................130
APPENDIX 6 .........................................................................................................134
APPENDIX 7 .........................................................................................................137
APPENDIX 8 .........................................................................................................139
APPENDIX 9
APPENDIX 10
V
LIST OF TABLES
Table 1 Proposed Dimensions of a Development Construct.............................15
Table 2 Survey Response Rates by Unit............................................................33
Table 3 Distribution of Job Grades....................................................................37
Table 4 Employees who Received Merit Pay and Promotion.......................... 43
Table 5 Tumover Rates by Year for all Employees..........................................49
Table 6 Summary of Hypotheses and Analyses................................................ 50
Table 7 Regression Results for Organizational Commitment.......................... 61
Table 8 Regression Results for Intention to Turnover......................................62
Table 9 Hierarchical Regression Results for Intention to Turnover: Archival
Data...................................................................................................... 68
Table 10 Earning a Degree, Timing of Rewards and Intention to Turnover... 70
Table 11 Rates of Voluntary Turnover Across the Samples.............................. 73
Table 12 Cox Regression Results........................................................................80
Table 13 Timing of Promotion Relative to Degree and Quit Rates................... 84
Table 14 Timing of Merit Pay Relative to Degree and Quit Rates.................... 84
Table 15 Time-Dependent Cox Regression Results...........................................86
Table 16 Summary of Hypotheses and Findings................................................ 96
LIST OF FIGURES
Figure 1 Employee Ratings of Development Usefulness..................................59
Figure 2 Participation in Tuition-Reimbursement and Intention to Turnover. 65
Figure 3 Hazard Estimates for Earning a Degree vs. Not Earning a Degree ..76
Figure 4 Earning a Degree, Receiving Rewards, and Quitting the Company. 78
ABSTRACT
Existing literature on workplace training and development takes the view that
skill development is a benefit of employment and employees tend to respond with
positive attitudes and less turnover. Based on the turnover and human capital
economics literature, I propose that certain types of skill development also create
economic incentives that may lead to decreased organizational commitment and
contribute to turnover. Categorical, regression and survival analyses are used to test
twelve hypotheses regarding the effects of on-the-job training, company-based
classes and tuition-reimbursement eligible classes on organizational commitment,
intention to turnover, and actual turnover. The data include a combination of 979
survey responses and archival data on 12,000 current and former employees of a
high-technology manufacturing firm from 1996 to 2000. My dissertation finds that
the effects of training and development on the commitment and retention of
employees are moderated by the type of development activity and whether or not
employees are rewarded for developing new skills.
1
I. THE EFFECTS OF TRAINING AND DEVELOPMENT ON THE
COMMITMENT AND RETENTION OF EMPLOYEES
Employee training and development is a significant and growing part of work
organizations. Estimates of total expenditures on development range from $16
billion to more than $55 billion in direct costs (Frazis, Herz, & Horrigan, 1995; Van
Buren, 2001). Throughout the 1990’s there was a steady growth in training budgets,
the percentage of employees receiving training, and the types of training offered in
companies (Lawler, Mohrman, & Benson, 2001; Van Buren, 2001). In 1995 the
Bureau of Labor Statistics (BLS) estimated that 93% of establishments with over 50
employees actively trained employees and that 70% of all employees in the U.S. had
experienced some sort of company-sponsored development within the last 12 months
(Frazis, Herz, & Horrigan, 1995).
Training and development (T&D) serves many functions. It is most
importantly the means by which employees learn their jobs. Many studies have
shown that training increases individual job performance when conditions are
appropriate for the transfer of learned skills to the job (Bartel, 1989; Rouiller, &
Goldstein, 1993; Tracey, Tannenbaum, & Kavanagh, 1995). Multi-company
research has also shown that employee development offered as part of a package of
supportive human resource management practices benefits the productivity and
financial performance of firms (Cappelli, 2000; Gardner, Moynihan, Park, & Wright,
2000; Huselid, 1995; Koch, & McGrath, 1996; Lawler, Mohrman, & Benson, 2001).
The management of individual skills is an important aspect of doing business today,
and employee development will likely grow in the future (Noe, 1999).
The benefits of employee development extend beyond the actual skills gained
and their contribution to an individual’s productivity. There is a growing
appreciation that skill development is an important part of a human resource strategy
for companies that require highly skilled and committed employees (Lepak & Snell,
1999; Tsui, Pearce, Porter, & Tripoli, 1997). Development of individual skills is
now seen as an integral part of the employment relationship (Rousseau, 1995) and
the package of benefits that employers are able to offer (Meyer, & Smith, 2000;
Nordhaug, 1989). Companies are also recognizing that development affects how
employees feel about their employer as well as their decisions to remain with the
company (Cappelli, 2000; “Recruiting,” 1999).
Researchers have long argued that the importance of training and
development extends beyond the skills and knowledge needed to carry out one’s job
(Goldstein, 1989; Tannenbaum, Mathieu, Salas, & Canon-Bowers, 1991). However,
there are several unresolved theoretical and empirical questions surrounding the
relationship between employee development, organizational commitment and
3
retention. Models of training evaluation do not include turnover, and have
historically overlooked affective outcomes unless they are the explicit goal of a
training program (Kraiger, Ford, & Salas, 1993). Moreover, theories from
management, psychology and labor economics yield different and sometimes
contrary predictions about the impact of development experiences on the employee’s
attitudes and willingness to remain with the company.
Management and psychology studies generally assume that skill development
and the opportunity to take training are viewed by employees as benefits, and as
signals that they are valued by their employers (Birdi, Allen, & Warr, 1997; Gaetner,
& Nollen, 1989; Noe, Wilk, Mullen, & Wanek, 1997; Nordhaug, 1989; Tsui et al.,
1997). Perceptions of company support in turn contribute to an employee’s
organizational commitment and willingness to remain with the firm (Farrell, &
Rusbult, 1981; Meyer, & Smith, 2000). Labor economics takes a very different view
and predicts that workers with upgraded general skills are likely to turnover unless
their wages are increased (Becker, 1965). The turnover literature also suggests that
development activities may have negative effects on attitudes and retention by
increasing an employee’s perceptions of external job opportunities (Gerhart, 1990;
Griffeth, Horn, & Gaertner, 2000).
This dissertation integrates these different perspectives on the effects of
employee development on organizational commitment and retention by arguing that
development may have either positive or negative effects depending on the nature of
4
the development activity and what happens to employees when they gain new skills.
The effects of development are hypothesized to differ based on the level of benefit
that they afford, the types of skills developed, and whether or not it is rewarded by
the organization. Twelve hypotheses are developed and tested using a combination
of survey and archival data from a high-tech manufacturing company. The results
should both contribute to existing theory and research in employee development, and
aid individuals and employers in getting the most out of their large mutual
investment in skill development activities.
Literature and Previous Research
There are two important sets of literature that address the potential effects of
employee training and development on employee commitment and retention. Theory
and research from management and psychology include a few studies of the effects
of employee development on organizational commitment. However, turnover is not
generally included in these models of the outcomes of development, and the research
addressing this question is limited to a few qualitative case studies (“Recruiting,”
1999). Research on this relationship comes primarily from labor economics, which
is the second major set of literature that deals with employee development.
In the management literature, the outcomes of training and development
have historically been evaluated in terms of skill acquisition and performance. There
has been awareness for some time, however, that participation in employee
development has other potential effects. Kirkpatrick (1976) proposed that training
be evaluated at four levels: by trainees’ reactions, learning, behavior change, and
then subsequent organizational results. Since then many others have argued that
changes in the attitudes of employees, such as organizational commitment and job
satisfaction, should also be examined (Feldman 1989; Noe et al., 1997; Tannenbaum
et al., 1991). Only recently have these other potential attitudinal outcomes been
incorporated into empirical studies of training and development (Birdi et al., 1997).
In predicting the effects of development on organizational commitment,
contemporary theories of employee development from management and psychology
are dominated by the assumption that development is perceived as a benefit by
employees who then respond positively with increased commitment to the
organization (Birdi et al., 1997; Noe et al., 1997). This predicted relationship
between the provision of development and organizational commitment is based on a
social exchange model. Blau (1964) defined social exchange as cooperation between
two or more individuals for mutual benefit. Such cooperation is governed by a norm
of reciprocity that holds that an individual will seek to respond in kind when he or
she receives a benefit from another individual or an organization (Gouldner, 1960).
Employee development research based on the assumption of social exchange
has taken two forms in the literature. First, studies have examined whether the
provision of development opportunities as part of a larger set of HR practices is
6
related to commitment (Meyer, & Smith, 2000; Tsui et al., 1997). Although the
findings suggest that development is likely to be positively related to employee
attitudes, this conclusion is tempered by the fact that employee development is
measured and analyzed in conjunction with other HR practices. For example, Meyer
and Smith (2000) found “employee friendly” HR practices including training to be
positively related to commitment. Tsui et al. (1997) used a similar methodology and
measured skill training as one of four practices included in a scale of “employer
investment.” Tsui et al. (1997) investigated the balance of contributions and
inducements in the employment relationship and found that development is an
important “inducement” offered to employees, which has the potential to impact
commitment.
Second, there have been a small number of studies that have measured actual
participation in different types of development activities (Birdi et al., 1997; Gaertner,
& Nollen, 1989). In the study most relevant to the present research, Birdi et al.
(1997) analyzed training time spent in four categories of development. The authors
found that job satisfaction and organizational commitment were both positively
correlated with prior participation in work-based development and required training
courses. Classes taking place on the employee’s own time were unrelated to these
same outcomes. The findings of Birdi et al. (1997) support the social exchange
model of development in which participation (at least during paid work hours) leads
to increased organizational commitment.
Research in labor economics and human capital theory presents a different
picture of the likely effects of development on employee retention. The reason is
that the labor economics literature assumes that training and skill development are
elements of economic exchange in organizations. That is, an individual’s skills are
valuable human capital that he or she offers to employers in return for wages. Based
on the work of Becker (1965) and others, human capital theory predicts that training
which serves to increase the skills of employees makes them more likely to be bid
away by other firms if their wages are not increased to compensate for this increase
in skills. Becker (1965) defined those skills that were useful to other firms as
“general” skills.
Because providing training and development is a direct cost to the employer,
Becker (1965) concluded that a firm would only provide general skills training if an
employee were willing to bear the cost of that training by accepting a lower initial
wage. From the firm’s perspective, providing employees with skills is an investment
similar to physical capital, but human capital is owned by the employee and is
therefore mobile. Employees are free to seek out the best return for their skills
regardless of whether the skills were developed through company-sponsored
activities. This means that companies risk losing their investment in development
when an employee leaves for a better job.
General empirical support has been found for a number of Becker’s (1965)
human capital predictions, including the correlation between training and wage
8
growth (Bartel, 1995; Lynch, 1992; Lynch, & Black, 1998). However, the evidence
contradicts a central contention of human capital: that companies will not pay for
general skills training to their employees. Several studies have found that employers
are devoting substantial resources to building skills that are useful to other firms
without lowering trainee wages (Bishop, 1997; Lengermann, 1996; Loewenstein &
Spletzer, 1998). This raises the questions of why firms make this investment despite
the turnover risk and of whether providing such development actually leads to
increased employee turnover.
Human capital research has mixed results regarding the relationship between
employee training and turnover depending on the type of development activity
provided. Lynch (1991) used a national longitudinal survey to investigate the
relationship between participation in different types of training and the probability of
employee turnover among young women. She found that employees who had
experienced formal on-the-job training were much less likely to leave their
employer, while those who participated in some form of company-sponsored training
outside of the workplace were more likely to leave. Loewenstein and Spletzer
(1997) used U.S. census data to reach the similar conclusion that individuals who
participated in company training are less likely to leave their jobs. They found no
differences, however, for employees who participated in school-based training
outside of work.
Taken together these two sets of literature from management and labor
economics suggest two important conclusions regarding the outcomes of employee
development on commitment and retention. First, it is clear that not all development
activities are the same, and activities such as company training, on-the-job training,
and tuition-reimbursement classes need to be considered independently. Studies
from both management (Gaertner, & Nollen, 1989) and labor economics (Krueger, &
Rouse, 1998) that do not distinguish between different types of development
activities together have not found significant relationships to either commitment or
turnover. Second, there is a strong need for theory that integrates the different
assumptions regarding employee development as an element of either social or
economic exchange. Theory and research from management and psychology find
that providing development opportunities to employees serves to increase their
organizational commitment (and therefore retention). Labor economics argues the
opposite — that providing employees with new skills increases the likelihood that
they will leave the firm. These differences in the existing literature must be
reconciled in order to gain a comprehensive picture of the likely effects of
development on commitment and retention of employees.
Theory and Hypotheses
To integrate these two important literatures into a new conceptual framework,
twelve hypotheses are proposed that predict either positive or negative effects of
employee development according to the level of benefit provided by the activity, the
10
type of skills gained by the employee, and whether or not the employee is rewarded
after participating. Hypotheses are developed for the three most common forms of
formal company-sponsored development activities: company training classes, formal
on-the-job training, and tuition-reimbursement programs. These three training
activities were selected because they represent the main development activities
sponsored by companies and are widely accepted and used by researchers and
practitioners alike to categorize the different company-sponsored development
activities undertaken by firms (Birdi et al. 1997; Loewenstein, & Spletzer, 1997;
“Recruiting,” 1999). Each of these activities is examined separately for its effects on
organizational commitment and retention of employees.
Outcomes Beyond Skill Development
The relationships between employee development, commitment and retention
are growing in importance as companies seek new ways to attract and retain
employees. Employee development holds great promise if it can be used effectively
to create organizational commitment and motivation to remain with the company
(Cappelli, 2000; “Recruiting,” 1999). The potential for employee development to
affect these outcomes, however, should be considered in terms of existing theory on
organizational commitment and employee retention.
Organizational commitment. Commitment is the strength of an individual’s
identification with and attachment to an organization (Mowday, Porter, & Steers,
1982). It has been defined as “(a) a strong belief in and acceptance of the
11
organization’s goals and values, (b) a willingness to exert considerable effort on
behalf of the organization, and (c) a desire to maintain organizational membership”
(Porter, Steers, Mowday, & Boulian, 1974: 604). Since this definition was proposed,
others have found that an employee’s commitment to an organization can take
several forms (Mathieu, & Zajac, 1990; Meyer, & Allen, 1997; O’Reilly, &
Chatman, 1986). Allen and Meyer (1990) identified commitment as taking three
forms: affective, normative and continuance. Affective commitment reflects the
heart of the Porter et al.’s (1974) definition of attachment to and identification with
an organization. Normative commitment is a sense of obligation on the part of the
employee to maintain membership in the organization. Finally, continuance
commitment is based on the notion of “side bets” or the perceived losses associated
with discontinuing employment with the organization (Meyer, & Allen, 1984).
Previous studies of the effects of employee development, and of other HR
practices in general, have typically measured affective commitment. However, some
have suggested that HR and development might also affect normative and
continuance commitment (Meyer, & Allen, 1997). In fact, the development of skills
that are difficult to market outside the company was included in Allen and Meyer’s
(1990) original measure of continuance commitment. They argued that the
development of company-specific skills and knowledge is one of the “side bets” or
investments that employees make in their employers.
While employee development has been suggested as an antecedent of both
affective and continuance commitment, different outcomes are associated with these
forms of commitment. The consequences of high continuance commitment can
actually be negative (Allen, & Meyer, 1996; Meyer, & Allen, 1997). Continuance
commitment can make an employee feel trapped in an unsatisfying employment
situation and can lead to withdrawal and have a negative affect on attitudes (Meyer,
& Allen, 1997). Employers seeking the benefits of commitment generally focus on
affective commitment and only affective organizational commitment is included as
an outcome of interest in this dissertation. Affective organizational commitment
(hereafter referred to as “organizational commitment”) is important because of its
empirical linkages to work performance (Meyer, Paunonen, Gellatly, Goffin, &
Jackson, 1989), customer satisfaction (Benkhoff, 1997; Bowen, & Schneider, 1999;
Oakland, & Oakland, 1998), and behaviors including absenteeism and turnover
(Blau, 1989; Blau, & Boal 1987, Huselid, & Day, 1991), prosocial behavior
(O’Reilly, & Chatman, 1986) and extra-role innovativeness (Katz, & Kahn, 1978).
Employee retention. Employee retention is defined as the opposite of
voluntary turnover and has received increased attention from companies in the tight
labor markets of the 1990’s. Historically low unemployment rates created real
problems for companies in keeping their best employees. One way that companies
have been trying to cope with the problem is by increasing investments in employees
and in employee development (“Recruiting,” 1999). It has become common for
13
companies to offer generous employee development benefits intended to help retain
employees. However, the direct effect of skills development on turnover has
received little attention from the hundreds of published turnover studies (Griffith,
Horn, & Gaertner, 2000; Horn, Caranikas-Walker, Prussia, & Griffeth, 1992).
In addition to voluntary turnover behavior, this dissertation also examines
intention to turnover. Models of turnover treat intention to leave as the final
cognitive variable immediately preceding actual turnover (Mobley, Griffeth, Hand,
& Meglino, 1979). Intent to turnover as a theoretical construct in models of turnover
is drawn from arguments of reasoned action made by Fishbein and Ajzen (1975) and
others that intentions are a direct precursor of behaviors. Intention to turnover
reflects an individual’s direct motivation to stay or leave (Cotton, & Tuttle, 1986),
and has been shown in research to be the strongest single predictor of actual turnover
behavior (Griffeth, Horn, & Gaertner, 2000; Johnston, Griffeth, Burton, & Carson,
1993).
Models of turnover at the individual level generally include affective
variables such as job satisfaction or organizational commitment along with
demographic and organizational characteristics as the primary antecedents of
voluntary turnover (Bluedom, 1982; Mobley et al., 1979). These predictors then
interact with perceived external opportunities or ease of movement to predict
intentions to turnover and eventually turnover behavior (Gerhart, 1990; Trevor,
2001). The causal ordering of commitment, intention to turnover, and external
14
opportunities within general turnover models have been widely debated in the
literature (Blau, & Boal, 1987; Huselid, & Day, 1991; Somers, 1995; Tett, & Meyer,
1993). This literature finds that although organizational commitment and intention
to turnover are negatively correlated, they are distinct constructs. This dissertation is
not intended to address the causal ordering of organizational commitment, turnover
intentions, and turnover behavior but rather to examine the effects of employee
development on all three as potential outcomes.
From a human capital perspective, all three are critical because the individual
skills and expertise gained through employee development can provide a competitive
advantage to the company only if employees are committed to the goals of the
organization and are willing to remain (Ulrich, 1998). Human capital views
individual skills as a critical part of the firm-specific capabilities that are the primary
determinants of productivity and the competitiveness of firms (Strober, 1990).
Voluntary turnover and behavioral withdrawal in the form of decreased commitment
to the organization reduce the value of human capital and ultimately the productivity
of the firm. When an employee voluntarily leaves an organization, the firm not only
incurs the costs of recruiting and training a new employee to fill the vacancy created
by turnover, it also loses the opportunity to realize the return on the investments
made in the employee who has left the organization.
15
Predicting Commitment and Retention
Noe et al. (1997) proposed a number of dimensions of employee
development that they hypothesized to explain differences in outcomes, including
learning and work attitudes. Birdi et al. (1997) built on this work with a similar list
to define employee development in their study of a broad array of development
outcomes. The dimensions presented by each of these authors are presented in Table
1.
Table 1
Proposed Dimensions of a Development Construct
Noe, Wilk, Mullen, & Wanek (1997)
Voluntary vs. Involuntary
Formal vs. Informal
Current vs. Future-oriented
Incremental vs. Framebraking
Introspective vs. Interactive
Birdi, Allan, & Warr (1997)
Voluntary vs. Required
Formal vs. Informal
Current vs. Future-oriented
Job focus vs. Non-job focus
Work time vs. Non-work time
This dissertation is not intended to test a complete development construct, but
instead focuses on two dimensions drawn from this existing theory and labor
economics because they are particularly relevant to explaining outcomes in attitudes
and retention. These are whether the activity is a company-provided benefit, and the
nature of the skills gained through development. In the following sections, the
theoretical foundations of each of these dimensions are reviewed and hypotheses are
16
developed to describe how on-the-job training, company classes, and tuition-
reimbursement classes should lead to different effects on commitment and retention.
Finally, whether or not employees receive rewards after they gain new skills is
proposed to moderate the effects of general skills development.
Company-provided benefit As described previously, contemporary models
of development are based on the notion that employees perceive development
opportunities as benefits provided by the company. In this case the degree to which
participants perceive the development activity as a company-provided benefit is
theoretically the most important feature for explaining commitment and retention.
Past research has shown that people perceive the provision of training to be a benefit
offered by their employers in addition to pay and along with other fringe benefits
(Nordhaug, 1989). Birdi et al. (1997) operationalized this benefit in terms of the
amount of personal time that an employee has to expend to participate in the
development activity. They found that development taking place during work time
(as opposed to the employee’s own time) is likely to lead to increased satisfaction
and organizational commitment. Company classes and on-the-job training both take
place completely on company time and are therefore likely to be seen as benefits
provided by the company. Participation in each of these activities is hypothesized to
be positively related to organizational commitment:
H I: Participation in company training classes is positively
related to organizational commitment.
17
H2: Participation in formal on-the-job training is positively
related to organizational commitment.
If participation in company training and on-the-job training is positively
related to commitment, then models of employee turnover predict that participation
may also affect turnover and turnover intentions (Lee, & Mowday, 1987). Research
on voluntary turnover finds that commitment, along with other positive employee
attitudes, influences turnover (Guzzo, & Noonan, 1994; Mathieu, & Zajac, 1990;
Mowday, Porter, & Steers, 1982). Employees with low commitment are most likely
to withdraw (Bluedom, 1982) and eventually leave an organization (Johnston et al.,
1993). Employees with high organizational commitment, on the other hand, turn
over less often (Cohen, 1993). Participation in on-the-job training and company
classes are hypothesized to increase organizational commitment through social
exchange. This implies that participation in these two development activities might
also lead to reduced intention to turnover and actual turnover.
H3: Participation in company training classes is negatively
related to intention to turnover.
H4: Participation in formal on-the-job training is negatively
related to intention to turnover.
H5: Participation in company training classes is negatively
related to voluntary turnover.
H6: Participation in formal on-the-job training is negatively
related to voluntary turnover.
18
While predicting the social exchange effects for these two types of employee
development is relatively straightforward, the case of tuition-reimbursement is much
less clear. Birdi et al. (1997) were unable to link tuition-reimbursement classes to
commitment and satisfaction and suggest that employees perceive tuition-
reimbursement as less of a benefit because it takes place on the employee’s own time
and therefore they feel less obligation to reciprocate. Even in the case where a
company gives some time off for classes, taking classes and earning a degree
through tuition-reimbursement often requires a considerable amount of the
employee’s own time.
However, a pure social exchange model of development would still predict
some positive effects of tuition-reimbursement on commitment and retention. The
reason is that tuition-reimbursement is still clearly a benefit offered by employers
even though it requires more time from the employee than company training classes
or on-the-job training. Companies pay millions of dollars in tuition every year, and
tuition-reimbursement is even more likely to be viewed as a benefit in companies,
such as the one under investigation, which pay full tuition costs and offer time off for
classes. Differences in the relationships found between the types of development
may instead be influenced by the skills gained rather than the expenditure of personal
time and resources.
General versus specific skills. The second major characteristic of
development that affects the outcomes of participation is the nature of the skills
19
developed. Birdi et al. (1997) describe the differences in the types of skills gained
through development in terms of both “job focus to non-job focus” and “current to
future orientation.” These dimensions are useful for describing the learning
outcomes of development, but are less effective for capturing how training and
development affects commitment and retention. The characteristic of skills gained
through development that is most important for predicting these outcomes is whether
the skills increase the external job opportunities of the employee. This aspect of
employee development is best described by labor economics, which contrasts firm-
specific” skills with “general” skills.
Becker (1965) first defined “general” skills as skills that increase the
marginal productivity of labor at all firms equally, while “specific” training increases
the productivity of labor at a single firm. This definition has been slightly modified
since these original categories were defined as absolutes that seldom occur in
practice. As Loewenstein and Spletzer (1999, 730) note, “One is hard pressed to
come up with good examples of training that provides skills that are useful at only
one employer.” Stevens (1996) suggested that “transferable” training be used to
describe the development of skills that were in between the categories of general and
specific. Transferable training is useful at some, but not all, firms.
The general versus specific or transferable nature of skills depends on how
the skills are perceived by other potential employers rather than whether or not the
skills will actually raise productivity at a different firm. For training to be general,
20
alternative employers must perceive that an employee’s skills gained through
development would in fact raise his or her productivity at their firms. Asymmetric
information between employees and other perspective employers on the value of the
skills gained through development has also been offered as one explanation of how
employers can offer general skills training to employees without sharing the cost of
training through reduced wages (Acemoglu, & Pischke, 1999; Bishop, 1997;
Hashimoto, 1981).
This information asymmetry means that the market for transferable skills is
likely to be imperfect because the value of the skills is not known definitively by
other potential employers until they are actually hired. As a result, an employee may
work for less than they are actually worth after participating in a development
program. This situation also creates the potential for employees to be bid away or
“poached” by other companies who then receive the return on the investment in
employee development (Stevens, 1996). Research shows that development of
transferable skills increases the likelihood that the employee will take a better job
elsewhere (Lynch, 1991; Loewenstein, & Spletzer, 1997).
This research also shows that tuition-reimbursement, company classes, and
on-the-job training are significantly different in terms of both the types of skills that
they impart and how prospective employers perceive them. Loewenstein and
Spletzer (1999) use individual survey data to conclude that company training is the
most specific type of training, tuition-reimbursement is the most general, and
21
company-sponsored seminars fall somewhere in between. The differences between
company training and tuition-reimbursement are confirmed by research on the wage
effects of different types of training. Loewenstein and Spletzer (1997) found that
general skills development raised the future wages for workers who switch
employers more than for workers who remain with the employer that initially
provided the training. Using similar data, Lynch (1991) concluded that prospective
employers view formal company training as highly firm-specific and training
received from schools through tuition-reimbursement as significantly more general
in nature. Based on these findings, some studies within the human capital literature
have gone so far as to use “on-the-job” versus “off-the-job training” as proxies for
general versus specific development (Lynch, 1991; 1992).
The general nature of tuition-reimbursement is also supported by economic
theories of “signaling” which suggest that firms use formal education in hiring
decisions as a sign of the future productivity of the applicant (Spence, 1974; Strober,
1990). The quality of the development experience is more easily judged when the
employee attends a school or university rather than a work-based development
program. For this reason, participation in a tuition-reimbursement program is likely
to provide more skills that are more valuable to other firms than on-the-job training
or company programs. In the case that an employee earned a formal degree or
certificate through tuition-reimbursement, signaling theory suggests that the skills
would be perceived as even more general or transferable. A completed degree
22
provides easily comparable information about the nature of the employee’s skills and
potential contribution to a potential employer. From the economic perspective,
earning a degree would further increase the likelihood that an employee will quit the
company.
The general versus specific nature of development is also relevant to models
of commitment and turnover from management, because general skills development
may influence the perceptions of employees concerning their job alternatives. The
availability of different jobs and the perception of external opportunities have been
an important part of the turnover literature since March and Simon (1958) noted that
“ease of movement” would influence an individual’s decision to leave an
organization. Research has shown that increases in an individual’s perceptions of
external career opportunities are negatively related to affective commitment
(Bateman, & Strasser, 1984; O’Reilly, & Caldwell, 1981), positively correlated with
intention to turnover (Gerhart, 1990; Kirschenbaum, & Nano-Negrin, 1999), and
positively related to actual turnover (Griffeth, Horn, & Gaertner, 2000; Trevor,
2001).
Developing new skills that are in demand at other companies increases
employees’ perceptions that they can find another job easily. Findings from labor
economics and the turnover literature suggest that the economic incentives created
by increased job opportunities predominate over any potential positive social
exchange benefits of providing tuition-reimbursement to employees. General skills
23
development has been shown to be positively related to employee turnover (Lynch,
1991) and unrelated to organizational commitment (Birdi et al, 1997). For all these
reasons, participation in tuition-reimbursement classes is hypothesized to lead to
decreased organizational commitment, increased intention to turnover, and increases
voluntary turnover.
H7: Participation in tuition-reimbursement classes is
negatively related to organizational commitment.
H8: Participation in tuition-reimbursement classes is
positively related to intention to turnover.
H9: Participation in tuition-reimbursement classes is
positively related to voluntary turnover.
This research on attitudes and retention suggests that when employees gain
general skills that are marketable to other companies, they seek returns for new
skills. Leaving the company that provided those skills, however, is not the only way
to realize the benefits of new general skills. The worker may also remain with the
same company and be rewarded for their new skills. In this case the company bears
the full cost of providing the training, but at least it has the opportunity to gain some
of the benefits of providing the training by keeping the employee with the company.
If an employee is rewarded after participating in general skills development, the
external job opportunities no longer appear quite as lucrative and pressure to
turnover is reduced. For this reason, whether or not employees are rewarded for new
24
skills within their own firms is likely to moderate the hypothesized negative effects
of tuition-reimbursement on commitment and retention.
Rewards after gaining skills. Although the training and development
literatures from management and economics have different assumptions about the
likely effects of providing development to employees, each of these literatures
suggests that the outcomes depend, at least in part, on what happens to the employee
following the development experience. The management literature on training and
development provides ample evidence that employees participate in voluntary
development activities because they wish to benefit their careers (Fujita-Starck,
1996; Maurer, & Tarulli, 1994). This suggests that whether or not the employee
experiences subsequent career advancement should affect how the employee
responds in terms of commitment and willingness to remain with the company.
Instead of career advancement, human capital theory predicts employee turnover in
terms of the wage effects of increased general skills. For example, research finds
that employees who participate in general skills development increase their wages if
they switch employers (Loewenstein, & Spletzer, 1997; Lynch, 1991). Based on
either career advancement or wage increases, both literatures suggest that the
outcomes of employee development might depend on whether the employee is
rewarded following participation in general skills development.
From the employees’ perspectives, general skills development makes them
more marketable, and provides stepping-stones to better jobs and greater rewards
25
(Martin, Pate, & Beaumont, 2000). This dissertation takes a more general view that
wage increases and career advancement are both rewards offered to employees by
the firm for increased skills. From an exchange perspective they go together into a
total package of inducements that a firm offers in return for employees’ contributions
(March, & Simon, 1958). The present study operationalizes rewards as either
promotions or merit awards and does not attempt to compare the differences between
career advancement versus wage increases.
In the case where employees are not rewarded for new skills, equity theory
provides further theoretical support for predicting negative effects on commitment
and positive effects on turnover. Equity theory predicts that developing new general
skills is likely to impact the employee’s perceptions of his or her inputs compared to
a similar person (Adams, 1965). When a worker’s ratio of inputs to outputs diverges
from that of a comparison person, negative and uncomfortable feelings are aroused
that manifest themselves in the form of negative attitudes and behavioral withdrawal
from the organization (Greenberg, 1988). Schwarzwald, Koslowsky, and Shalit
(1992) found that failure to receive a promotion after a self-initiated candidacy for a
new job within an organization was associated with feelings of inequity, decreased
commitment, and increased absenteeism. Voluntarily participating in tuition-
reimbursement and developing general skills can be viewed in some respects as
similar to self-nomination for a promotion. Failure to receive rewards of any type in
26
this case is likely to result in feelings of inequity with corresponding negative effects
on organizational commitment and retention.
In the case that employees are rewarded following general skills
development, receiving rewards first alleviates the economic pressures and
perceptions of inequity that are hypothesized to have negative effects on
commitment and retention. This would allow the social exchange effects of receiving
tuition-reimbursement to exert positive influences on commitment and retention.
Second, developing general skills from company-sponsored activities and
subsequently receiving rewards from the company should have greater positive
effects on commitment (and therefore turnover) than if either occurred independently
or not at all. The reason is that developing employee skills and then rewarding
employees communicates to the employees that they have career opportunities and
are valued by their employer.
Both the literature on commitment and perceived organizational support find
that the relationship between HR practices such as employee development and
affective outcomes depends on employees’ assessments of how they are regarded by
the company (Eisenberger, Huntington, Hutchison, & Sowa, 1986; Meyer, & Smith,
2000). In other words, commitment depends more on the messages that
development policies convey rather than the practices themselves (Guzzo, &
Noonan, 1994; Meyer, & Smith, 2000). Studies have found that commitment
depends on whether or not the employee interprets the motivation behind HR
27
practices as genuine concern for employees’ well being and desire to attract and
retain good employees (Kinicki, Cason, & Bohlander, 1992; Koys, 1988).
Based on this research, rewards are likely to moderate the effects of
development on commitment and retention because they play a role in determining
the employees’ interpretation of the company’s employee development programs and
their own development experiences. When employees participate in tuition-
reimbursement and develop general skills, they are not likely to view it as a signal
that the company is genuinely interested in their careers unless the company rewards
them when they gain those skills. Research in perceived organizational support
shows that company actions foster more positive feelings of support in employees if
those actions are discretionary and directed at the individual employee (Eisenberger
et al., 1986; Shore, & Shore, 1995; Wayne, Shore, & Liden, 1997). When an
employee develops general skills and receives rewards, the interaction of the two
events fosters perceptions of support and interest from the company. For these
reasons, receiving rewards following participation in tuition-reimbursement is
hypothesized to moderate the relationship between general skills development,
commitment, and employee turnover:
H10: Receiving rewards following participation in tuition-reimbursed
courses moderates organizational commitment in such a way that, when an
employee receives a reward, participation in tuition-reimbursed courses is
positively related to organizational commitment and when an employee does
not receive subsequent rewards, participation in tuition-reimbursed courses is
negatively related to organizational commitment.
28
HI 1: Receiving rewards following participation in tuition-reimbursed
courses moderates intention to turnover in such a way that, when an
employee receives a reward, participation in tuition-reimbursed courses is
negatively related to intention to turnover and when an employee does not
receive subsequent rewards, participation in tuition-reimbursed courses is
positively related to intention to turnover.
H I2: Receiving rewards following participation in tuition-reimbursed
courses moderates voluntary turnover in such a way that, when an employee
receives a reward, participation in tuition-reimbursed courses is negatively
related to voluntary turnover and when an employee does not receive
subsequent rewards, participation in tuition-reimbursed courses is positively
related to voluntary turnover.
The preceding twelve hypotheses are intended to integrate the assumptions of
the social and economic exchange effects of employee development on commitment
and retention. See Appendix 1 for a summary of hypotheses. They predict that
company-provided classes, on-the-job training and tuition-reimbursement have
different effects on organizational commitment, intention to turnover and voluntary
turnover. These differences depend on whether the development is a company-
provided benefit, the nature of the skills developed, and whether the employee is
rewarded following participation.
Prior to now, management researchers have overlooked altogether the
economic effects of development. Labor economists, on the other hand, do not
account for social exchange and the potential positive effects on commitment and
retention in the case of developing general skills and receiving rewards. Using
organizational rewards to moderate the effects of general skills development makes it
29
possible for an activity such as tuition-reimbursement to have either positive or
negative effects on employee commitment and retention when compared to
employees who do not participate in tuition-reimbursement. Participation in tuition-
reimbursement creates external opportunities, economic incentives and feelings of
inequity that are likely to lead to decreased commitment and retention unless the
employee’s new skills are rewarded. Being rewarded for new general skills by the
company that provided the development not only reduces these external job
opportunities and feelings of inequity, but also has an additional positive effect of
being a strong communication that the employee is valued. Employees who
experience both of these events together (development and rewards) should therefore
be more committed and turn over less often than an employee who never participated
in development at all.
30
II. METHODS
The effects of employee development on organizational commitment,
intention to turnover and retention are tested using electronic personnel records and
survey data from the U.S.-based employees of a large high-technology
manufacturing firm. The company has approximately 10,000 full-time salaried
employees, with headquarters in the United States and repair facilities located around
the world. In 1996 the firm launched an effort to significantly increase development
opportunities for employees including extensive in-house training, rotational
programs and other structured on-the-job training. The cornerstone of the
company’s development efforts is an ambitious tuition-reimbursement program that
pays for all courses and gives a portion of class time-off to employees to pursue any
degree or professional development program, regardless of their current job or
business need. When employees complete any type of degree, they also receive a
significant one-time bonus of $10,000 in company stock.
This emphasis on development opportunities came in response to a number of
large layoffs by the company. Although the largest occurred in the early 1990’s, the
threat of significant workforce reductions has persisted over the last five years. In
31
1999, the company trimmed 11.8% of its salaried workforce following a decision to
close its second largest facility and relocate employees. With job security at the firm
clearly declining, managers found they needed to provide an additional incentive
beyond salary for talented engineers and business managers to join the company and
remain long-term. The company has since made a concerted effort to feature
educational and development opportunities as an incentive to attract and retain new
employees, as well as to inspire commitment among long-tenured employees.
Although workers readily agree that these development programs are a significant
benefit, important questions remain as to the actual effects of these policies on the
commitment and retention of employees.
Sample and Data Sources
The company provided electronic personnel records for all salaried
employees for the period January 1996 - June 2000. Records were provided for a
total of 14,040 full-time permanent employees. Employees who were rehired from a
previous layoff during the study period were excluded, leaving 12,360 current and
former employees available for analysis.1 These electronic work histories included
data on job grade, salary, pay raises, promotions, start date and termination date.
They were matched by employee identification number with electronic records on
1 A total of 1680 employees (11.7% of the employees active 1996-2000) were excluded because the
attitudes and voluntary turnover behavior of these employees is likely to be affected by their previous
experience of a layoff.
32
the firm’s tuition-reimbursement program. These data provided detailed information
on employee participation in the program, whether or not the employee earned a
degree through the program, and the date on which the bonus was awarded.
Survey data were collected as part of an on-going multi-company study
conducted by the Center for Effective Organizations at the University of Southern
California. This larger study was focused on technical employees, and units of the
company were selected for survey because of their large numbers of engineers and
technical managers. Nearly all of the surveys were sent to employees who worked in
the Research, & Development and Aftermarket, & Repair units of the company.
Because of the large numbers of engineers in these units, the sample effectively
included all salaried employees in the units except for a small number of support
staff. Employees reported the amount of time they spent in various development
activities over the past year. Responses to a number of questions with five-point
Likert scales were then used to create attitudinal scales for the dependent variables of
organizational commitment and intention to turnover. Table 2 details the response
rates of the units surveyed.
Table 2
Survey Response Rates by Unit
33
Date Surveyed Responses Rate
R, & D Facility 1 (Relocated) January 590 207 35.1%
R, & D Facility 2 January 1154 481 41.7%
Aftermarket and Repair June 829 291 35.1%
Total 2573 979 38.1%
A total of 2,573 employees were surveyed in multiple organizational units.
The same survey was administered to different units of the company in January and
June of 2000. Usable surveys were returned from 979 employees, for an effective
response rate of 38%. In January, 688 surveys were returned from Research, &
Development employees. These respondents worked in two locations, with 207
employees located at the facility where relocation efforts were already underway.
Five months later another 291 surveys were collected from engineers and managers
based mostly in Aftermarket, & Repair units of the company spread out among
multiple facilities around the U.S.
Measures
Control Variables
The relationship between employee development, attitudes and retention is
particularly complex because there are several individual characteristics and
34
organizational factors that contribute to the likelihood that an employee will
participate in development activities. These factors may also affect how an
employee interprets the development activity and translates the experience into
commitment, turnover intentions, or turnover behavior. Several control variables are
used in the analysis to account for as much of this individual and organizational
variance as possible. These control variables are organizational tenure, job grade,
and business unit.
Tenure. Tenure is included as a control variable because of its likely effect
on both participation in development and the outcomes of organizational
commitment and turnover. Research has shown job and organizational tenure to be
significant factors in predicting participation in development activities (Kozlowski,
& Farr, 1988; Kozlowski, & Hults, 1987). This may be due to increased
opportunities for long-tenured and managerial employees to participate in
development (Camevale, Gainer, & Villet, 1990). Organization tenure is also
significantly correlated with organizational commitment (Mathieu, & Zajac, 1990).
Employees who spend years with a company tend to build attachment over time. It
may also be the case that employees with low commitment to an organization tend to
have shorter tenure with that company because they leave sooner than their
counterparts with high organizational commitment. Finally, multiple studies have
concluded that tenure is negatively correlated with turnover and turnover intentions
(Griffeth, Horn, & Gaertner, 2000). This may come from the positive effects of
35
affective commitment that build during employment with an organization. It has
also been argued that long-tenured employees have greater investments in an
organization (such as pension plan contributions and firm-specific skills) which act
as “side-bets” and discourage an employee from leaving (Meyer, & Allen, 1984).
Tenure is coded from personnel records as the number of years between the
employee’s original start date and June 2000. Tenure was counted at the date of
termination for those employees leaving the company during 1996-2000. For
employees who had left the company and been rehired, their original start date is
used to calculate organization tenure.2 Tenure ranges from less than a month to 54
years for the entire salaried population 1996-2000, with an average of 16.9 years for
the 12,360 employees. Although the survey sample was similar to the archival
sample, a mean tenure of 13.9 years indicates that survey respondents were overall
slightly less experienced employees.
Job grade. Dummy variables were created for six job grades taken directly
from the HRIS personnel records provided by the company. These included three
levels of engineers and three managerial grades. Job grade was measured as the
grade the employee held in June 2000 or the job grade that the employee held when
he or she left the company. In each of the regression models using job grade as a
2 The number of rehires included in the sample examined was 43 employees or .2% of the total
sample. All of these employees were rehired before 1996 and the beginning of the study.
3 The bottom three salaried job grades (703, 749 and 550 employees respectively) were collapsed
together into a single job grade. These lowest grades represented salaried manufacturing engineers
and contained relatively smaller numbers of employees. Employees from these three grades were not
significantly different in either their attitudes or participation in development.
36
control variable, one of the engineering grades (Grade 3) was excluded as a referent
category. The coefficients estimated for each of the job grade dummy variables
should be interpreted in comparison with Grade 3, the top non-managerial job grade.
Table 3 details the distribution of the job grades for the archival and survey samples.
The job grade dummy variables are used as control variables in all
multivariate analyses because of the potential effects on participation in
development, as well as the outcomes of commitment and turnover. First of all,
research has shown that participation in employee development often differs by the
level of the job within an organization. In general, managerial and professional
employees are more likely to receive training (Lynch, 1992; Tharenou, 1997). It has
been suggested that this is because managerial employees have greater opportunities
to attend training than lower-level employees (Camevale, Gainer, & Villet, 1990).
However, only salaried employees are included in the present study, which reduces
this potential variance in development participation.
37
Table 3
Distribution of Job Grades
Archival
Percent
Survey
Percent
Grade 1 - Engineer 1 1902 15.4% 74 7.6%
Grade 2 - Engineer 2 1219 9.9% 145 14.8%
Grade 3 - Engineer 3 3007 24.3% 202 20.6%
Grade 4 - Manager 1 4204 34.0% 345 35.2%
Grade 5 - Manager 2 1239 10.0% 131 13.4%
Grade 6 - Manager 3 789 6.4% 82 8.4%
Total 12360 100% 979 100%
This focus on salaried employees is also likely to reduce the variance in
organizational commitment among the different job grades. In terms of
commitment, research suggests employees at higher job levels tend to commit more
strongly to their employing organization (Kallenberg, & Reve, 1992). It is possible
that an employee views higher-level jobs as a greater amount of trust and investment
from the organization, which generates increased commitment to the organization
(Eisenberger et al., 1986). In the present case there are no significant links between
job grade and either voluntary turnover or organizational commitment.
The primary reason for including job grade as a control variable, however, is
its relationship to the rewards variables used as moderators in the analyses.
Employees in the highest job grades are generally more likely to have received
promotions and merit pay from the company than employees in lower job grades.
38
Fifty-eight percent of employees in Grade 5 had been promoted during the study
period, making them the most likely to have received a promotion 1996-2000. On
the other hand, only 28% and 33%, of Grades 1 and 3 respectively, had received
promotions. More than 90% of the employees in the top two job grades received a
merit award over the study period, while 80% or fewer employees in the lowest three
job grades had received merit pay.
Business unit. The final control variables included in analyses of the survey
data are two dummy variables used to control for variance among the different units
within the company. Dummy variables are used to indicate whether the employee
worked in one of the two Research, & Development facilities. The coefficients of
these two variables should therefore be interpreted in comparison to Aftermarket, &
Repair as the unmeasured comparison category. These two dummy variables are
included as controls in the survey analyses, but are unavailable for the archival data
analysis. However, the amount of variance among voluntary turnover that can be
attributed to differences among business units declines when the sample includes all
salaried employees across a 54-month time period.
The primary reason for including these dummy variables is that facility
closure and layoffs impacted some units of the company more than others.
Downsizing during the study period is likely to have affected employee attitudes and
voluntary turnover. In particular, the second dummy variable represents employees
who worked at a large facility that was in the process of closing in an effort to cut
39
costs and reduce headcount. The survey was conducted at a time after the closure
had been announced and before the process had been completed. The process of
closing the facility included offering a proportion of the employees the opportunity
to move to new jobs at the headquarters facility, which is located in a different part
of the country. Job offers were made department-by-department on a rolling
timetable with the remainder subject to involuntary layoffs. The result was a
considerable amount of uncertainty, dissatisfaction, and voluntary turnover in the
employees who remained.
Predictor Variables
Participation in development. Individual participation in training and
development is measured using both company-provided data and self-reports from
employees. Self-reports from the 979 survey respondents measure the number of
days spent in three different development activities in 1999 (see Appendix 2 for the
development measures). These activities are company training classes, tuition-
reimbursement courses and on-the-job training. Participation in employee
development is measured first as a continuous measure of time spent because the
level of company-provided benefit is operationalized as the amount of the
employees’ personal time versus the amount of company-time spent in development.
This method follows Birdi et al. (1997) and other models of the outcomes of
employee development based on social exchange, which imply an increasing (or
40
decreasing in the case of tuition-reimbursement) relationship between employee
development, organizational commitment, and employee retention.
Company training classes, on-the-job training, and tuition-reimbursement
classes are the most common types of development activities in organizations and
account for the vast majority of time that employees spend developing new skills
(Frazis, Herz, & Horrigan, 1995). Although training information systems have
drastically improved, many companies do not collect systematic information on
development activities in general and on-the-job training in particular. The company
studied only collected systematic data on participation in tuition-reimbursement.
Because accurate objective training time data are rare, self-report data are commonly
used in training research studies (Maurer, & Tarulli, 1994; Noe, & Wilk, 1993).
The self-report data show that tuition-reimbursement and structured on-the-
job training have similar distributions among the employees studied. Between one-
quarter and one-third of the employees surveyed participated in these activities
during the preceding twelve months (24% for tuition-reimbursement and 34% for on-
the-job training). Although more employees experienced on-the-job training, the
average length of time spent in actual development activities was greater for tuition-
reimbursement courses. Company training classes were the most common
development activity reported by employees, with nearly 82% participating. It
should be noted that the company requires participation in some form of company-
sponsored training and that several of the business units maintained goals of up to 20
41
hours or more annually per employee. Adherence to these policies varies across
units.
Development participation is also measured from the company’s archival
records on the company’s tuition-reimbursement program 1996-2000. These data on
(1) whether the employee took courses without earning a degree, and (2) whether the
employee earned a degree and received the bonus award from the company were
coded as dummy variables and matched with personnel records. Taking classes
without earning a degree was correlated at .32 (p < .001) with the self-reported
participation in tuition-reimbursement classes even though the survey measured
participation over the past 12 months compared to the archival data which measured
participation over 54 months. During the study period, 38.4% of all employees
participated in tuition-reimbursement without earning a degree, and 8.9% earned a
degree through the company.
Participation in tuition-reimbursement is hypothesized to be negatively
related to organizational commitment and employee retention. No separate
hypotheses are proposed for these two measures of participation. However, since the
signaling effects of earning a degree are likely to create greater external job
opportunities for employees than participation alone, the effects on commitment and
retention of completion should be greater than simply working towards a degree. In
each of the analyses of past participation in tuition-reimbursement, separate variables
42
were used for those employees who have earned a degree through the company and
those who participated without yet earning a degree.
Rewards. Rewards from the company are measured as the receipt of merit
pay or a promotion. For those employees who earned a degree through tuition-
reimbursement, merit awards or promotions were received in addition to the $10,000
bonus paid for the completion of a degree. The company studied is somewhat
unique in that there are no within-grade raises, meaning that merit pay and
promotions are the only ways in which employees within the company can increase
their compensation.
Variables for receiving a promotion and receiving a merit award were coded
directly from the HRIS personnel records provided by the company. Coding was
done by examining the entries into each employee’s electronic personnel file over
the 54-month study period. Every time a change is entered into the company’s
database, the date and the reason for the entry is recorded from a list of possible
actions that would require the employee’s personnel file to be updated. Table 4
describes the percentage of survey respondents and all employees who received
merit pay and promotions at any time during the study period.
43
Table 4
Employees who Received Merit Pay and Promotion
N
Received
Merit Pay
Received a
Promotion
All Employees 1996-2000 12360
78.7% 36.5%
1996 Active Employees 9243
92.4% 36.8%
Survey Respondents 979
77.2% 43.6%
The dummy variable for receiving a promotion was coded based on whether
an employee’s personnel file had been updated with actions named “Promotion” and
“Promotion w/ Department Change.” The dummy variable for receiving merit pay
was coded based on whether an employee’s personnel file had been updated with the
actions “Merit Pay” or “Lump Sum Payment.” Merit pay and promotions are
positively correlated (.18, p < .01) but not directly linked.
Both wage increases and career advancement have been shown to have direct
effects on the commitment and retention of employees. For example, research has
consistently shown a positive relationship between promotions and organizational
commitment (Johnston et al., 1993; Romzek, 1989). In terms of merit pay, a positive
relationship has been demonstrated between the level of an individual’s salary, and
the receipt of discretionary rewards on organizational commitment (Mathieu, &
Zajac, 1990). Conversely, research has also found that failing to receive a promotion
44
has an impact on an employee’s desire to leave an organization (Shwarzwald,
Koslowsky, & Shalit, 1992).
There is some uncertainty, however, concerning whether the effects of
rewards such as a promotion or merit award on commitment and retention are
derived from career advancement or wage increases. For example, Trevor, Gerhart
and Boudreau (1997) found that the effect of receiving a promotion on turnover is
insignificant when the increase in salary is controlled. Unfortunately the importance
of career advancement versus wage increases cannot be addressed in the present
study because company policy limits within-grade salary increases. Career
advancement (such as promotion) and wage increases occur together and therefore
cannot be compared.
Instead, merit pay and promotions are considered as two separate measures of
rewards that include elements of both career advancement and wage increases.
Receiving either of these two rewards after participating in tuition-reimbursement
classes is hypothesized to moderate the effects of developing general skills on
commitment and employee retention. No theoretical predictions are made, however,
about the relationship between these two measures of rewards, and they are used
separately in the analyses. Regressions are estimated for each of these two variables
I
to provide multiple perspectives on the same phenomenon of rewarding employees
when they gain new general skills through tuition-reimbursement.
45
Dummy variables indicating receipt of merit pay or a promotion at any time
during the five-year study period are included in all regression analyses to control for
the main effects of receiving these rewards on employee attitudes and turnover.
Interaction terms are then used to test the effects of participating in tuition-
reimbursement together with receiving either a promotion or a merit award from the
organization. Interactions terms were coded in two ways depending on whether the
employee earned a degree through tuition-reimbursement or participated in tuition-
reimbursement without earning a degree.4 For those who participated without
earning a degree, the dummy variable for taking classes 1996 - 2000 was multiplied
with the dummy variable for either receiving a promotion at any time 1996 -2000 or
receiving merit pay at any time 1996-2000.5 For those who earned a degree through
tuition-reimbursement, the interaction is coded based on whether employees received
a promotion or a merit award after they had earned a degree through the company.
The date that a promotion or merit award was entered into the company’s database
was compared with the date that an employee received a degree through the tuition-
reimbursement program.
4 Earning a degree is associated with a specific date (i.e. graduation) and can be compared to the date
that a promotion or merit award was received. Taking classes through tuition-reimbursement without
earning a degree is not associated with specific dates in data. Therefore, it was impossible to code
whether the reward had occurred after participating in tuition-reimbursement.
5 Interactions between participating in tuition-reimbursement without earning a degree and the
rewards variables test whether the employee experienced both of these events at any time 1996 - 2000
regardless of the sequence in which they occurred.
46
Dependent Variables
Organizational commitment and intention to turnover were measured in the
employee survey using five-point Likert scale items. Factor analyses using Varimax
rotation demonstrated that the survey items selected formed independent and reliable
scales for two dependent variables (please see Appendix 3 for the rotated component
matrix). Voluntary turnover was gathered for all 12,364 salaried employees using
personnel records January 1996 to June 2000. The turnover of employees who
participated in the employee surveys in either January or June 2000 was also
measured at the end of December 2000.
Organizational commitment Based on an exploratory factor analysis of the
survey items used in the study, six items were selected from Porter et al.’s (1979)
nine-item scale to measure organizational commitment. Items were dropped to
reduce the items to a single factor for affective commitment and eliminate cross
loading with intention to turnover. These items dropped from the scale were: “I
would accept any type of job to continue working for this organization,” “I really
care about the fate of this organization,” and “I am willing to put in effort beyond the
norm for this organization.” The remaining items include, “I am extremely glad to
have chosen this organization to work for over other organizations,” and “For me
this is the best of all organizations for which to work.” The measure of commitment
shares at least one item, “I am proud to tell others that I am part of this organization,”
with the three-item measure used by Birdi et al. (1997) in their study of the
47
relationship between development participation and attitudes. The survey sample
reported a reliable scale with Chronbach’s alpha of .87 for organizational
commitment.
Intent to turnover. Intention to turnover is measured using two survey items
where respondents indicated the extent to which they agreed with the statements: “I
would be willing to change companies right now for career advancement” and “I
plan to look outside my organization for a new job within the next year.” These
items were written for this survey, and are similar to items such as, “As soon as I
find a better job I will quit,” (Begley, & Czajka, 1993) and “What are the chances
you will quit your job in the next 12 months” (Davy, Kinicki, & Scheck, 1997; Horn,
& Griffeth, 1991; Horn, Griffeth, & Sellaro, 1984; Johnston et al., 1993). This
dependent variable is intended to capture the willingness of employees to leave the
company for other jobs and their intention to begin searching for a new job. This
follows Horn and Griffeth’s (1991) confirmatory factor analyses that found that the
variety of withdrawal cognitions (e.g., thoughts of quitting, search intentions, and
intentions to quit) are best represented as a single measure. Chronbach’s alpha for
the intention to turnover scale was .71.
Turnover. Turnover was coded from electronic personnel records that
included termination date and reason why the individual left the company. The
reasons were categorized between voluntary turnover or “quitting,” and involuntary
turnover that includes layoffs and terminations. The distinction between voluntary
48
and non-voluntary sources of turnover is supported by the literature and more recent
research into employee turnover (Gupta, & Jenkins, 1991; Lee, & Mauer, 1999; Lee,
Mitchell, Holton, McDaniel, & Hill, 1999). However, distinguishing between
voluntary and non-voluntary turnover is more difficult in this case because the
company underwent downsizing, which included the opportunities for certain
employees to take severance packages or early retirement plans. These employees
accounted for 22% of all turnover (896 of 3975 employees) over 54 months for
which data were available. This study focuses only on the employees who quit the
company voluntarily without early retirement. These employees accounted for
33.5% of all turnover (1334 of 3975 employees) during the study period. Employees
who left the company for other reasons were used for comparison, but not for
analyzing the effects of employee development on voluntary turnover.
Voluntary turnover is the outcome of interest in individual-level models of
turnover because it represents an employee’s decision to leave an organization
(Shaw, Delery, Jenkins, & Gupta, 1998). In the case of severance packages or early
retirement, the decision to leave the organization is clearly influenced by other
factors, such as the fact that severance and early retirement packages tend to be
offered only to long-tenured employees. In fact, employees who took early
retirement or severance packages averaged 33.5 years with the company, compared
with 10.3 years for those counted as voluntary turnover. In order to focus the
analyses on how participation in employee development affects an employee’s
49
decision to quit the company (as opposed to retire early), those employees taking
severance or early retirement are not counted as voluntary turnover.
For analysis of the archival data on participation in tuition-reimbursement,
turnover is measured over the period of 54 months from January 1996 until June
2000. Table 5 presents annualized rates for all 12,360 employees for quits and all
turnover during this time. For analysis of the self-reported participation in tuition-
reimbursement measured through the survey, turnover is seven or eleven months
later depending on whether the survey was administered in January or June of 2000.
The turnover rate for the 979 survey respondents was 4.5% or 44 employees.
Table 5
Turnover Rates by Year for all Employees
1996 1997 1998 1999 2000
Active Employees 9141 9453 9742 10337 9270
Quits 169
(1.4%)
208
(1.7%)
211
(1.7%)
304
(2.5%)
184
(1.5%)
All Turnover* 337
(2.8%)
757
(6.2%)
389
(3.2%)
1439
(11.8%)
317
(2.6%)
* Includes layoffs, voluntary separations and terminations.
Analyses
Twelve hypotheses are tested using several different combinations of data
and methods, which are summarized in Table 6.
50
Table 6
Summary of Hypotheses, Data Sources and Analyses
Hypo’s Regression
Method
Development
Data Source
Development
Activity
Hypothesized
Relationship
Predicted
Outcome
1 OLS Self-report Co. Classes Positive Commitment
2 OLS Self-report OJT Positive Commitment
3 OLS Self-report Co. Classes Negative Intention
To Turnover
4 OLS Self-report OJT Negative Intention
To Turnover
' 5 Logistic Self-report Co. Classes Negative Turnover
6 Logistic Self-report OJT Negative Turnover
7 OLS Self-report Tuition-
reimbursement
Negative Commitment
8 OLS Self-report Tuition-
reimbursement
Positive Intention
To Turnover
9 Cox Archival Tuition-
reimbursement
Positive Turnover
10 OLS Archival Tuition-
reimbursement
Moderated
by Rewards
Commitment
11 OLS Archival Tuition-
reimbursement
Moderated
by Rewards
Intention
To Turnover
12 Cox Archival Tuition-
reimbursement
Moderated
by Rewards
Turnover
The analyses can be divided into four distinct sets depending on the time
period examined and whether the self-reported training participation data or the
archival data on tuition-reimbursement are examined. The first is a correlational
study that uses self-report data to predict attitudes from the same survey conducted in
early 2000. The second uses archival data on tuition-reimbursement to predict
51
attitudes measured in the survey. Both of these analyses of employee attitudes use
hierarchical OLS regression. The third uses self-report development data in a
logistic regression to predict turnover seven to eleven months later in December
2000. The fourth analysis uses archival tuition-reimbursement data in contingency
tables and Cox proportional hazards regression to examine turnover January 1996-
June 2000.
Development and Attitudes
The relationship between development and employee attitudes is examined in
Chapter 3 using OLS multiple regression to test both survey and archival measures
of development participation. The first analysis tests Hypotheses 1-4, 7, and 8 using
survey data and hierarchical OLS regression to predict the influence of the three
types of development on organizational commitment and intention to turnover.
Similar to Birdi et al. (1997), self-reported time in development measured as a
continuous variable is used to predict attitudes also measured in the employee
survey. Entered first into the hierarchical regression equations are company tenure,
job grade, and business unit. Following these individual and organizational control
variables, time spent in company training classes, on-the-job training, and tuition-
reimbursement classes over the preceding twelve months are entered together into
the regression. Separate models are estimated for the effects of employee
development on organizational commitment and intention to turnover.
52
The second analysis uses hierarchical OLS regression to test archival
measures of tuition-reimbursement participation. This analysis tests Hypotheses 7
and 8 along with Hypotheses 10 and 11 regarding whether receiving rewards
following participation in tuition-reimbursement classes moderates the effects of
participation on employee attitudes. Hierarchical OLS regression models are used to
predict both organizational commitment and intention to turnover using dummy
variables for taking classes without earning a degree and earning a degree through
tuition-reimbursement. These two participation variables are entered in step 1 of the
regression along with control variables for company tenure, job grade, and business
unit. In step 2a and 2b, a dummy variable for either receiving a promotion or a merit
award 1996-2000 is entered into the model along with interaction variables for (1)
taking classes without earning a degree and receiving a promotion or merit award
1996-2000, and (2) earning the degree through tuition-reimbursement and then
receiving either a promotion or merit award. The interaction terms test whether
receiving either a promotion or merit award moderates the relationship between
tuition-reimbursement, commitment, and intention to turnover.
Development and Retention
Employee retention as an outcome of training and development is examined
in Chapter 4. A combination of categorical analysis, logistic regression and Cox
proportional hazards regression are used to test the relationships between general
53
skills development, merit pay, promotions, and employee retention. The first
analysis uses survey data on self-report training time in a logistic regression to
predict employee turnover measured seven to eleven months later to test Hypotheses
5, 6 and 9. Turnover was measured at the end of the 2000 calendar year for those
employees responding to the survey earlier that year in January and June. In step 1,
control variables for company tenure, job grade, and whether the employee worked
in the relocated facility are entered in the model. Following these control variables,
step 2 of the regression tests the effects of participating in company training classes,
on-the-job training and tuition-reimbursement on turnover by entering variables for
the time employees reported spending in each of the three development activities
over the past 12 months.
The second analysis tests Hypothesis 9 on the direct effects of tuition-
reimbursement participation on turnover and Hypothesis 12 regarding the
moderating effects of receiving rewards following participation using Cox
proportional-hazards regression (Hosemer, & Lemeshow, 1999; Morita, Lee, &
Mowday, 1993; 1989). Cox proportional hazards regression (hereafter called “Cox
regression”) is an increasingly popular approach to analyzing organizational turnover
(Lee, & Maurer, 1999; Somers, 1996; Trevor, 2001). Instead of approaching
turnover as a simple binary event that differentiates between “quit” and “not quit,”
Cox regression includes information on the duration of employee retention. Using a
measure of duration permits the estimation of the conditional probability of leaving
54
over time, which is called a hazard function (Hosmer, & Lemeshow, 1999). Cox
regression tests whether independent variables contribute to the predictive power of a
multivariate hazard function (Hosemer, & Lemeshow, 1999; Morita, Lee, &
Mowday, 1993). The significance of the independent variables is tested using Wald
statistics and the coefficients are interpreted in terms of exponentiated odd ratios.
Two sets of Cox regressions are estimated using the 9,243 employees who
were active at the beginning of 1996 when the company’s new liberalized tuition-
reimbursement policy was announced and the first degrees were awarded. Step 1 of
first Cox regressions includes control variables for company tenure and job grade, as
well as dummy variables for taking classes without earning a degree, and earning a
degree during the study period. In step 2a and 2b, the effects of receiving either a
merit award or a promotion at any time during the study period 1996-2000 are tested
using dummy variables and interactions between the two tuition-reimbursement
dummy variables (earning a degree and taking classes without earning a degree) and
the reward dummy variable (received either promotion or merit pay 1996-2000).
Although the first Cox regressions test the effects of participation in tuition-
reimbursement with and without earning a degree and receiving rewards together,
the data are coded so that the models do not take into account the sequence in which
the two events (participating in tuition-reimbursement and receiving a reward)
actually occur. Theoretically, the more interesting case is when rewards actually
follow participation in tuition-reimbursement generally, and earning a degree in
55
particular. The second Cox regressions use a different method to calculate the
interaction between earning a degree and receiving either a promotion or merit award
to test the importance of receiving rewards after earning a degree through tuition-
reimbursement.6
This is done by analyzing receiving a degree through tuition-reimbursement
and receiving a subsequent merit award or promotion as time-dependent covariates.
Time-dependent covariates are created by interacting elapsed time with a categorical
independent variable in addition to predicting time-to-tumover as the dependent
variable (Morita et al., 1993). Step 1 of the second Cox regressions include the same
control variables (company tenure and job grade) and dummy variables for whether
or not the employee took classes without earning a degree, or earned a degree
through tuition-reimbursement. In steps 2a and 2b of these Cox regressions,
however, the interaction between earning a degree and receiving either a promotion
or merit award is calculated using the elapsed time between the degree and when the
employee received a promotion or merit award. Again, two separate models are
estimated for the effects of receiving a merit award and receiving a promotion.
The time-dependent covariate tests whether the length of time elapsed
between receiving a degree and receiving either a merit award or promotion has
additional predictive power beyond receiving the reward without regards to its
6 Earning a degree is the only measure of tuition-reimbursement that is analyzed using the time-
dependent method because it can be associated with a specific graduation date. Taking classes
without earning a degree is not date specific and cannot be coded and analyzed as a time-dependent
covariate.
56
relative timing to earning a degree through tuition-reimbursement. The use of time-
dependent covariates within the Cox regression models provides additional tests of
Hypotheses 9 and 12 and the notion that rewards need to follow earning a degree
through tuition-reimbursement in order to moderate the effects of general skills
development on the retention of employees.
57
III. DEVELOPMENT AND EMPLOYEE ATTITUDES
The effects of development on employee attitudes were tested using
hierarchical multiple regression and ANOVA in which organizational commitment
and intention to turnover were the dependent variables. The results are divided into
sections, with self-reported survey data presented first, followed by archival data on
participation in tuition-reimbursement 1996-2000. A total of eight hypotheses are
tested.
Results
Self-Reported Development Time and Attitudes
Preliminary results. Descriptive statistics are presented in Appendix 4, and
correlations are in Appendix 5. Initial inspection of the data indicates that all three
self-report measures of development time are negatively correlated with company
tenure and job grade. That is, participation in development activities is more
prevalent among those in lower job grades and those with less experience. This may
appear to go against previous empirical research, which finds that training is
58
generally more available to higher-level and long-tenured employees (Camevale,
Gainer, & Villet, 1990). However, this finding should be interpreted in the context
of the sample of employees surveyed. All of the employees included in the study are
salaried engineers and managers, whom previous research finds to have more
opportunities for training than hourly workers and non-technical workers (Frazis,
Herz, & Horrigan, 1995). All of the employees studied also work in technical jobs in
a highly specialized industry that requires a great deal of up-front learning in order to
be effective. In this case it is not surprising to find less-experienced and lower-level
employees participating more often in development activities.
Examination of the three self-report training participation measures revealed
that the distributions for each of the development activities were skewed with several
large observations. Several individuals reported spending more than ten times the
average amount of time in tuition-reimbursement courses, company-sponsored
classes, and on-the-job training. To deal with this problem, each of these variables
was normalized using a natural log (In) transformation. The effectiveness of this
transformation is demonstrated in p-p plots in Appendix 6 of the development time
variables before and after transformation.
Initial analysis of survey responses from employees provides some evidence
that employees view the developmental content of tuition-reimbursement classes, on-
the-job training, and company training classes differently in terms of how useful they
are to an employee’s marketability for other jobs. On a five-point scale employees
59
rated for-credit courses (tuition-reimbursement) as the most useful (3.81), with
structured on-the-job training (3.02) and company classes (3.02) rated lower. Figure
1 lists the average usefulness ratings of the three different development activities.
Figure 1
Employee Ratings of Development Usefulness.*
4.50
S 3.50 —
T O
•4—*
CD
u
(O
3
L L
Tuition- Company training On-the-job training
reimbursement classes
*E m ployees w ere asked to evaluate developm ent activities in term s o f
“Their usefulness to yo u r fu tu re marketability. ” R esponses w ere given on
a five-p o in t scale from “N ot useful a t a ll” to “Extrem ely useful. ”
Tuition-reimbursement differs significantly from company classes and on-
the-job training in terms of its usefulness to future marketability (ANOVA, p <
.001). See Appendix 7 for the complete distributions of employee assessments of the
usefulness to future marketability of the three development activities. This finding
60
provides some evidence from the present sample of the validity of the assumption
that participation in tuition-reimbursement may be used as a theoretical proxy for
general skills development. This finding aligns with previous studies of general
skills training (Loewenstein, & Spletzer, 1997) as well as the methodology of
previous studies of general skills training and turnover, which has examined “on-the-
job” and “off-the-job” training to compare general versus specific skills (Lynch,
1991; 1992).
Hypotheses tests. Self-reported training time from the survey is used first to
test Hypotheses 1- 4, 7 and 8. Transformed training times for tuition-reimbursement
classes, company-sponsored classes, and on-the-job training are used in two
hierarchical OLS regression models predicting organizational commitment and
intention to turnover. In the first model predicting organizational commitment, the
addition of the development variables increased the variance explained over the
control variables by a small but significant amount, from 7.8% to 8.7% (F= 2.91, p
< .05). This increase can largely be attributed to the effect of participation in on-the-
job training, which was the only development activity significant in predicting
organizational commitment (t, p < .01). A coefficient of .09 supports Hypothesis 2
that on-the-job training is positively related to organizational commitment.
Hypotheses 1 and 7 are not supported, as the time spent in both company training
classes and tuition-reimbursement showed no significant relationship with
61
organizational commitment. Regression results for organizational commitment are
presented in Table 7.
Table 7
Results of Regression Analyses for Organizational Commitment
Step 1
Betaa t
Step 2
Beta3 t
Tenure1 * -0 .1 0 -2 .6 6
**
-0 .1 0 -2.61
**
Grade 1 0 .1 0 2.67
**
0.08 2.30
*
Grade 2 0.13 3.45
**#
0.13 3.29
***
Grade 4 0.06 1.59 0.06 1.66 +
Grade 5 0 .1 0 3.06
**
0.11 3.14
**
Grade 6 0 .1 2 3.60
***
0 .1 2 3.71
***
Business Unit 1 0.05 1.25 0.04 0.98
Business Unit 2 -0.15 -3.98
***
-0.15 -3.88
***
In (Tuit’n-reimbursement) -0.04 - 1.10
In (On-the-job training) 0.09 2.77
**
In (Company classes) 0.01 0.15
R2 (adjusted R2) .078 (.071) .087 (.076)
F 10.13
***
8 .2 0
***
D f
962 962
A R 2 .008
A F 2.91
*
N 973 973
°standardized.
+ P<.10
*P<. 05
* * P < .01
* * * P < .001
The self-report data are next used in a hierarchical regression predicting
intention to turnover using the same control variables. Regression results for
intention to turnover are presented in Table 8.
62
Table 8
Results of Regression Analysis for Intention to Turnover
Step 1
Beta3 t
Step 2
Beta3 t
Tenure3 -0.24 -6.42
***
-0.23 -6.37
***
Grade 1 -0.08 -2.32
*
-0.07 -1.99
*
Grade 2 -0.08 -2.03
*
-0.08 -2.15
*
Grade 4 -0.06 -1.54 -0.06 -1.76 +
Grade 5 0 .0 2 0.48 0.01 0.32
Grade 6 0 .0 0 -0.08 -0.01 -0.28
Business Unit 1 -0.13 -3.54
***
-0 .1 2 -3.26
***
Business Unit 2 0.18 4.87
***
0.17 4.67
***
In (Tuit’n-reimbursement) 0.14 4.67
In (On-the-job training) -0.15 -4.75
***
In (Company classes) 0 .0 2 0.58
R2 (adjusted R2) .108 (.1 0 0) .141 (.130)
F 12.77
***
12.87
***
D f 955 955
A R 2 .032
A F 11.85
***
N 973 973
0 standardized
+ P < . 10
* P < .05
**P< .01
* * * P < 001
The addition of the three development variables significantly increased
variance explained in the model of intention to turnover by 3.2%, from 10.8% to
14.1% (F=l 1.8, p < .001). The coefficients for on-the-job training and tuition-
reimbursement classes were significant (t, p < .001). These findings support
63
Hypothesis 4 that on-the-job training leads to decreased intention to turnover, as well
as Hypothesis 8 that participation in tuition-reimbursement is associated with higher
intention to turnover. The coefficient for company training is not significant and
Hypotheses 3 is not supported.
Tuition-Reimbursement and Attitudes
The second analysis of development and attitudes uses archival data from
January 1996 through June 2000 for the 979 employees responding to the survey to
test the relationship between participation in the company’s tuition-reimbursement
program, organizational commitment and intention to turnover. Hierarchical OLS
regression models test the relationship between tuition-reimbursement and attitudes
using dummy variables, which measure whether the employee has taken classes
without earning a degree and whether the employee has earned a degree through the
company’s tuition-reimbursement program.
Preliminary results. Descriptive statistics are presented in Appendix 4, and
correlations are in Appendix 5. Initial inspection reveals that correlations between
receiving a merit payment, receiving a promotion, organizational commitment and
intention to turnover are all significant. However, the correlations between receiving
a promotion and the outcomes are the opposite of what might be expected.
Employees who had been promoted or received merit pay indicated higher intention
to turnover and lower organizational commitment. Previous research suggests that
64
rewards such as promotions are related to higher commitment (Johnston et al., 1993;
Romzek, 1989).
These findings may reflect the curvilinear relationship between employee
performance and voluntary turnover that has been demonstrated in previous research
(Jackofsky, 1984; Jackofsky, Ferris, & Breckenridge, 1986). Studies have found that
the highest- and lowest-performing employees are generally more likely to turnover
(Trevor, Gerhart, & Boudreau, 1997). This is presumably because higher-
performing employees are more likely to have greater external job opportunities. In
the case of high performers who receive a merit award or a promotion, such a reward
might act as an additional signaling device to potential employers about the
productivity of the employee, which would translate into additional external job
opportunities. Increased job opportunities are associated with lower organizational
commitment and higher intention to turnover in employees (Bateman, & Strasser,
1984; Gerhart, 1990; Kirschenbaum, & Mano-Negrin, 1999).
Hypotheses tests. Similar to the analysis of the self-reported measure of
participation in tuition-reimbursement, no significant relationships were found
between the archival measures of tuition-reimbursement and organizational
commitment. The following results test the effects of tuition-reimbursement on
intention to turnover. Figure 2 illustrates differences in intention to turnover across
all employees surveyed, categorized by their level of participation in tuition-
reimbursement, without controlling for other variables.
65
Figure 2
Participation in Tuition-Reimbursement and Intention to Turnover
< 1 )
>
o
c
1 —
H
o
- # - »
c
o
c
< u
c
4.50
4.25
4.00
3.75
3.50
3.25
3.00
2.75
2.50
3.93
3.47
3.27
3.33
Not Took Classes Earned a Earned Degree then
Participated (N=395) Degree (N=95) Degree, No Reward
(N=480) Reward After (N=72)
(N=23)
*R eceiving a rew ard after earning a degree is defined as a prom otion o r m erit
aw ard in addition to the $10,000 bonus received b y a ll em ployees who earned a
degree through tuition-reimbursement.
Figure 2 supports Hypotheses 8 by showing a positive and significant
relationship between intention to turnover and participation in tuition-reimbursement
across all employees surveyed (ANOVA, p < .01). Average intention to turnover for
those employees who did not participate in tuition-reimbursement during 1996-2000
was 2.88 compared with 3.27 for those employees who took at least one class.
Employees who earned degrees through the program reported the highest intention to
turnover, with an average of 3.47. Tukey least squares difference (LSD) tests show
66
that the differences between each of first three groups presented in Figure 2 are
significant (t, p < .001).
The effects of receiving either a promotion or a merit award after
participation in tuition-reimbursement can tested by further examining the
experiences of the 95 survey respondents who earned a degree through the program.
The two bars on the right-hand side of Figure 2 represent the 95 employees who
earned a degree through tuition-reimbursement, split by whether or not they received
either a promotion or merit award in addition to the $10,000 bonus after they earned
their degree. Intention to turnover reported by employees who did not receive either
a promotion or merit award following their degree (3.93) was higher than employees
who received either a promotion or merit award (3.33) after graduation. A Tukey
LSD test indicates that this difference is only marginally significant (t, p < .10), but
supportive of Hypothesis 11. The weakness of this significance test is most likely
due to the small comparison group of 23 survey respondents who earned a degree but
did not subsequently receive either a promotion or merit award.
The archival tuition-reimbursement measures were also analyzed using
multivariate OLS regressions to test Hypotheses 7, 8, 10 and 11. Hypothesis 7
predicts a negative relationship between tuition-reimbursement and organizational
commitment, while Hypotheses 8 predicts a positive relationship between tuition-
reimbursement and intention to turnover. Hypotheses 10 and 11 predict that
receiving rewards after tuition-reimbursement moderates these effects. Regressions
67
are estimated for both organizational commitment and intention to turnover using
either promotions or merit pay to moderate the effects of tuition-reimbursement.
Entered in step 1 of the model are control variables for company tenure and
job grade, along with two dummy variables representing the employee’s tuition-
reimbursement history going back to January 1996. These dummy variables indicate
whether the employee took classes without earning a degree, or if the employee
earned a degree through tuition-reimbursement. In steps 2a and step 2b of the
regression, a dummy variable indicates whether the employee had received either a
promotion or a merit award at any time 1996-2000, along with two interaction terms.
For the interaction between taking classes without earning a degree and the reward,
the tuition-reimbursement is multiplied with the dummy variable for receiving either
a promotion 1996-2000 or a merit award 1996-2000. For the interaction between
earning a degree and the reward, a new variable was coded that indicates whether the
employee received either a promotion or a merit award after they earned a degree
through tuition-reimbursement.
In the first regressions predicting organizational commitment, coefficients for
both measures of participation in tuition-reimbursement were not significant (see
Appendix 8 for results). Hypothesis 7 is therefore not supported. The absence of a
direct relationship also means that Hypothesis 10 is not supported. Results for the
second two regressions predicting intention to turnover are presented in Table 9.
68
Table 9
Hierarchical Regression Results for Intention to Turnover: Archival Data
Step 1
Beta t
Step 2a
Beta t
Step2b
Beta t
Tenure3 -0.22 -6.13 *** -0.21
*% *
-0.27 -6.87 ***
Grade 1 0.13 2.12 * 0.13 2.09 * 0.09 1.32
Grade 2 0.01 0.25 0 .0 0 0 .0 0 0 .0 0 -0 .1 0
Grade 4 0.05 0.96 0.05 1.01 0 .0 2 0.39
Grade 5 0.10 1.99 * 0.08 1.48 * 0.06 1.22
Grade 6 0.08 1.75 + 0.07 1.57 + 0.05 1.08
Business Unit 1 -0.13 -3.59 *** -0 .1 2 -3.37 *** -0.13 -3.62 ***
Business Unit 2 0.17 4.75 *** 0.17
4 .7 4 ***
0.16 4.46 ***
Earned Degree 0.15 4.84 *** 0.17
4 35 ***
0.16 3.29 ***
Took Classes 0.15 4.86 *** 0.15 3.63 *** 0.19 2.92 **
Promoted 1996-2000 0.11 2.61 **
Progress X Promoted -0 .0 2 -0.30
Degree then Promoted -0.07 -1.69 +
Merit Pay 1996-2000 0.14 3.09 **
Progress X Merit Pay -0.07 -1.04
Degree then Merit Pay -0.03 -0.67
R1 (adjusted R2 ) .134 (.125) .145 (.133) .143 (.132)
F 14.9 *** 12.4 *** 12.2 ***
D f 969 969 969
N 985 985 985
“standardized
+ P<.10
* P < .05
* * P < .01
* * *P < .001
Supporting Hypothesis 8, the standardized coefficients for both measures of
tuition-reimbursement participation were significant in step 1 of the regression
showing a positive relationship with intention to turnover (t, p < .001). To test the
moderating effects of receiving rewards, separate models are estimated using the
69
interaction of merit pay and promotion with earning a degree in steps 2a and 2b. The
addition of the interaction between these two rewards and earning a degree through
tuition-reimbursement shows mixed support for Hypothesis 11. Receiving a merit
award after earning a degree does not significantly reduce intention to turnover.
Receiving a promotion after earning a degree, on the other hand, predicts intention to
turnover beyond the effects of earning a degree or a promotion alone. The
interaction term is only moderately significant (t, p < .10), but provides at least some
support for Hypothesis 11 given that only twenty-three employees earned a degree
and did not receive either a promotion or merit award.
Given the finding that receiving a promotion after earning a degree through
tuition-reimbursement reduces intention to turnover, the question arises as to whether
the effects of receiving rewards following tuition-reimbursement are more than just
the effects of receiving the merit award or the promotion. The multivariate results
presented in Table 9 demonstrate that the interaction of earning a degree and being
promoted has effects beyond receiving a promotion alone. Further examination of
the data suggests that the timing of receiving a promotion relative to earning a degree
also has an impact on intention to turnover of employees.7
7 Regression results presented in Table 9 indicate that receiving a merit award after earning a degree
through tuition-reimbursement does not predict intention to turnover. Only the timing o f receiving a
promotion relative to earning a degree is examined.
70
Table 10
Earning a Degree, Timing of Rewards and Intention to Turnover
Mean N
Std.
Deviation
Promoted no Degree 3.33 35 1.10
Degree Then Promotion 3.32 37 1.07
Promoted before Degree 3.63 30 1.06
Earned Degree, Never Promoted 3.93 23 0.83
ANOVA p<.001
Table 10 shows that the 30 employees who received a promotion before, but
not after, they earned a degree through tuition-reimbursement reported 3.63 for
intention to turnover. These employees who received a promotion before, but not
after, earning a degree reported higher intention to turnover than the 35 survey
respondents who had been promoted but had never participated in tuition-
reimbursement and earned a degree (3.33). Tukey LSD tests indicated the
differences were significant (t, p < .05).
In summary, there was mixed support for the hypotheses regarding
participation in development activities and employee attitudes. Only one hypothesis
related to organizational commitment was supported. The number of days spent in
on-the-job training over the past year were significantly related to organizational
commitment. On the other hand, three out of four hypotheses predicting a
relationship between development and intention to turnover were supported. On-the-
(i
71
job training was associated with lower intention to turnover. Self-reported time
spent in tuition-reimbursement was related to higher intention to turnover. The
positive relationship between tuition-reimbursement and intention to turnover was
also supported by archival data on whether the employee had taken classes through
tuition-reimbursement, or had earned a degree through the program from 1996 to
2000. Finally, receiving a promotion was found to moderate the positive relationship
between tuition-reimbursement and intention to turnover.
72
IV. DEVELOPMENT AND EMPLOYEE RETENTION
The effects of employee development on employee retention are tested using
categorical analysis, logistic regression, and Cox proportional hazards regression
with turnover data for two different time periods provided by the company. First,
turnover was measured at the end of 2000 for the 979 survey respondents. Since
there were two waves of survey administration in, January and June, turnover was
effectively measured seven or eleven months after employees reported participation
in development activities. Second, archival data on employee turnover were
provided for all salaried employees from January 1996 to June 2000. The rates of
voluntary turnover or “quitting” for the survey and archival samples are presented in
Table 11. Rates are provided for the survey sample, the 9,243 employees who were
active at the beginning of 1996, and the entire sample of 12,360 current and former
o
employees who were active at any time from January 1996 through June 2000.
8 The 12,360 em ployees were used in a categorical analysis o f voluntary turnover. The 9,243
em ployees who were active at the beginning o f 1996 were used in the C ox regression analyses to
avoid including em ployees w ho were hired during the study period and m ay have joined the firm to
take advantage o f the generous tuition-reimbursement program.
73
Table 11
Rates of Voluntary Turnover Across the Samples
Quit Total
Survey Sample 4.8%a 979
Archival (Active 1996-2000) 9.5%b 12360
Archival (Active 1996) 8.4%b 9243
a7 or 11 Months b54 Months
As shown in Table 11, the effective voluntary turnover rate differs among the
three samples and is much greater in the 979 employee survey sample. Nearly 5% of
the employees surveyed in 2000 quit the company after less than a year. The
relatively higher rate of voluntary turnover in the survey sample is most likely due to
the large number of survey respondents who worked in the facility that was in the
process of closing. The relocation created turmoil within the company, which may
have led a higher number of employees to quit during these 11 months. There is also
a difference between the voluntary turnover rates of the full sample of 12,360
employees and the 9,243 employees who were active at the beginning of 1996. A
total of 8.4% of the employees who were active at the beginning of 1995 quit the
company over the 54-month study period, compared to 9.5% of all the employees
who were active at any time between January 1996 through June 2000. This
difference is significant (x2 , p < .001) given the large sample size, and indicates that
74
employees who were active in January 1996 were less likely to quit the company
than those who were hired afterwards.
For the sample of 9,243 employees active at the beginning of 1996, the rate
of voluntary turnover remained relatively stable over the 54-month study period and
averaged 1.9% annually. Voluntary turnover rates for these employees ranged
between 1.5% and 2.6% per year between 1996 and 2000. As described in the
methods section, this study examines only non-layoff voluntary turnover or “quits.”
Total turnover during this same time varied significantly and in 1997 and 1999
included a number of early retirements and voluntary layoffs that were not included
in these analyses. Refer to Table 5 for the actual company voluntary and non
voluntary turnover rates 1996-2000.
Results
On-the-job Training, Company Classes, and Voluntary Turnover
The first analysis of turnover tests Hypotheses 5, 6 and 9 using the self-
reported survey data on the three development activities to predict future turnover.
A logistic regression was estimated with data on the voluntary turnover of the 979
survey respondents seven to eleven months later. This regression uses the three In
transformed measures of days spent in company training classes, on-the-job training
and tuition-reimbursement over the past year. No significant relationships were
75
found between self-reported development time and voluntary turnover (see Appendix
10 for these results). This leaves Hypotheses 5 and 6 regarding company training
and on-the-job training and voluntary turnover unsupported.
Tuition-reimbursement and Voluntary Turnover
Preliminary analysis. The second analysis of employee development and
turnover tests Hypotheses 9 and 12 predicting a positive relationship between tuition-
reimbursement and quitting moderated by receiving rewards from the company.
This analysis uses Cox proportional hazards regression to examine archival measures
of tuition-reimbursement. Descriptive statistics for the entire sample of 12,360
employees and the 9,243 employees active in 1996 are presented in Appendix 4, with
correlations in Appendix 5. The Cox regression model is based on the assumption
that the independent variables and covariates affect turnover at a constant rate over
time (Hosemer, & Lemeshow, 1999; Morita, Lee, & Mowday, 1993). That is, the
hazard functions for groups of employees need to be proportional at each level of the
covariates. The hazard function is the conditional probability of an employee
quitting the company over time (Hosemer, & Lemeshow, 1999). The assumption of
proportionality is checked by comparing estimates of hazard functions graphed for
each level of the variables used to predict voluntary turnover.
76
Figure 3
Hazard Estimates for Earning a Degree vs. Not Earning a Degree
.06
.05
.04
.03
■ g
N 02
c c
x
a >
1 0 1
D
E
a o.oo
Earned Degree
D E a r n e d D e g r e e
0 N o D e g r e e
40 50 10 20 30 60 0
Months
For example, Figure 3 shows Kaplan-Meier hazard functions for the 9,243
employees active in 1996 categorized by whether or not they earned a degree
through tuition-reimbursement. The top line represents the probability at any given
time during the 54-month study period that one of the 8,360 employees who did not
earn a degree through tuition-reimbursement has quit the company, with all other
variables held constant. The bottom line in Figure 3 represents the probability that
one of the 883 employees who earned a degree through the company has quit the
company. The Cox regression model is relatively robust and only a significant
curvature or crossed lines on the graph would violate the assumption of
proportionality. No problems were detected for the variables used in the Cox
77
regressions (see Appendix 9 for graphs of the remaining variables). For additional
details on this method, refer to Morita, Lee and Mowday (1993).
Hypotheses tests. Figure 4 shows the voluntary turnover for all 12,360
employees active at any point between January 1996 and June 2000 categorized by
their level of participation in tuition-reimbursement, without controlling for other
variables. The results show mixed support for Hypothesis 9. Earning a degree
through tuition-reimbursement is positively related to voluntary turnover and
supports Hypothesis 9. Those who participated in the tuition-reimbursement and
earned a degree were significantly more likely to quit the company (11.9%) than
those employees who did not participate (10.6%) in tuition-reimbursement (x2, p <
.01). Figure 4 also shows, however, that the positive effect on quitting does not
materialize until the employee actually earns a degree. Counter to Hypothesis 9,
those employees who participated without earning a degree (8.3%) were significantly
less likely to quit the company than those who did not participate (10.6%) in tuition-
reimbursement (x2 , p < .01). In other words, participation in tuition-reimbursement
without earning a degree shows a negative relationship with voluntary turnover.
The two bars on the right-hand side of Figure 4 indicates that the 11.9% of
employees who quit the company after earning a degree can be better understood by
categorizing employees by whether or not they received a merit award or promotion
following their degree in addition to the $10,000 bonus paid by the company. Only
9.3% of those employees who received a either a promotion or merit award after
78
earning a degree eventually quit the company, which was significantly lower (X2, p <
. 05) than the 10.6% voluntary turnover rate for those who did not participate. The
rate of voluntary turnover for those not receiving a promotion or merit award after
earning a degree was 17.7%, making them the most likely to quit the company (x2 , p
< .001). This finding supports Hypothesis 12 that receiving a promotion or merit
award after earning a degree through tuition-reimbursement decreases the likelihood
that employees will quit the company.
Figure 4
Earning a Degree, Receiving Rewards, and Quitting the Company
o
o
o
C M
i
C O
O )
CD
0
>
o
c
3
(-
C O
c
J3
o
>
20%
18%
16%
14%
12%
10% - U
8%
6%
17.7%
11.9%
10.6%
~ ~ 8 .3% “
'
9.3%
Not Took Classes Earned Degree Earned Degree then
Participated (N=4825) (N=1089) Degree, No Reward*
(N=6446) Reward After* (N=762)
(N=327)
*Receiving a rew ard after earning a degree is defined as a prom otion o r m erit
aw ard in addition to the $10,000 bonus received by a ll em ployees who earned a
degree through tuition-reimbursement.
79
The archival measures of tuition-reimbursement are also analyzed using Cox
regression to test Hypotheses 9 and 12. Using time-to-tumover as the dependent
variable, Cox regressions are estimated using dummy variables for whether or not
the employee has taken classes without earning a degree, and whether the employee
has earned a degree through tuition-reimbursement. The Cox regressions include
control variables for job grade and years of tenure, and interactions between the two
tuition-reimbursement variables, promotions, and merit pay to test whether these
rewards moderate the effects of tuition-reimbursement on voluntary turnover. The
basic Cox proportional hazards regression model is specified in Equation 1. Ht is the
baseline hazard function with X ’ s as the predictors and interaction terms.
(1) H (t:x) — H (t) exp [P ](X C ontrols) Pl(-^took classes) / % (Xdegree)
P 4(X reward) + P$(XdegreeXreward) Ps(took classesXreward)]
Cox regressions were estimated using the 9,243 employees who were active
in January of 1996 when the company’s new tuition-reimbursement program was
introduced. Results of the first Cox regression models are presented in Table 12.
80
Table 12
Cox Regression Results
Step
1
B Exp(B) Wald
Step
2 a
B Exp(B) Wald
Step
2b
B Exp(B) Wald
Tenure3 -1.56 0.21 585.25 *** -1.72 0.18 697.59 *** -1.49 0.23 553.18 ***
Grade 1 0.52 1.69 3.38 + 0.62 1.86 4.70 * 0 .2 0 1.22 0.46
Grade 2 0.72 2.06 27.06 *** 0.83 2.28 34.73 *** 0.35 1.42 6.11 *
Grade 4 0.72 2.05 76.96 *** 0.61 1.84 55.48 *** 0.64 1.90 62.09 ***
Grade 5 -0.28 0.76 3.51 + -0 .0 2 0.98 0 .0 2 -0.27 0.76 3.32 +
Grade 6 0.11 1.11 0.40 0.21 1.24 1.53 0.09 1.09 0.27
Earned Degree 0.82 2.27 53.33 *** 0.79 2.19 20.54 *** 1.66 5.27 13.26 ***
Took Classes -0.98 0.37 138.80 *** - 1.1 0 0.33 108.55 *** -2 .2 0 0.11 48.79 ***
Promoted 1996-2000 -0 .6 8 0.51 10.62 ***
Took Class X Promotion 0.53 1.69 8 .8 6 **
Degree X Promotion -0.46 0.63 3.81 *
Merit Pay 1996-2000 -1.46 0.23
9 95 **
Took Classes X Merit Pay 1.65 5.22 25.23 ***
Degree X Merit Pay -1.30 0.27 7.64 **
Global X 937.18 *** 1106.48 *** 2853.23 ***
Sample Size 9243 9243 9243
Censored (%) 91.6% 91.6% 91.6%
“ standardized
+ P < . 10
*P<. 05
**P< .01
***/>< .0 0 1
Step 1 of the regression includes control variables for tenure and job grade
along with the two tuition-reimbursement dummy variables. The moderating effects
of receiving a promotion or merit award are then examined in steps 2a and 2b. In
81
step 2a a dummy variable for whether or not the employee had been promoted at any
time 1996-2000 is entered into the regression along with two multiplicative
interactions between receiving a promotion and the dummy variables for taking
classes without earning a degree, and earning a degree through tuition-
reimbursement. In step 2b the same procedure is used to examine the effects of
receiving merit pay at any time 1996-2000.
The significance of the overall model is tested using a global x2 statistic, and
the estimated values of 1,106 and 2,853 for the promotion and merit pay models are
significant (p < .001). The raw coefficients in Cox regression are interpreted as log
odds ratios. Exponentiated coefficients are also presented to ease interpretation.
Exponentiated coefficients greater than one indicate a positive relationship with
voluntary turnover, and coefficients less than one indicate a negative relationship.
The exponentiated coefficients for the tuition-reimbursement variables
estimated in Step 1 are significant (p <.001). Support for Hypothesis 9 predicting a
positive relationship between tuition-reimbursement and voluntary turnover is again
split between the two measures of participation. An exponentiated coefficient of
2.23 for earning a degree supports Hypothesis 9 and means that employees who
earned a degree through tuition-reimbursement quit at roughly twice the rate of those
who did not earn a degree over the study period. A coefficient of .37 for taking
classes without earning a degree, however, shows a negative relationship with
voluntary turnover. Employees taking classes through tuition-reimbursement
82
without earning a degree are one-third as likely to quit the company as non
participants.
Steps 2a and 2b test the interactions of tuition-reimbursement with promotion
and merit pay. Hypothesis 12 is supported by significant coefficients for the
interactions of both measures of tuition-reimbursement (taking classes and earning a
degree) with receiving a promotion (p < .001, p < .01) and receiving merit pay (p <
.001, p < .01). To interpret the effect of the interaction variables, the exponentiated
coefficients of the two main effects are multiplied together with the coefficient for
the interaction.9 For example, multiplying the three coefficients together yields .29
(.33pr0 g re s s x .51pro m o tio n x 1 • 2to o k c ia ssc sX p ro m o tio n ) for taking classes together with a
promotion, and .69 (2.19degree x .51pr0 tn o tio n x .63degreexprom otion) for earning a degree
together with promotion. This means that employees who participated in tuition-
reimbursement (either by taking classes alone or earning a degree) and received a
promotion are one-third to two-thirds as likely to quit the company as someone not
experiencing tuition-reimbursement or a promotion. Using the same method to
interpret the interactions of tuition-reimbursement and merit pay produces
coefficients of .13 (.11 to o k classes x . 2 3 m er itp a y x 5 .2 2 to o k c ia s s e s X m e r itp a y ) for taking classes
without earning a degree and .16 ( 5 . 2 7 d e g ree x .11 m e rit p a y X .27degreeXmeritpay) f o r e a rn in g
a degree. This means that employees who took classes and received a merit award
9 The coefficients are m ultiplied because they are exponentiated. The same result can be achieved by
summing the raw coefficients and taking the exponent o f the result.
83
were .13 to .16 as likely to quit as those not participating in tuition-reimbursement
and receiving a merit award.
These results indicate that receiving either a promotion or a merit award
moderates the effects of participating in tuition-reimbursement on voluntary
turnover. However, using dummy variables for whether or not a promotion or merit
award was ever received between 1996 and 2000 overlooks when these rewards were
actually given to the employee. That is, the interactions presented in Table 12 test
the effects of receiving a degree and receiving a promotion at any time during the 54-
month study period. Closer examination of the work histories of the employees who
earned degrees through tuition-reimbursement suggests a promotion or merit award
needs to follow a degree in order to have a positive effect on employee retention.
Tables 13 and 14 compare the voluntary turnover rates of employees who earned a
degree between 1996 and June 2000.
Table 13 suggests that a promotion must follow a degree in order to realize a
positive effect on retention. In fact, the employees with the highest voluntary
turnover (17.0%) were those who had already received a promotion when they
earned a degree through tuition-reimbursement, but were not promoted again
afterward. Table 14 shows that merit pay which follows a degree through tuition-
reimbursement has a positive effect on retention only if the employee had received
merit pay previously. Both tables indicate that the timing of receiving promotions or
84
merit awards relative to earning a degree through tuition-reimbursement is a
potentially important predictor of voluntary turnover.
Table 13
Timing of Promotion Relative to Earning a Degree and Quit Rates
Quit Total
Earned Degree without Promotion 11.8% 356
Promoted before Degree only 17.0% 377
Promoted after Degree only 8.6% 197
Promoted before and after Degree 3.7% 159
^,p< .001
Table 14
Timing of Merit Pay Relative to Earning a Degree and Quit Rates
Quit Total
Earned Degree without Merit Pay 21.0% 62
Merit Pay before Degree only 15.9% 347
Merit Pay after Degree only 17.7% 79
Merit Pay before and after Degree 7.8% 601
, p < ■ 001
The first Cox regression models used dummy variables indicating whether
the employee had received either a degree or a promotion at any time during the
study period and multiplicative interaction with receiving a promotion or merit
award at any time 1996-2000. Calculating the interaction terms differently by using
85
elapsed time between earning a degree and receiving a promotion or merit award
instead of a yes/no dummy variable allows the regression to test whether a promotion
or merit award that follows a degree has an additional impact on voluntary turnover.
It also tests whether the length of time between earning a degree and receiving a
promotion or merit award affects the likelihood that the employee will quit the
company.
A second set of Cox regressions are estimated with earning a degree through
tuition-reimbursement and receiving either a promotion or merit award following the
degree analyzed as time-dependent covariates (Morita, Lee, & Mowday, 1993). Step
1 of the regression includes control variables for company tenure and job grade and
the two dummy variables indicating whether the employee had taken classes without
earning a degree, or earned a degree through tuition-reimbursement. In steps 2a and
2b a dummy variable for whether the employee had received either a promotion or a
merit award 1996-2000 was entered into the model along with two interaction terms.
The rewards dummy variables test the effects of receiving either a promotion or a
merit award at any time during the study period. To test the effects of receiving a
reward after earning a degree, a time-dependent interaction is also entered. In step
2a that interaction measures the length of time between the degree date and receiving
a promotion. In step 2b that interaction measures the length of time between the
degree date and receiving a merit award. The interaction between the reward
variable and took classes without earning a degree was calculated the same as before.
86
Table 15
Time-Dependent Cox Regression Results
Step
1
Step
2a
Step
2 b
B Exp(B) Wald B Exp(B) Wald B Exp(B) Wald
Tenure3 -1.57 0 .2 1 589.56 *** -1.73 0.18 698.22 *** -1.51 0.22 569.15 ***
Grade 1 -0.05 0.95 0.16 -0.10 0.91 0.56 -0.20 0.82 2.32
Grade 2 0.69 1.99 23.38 *** 0.79 2 .2 1 30.80 *** 0.24 1.27 2.70 ***
Grade 3 0 .6 8 1.98 61.67 *** 0.57 1.78 42.89 *** 0.58 1.78 43.25 ***
Grade 4 -0.31 0.73 4.15 * -0.08 0.92 0.28 -0.32 0.72 4.56 *
Grade 5 0.08 1.08 0 .2 0 0.18 1.19 1.03 0.05 1.05 0.07
Took Classes -0.98 0.38 137.46 *** -1.05 0.35 100.40 *** -2.04 0.13 42.36 ***
Earned Degree 0.81 2.24 51.81 *** -0.61 0.54 6.75 ** -0.50 0.61 2.35
Promoted 1996-2000 -0.95 0.39 92.45 ***
Took Class X Promotion 0.53 0.35 1.42
Time to Promotion 0.04 1.04 25.23 ***
Merit Pay 1996-2000 -2.61 0.07 664.55 ***
Took Class X Merit Pay -0.01 0.02 1 .0 2
Time to Merit Pay 1.43 4.20 19.40 ***
Global X 1088.91 *** 1127.88 *** 2365.30 ***
Sample Size 9243 9243 9243
Censored (%) 91.6% 91.6% 91.6%
“standardized
+ P < .10
* P < ,05
**P< .01
***P< .001
Results of the time-dependent Cox regression models are presented in Table
15. Estimated x2 measures of 1,088 for the promotion model and 2,365 for the merit
pay model are significant (p < .001). In step 1 of the regression, dummy variables
for taking classes without earning a degree and earning a degree are again significant
in opposite directions. The coefficients indicate a positive relationship between
earning a degree and quitting, and a negative relationship between taking classes
without earning a degree and quitting. In steps 2a and 2b the interactions between
taking classes and receiving either a promotion or merit award were not significant.
The coefficients for the time-dependent interactions between earning a degree and
receiving a promotion in step 2a and a merit award in step 2b were both significant
(p < .001). Significant coefficients for these interaction terms mean that the amount
of time that elapsed between earning a degree and receiving either a promotion or
merit award had an effect on quitting beyond the effect of receiving a promotion or
merit award. Exponentiated coefficients of 1.04 for the time elapsed between
earning a degree and receiving a promotion and 4.20 for the time elapsed between
earning a degree and receiving a merit award indicate that the likelihood of quitting
increases each month that passes after an employee earns a degree through tuition-
reimbursement and does not receive either a promotion or merit award.
The coefficients of the time-dependent interactions in step 2a and 2b are
interpreted by multiplying them with the coefficients for main effects of earning a
degree 1996-2000, and receiving a reward (either a promotion in step 2a, or a merit
88
award in step 2b), for any given value of the time-dependent variable. For example,
an employee who earned a degree and was promoted one month later was .21
( . 5 4 eam ed degree X •3 9 p ro m o tio n X (1) 1. 0 4 mOnths to prom otion) likely tO C juit the C o m p a n y 3S
an employee who had not earned a degree or promotion. An employee who earned a
degree and was promoted ten months later, on the other hand, was 2.10 ( . 5 4 eanied degree
X .3 9 p ro m o tio n x ( 1 0 ) 1 . 0 4 m Onths to prom otion) as likely to quit the company as an employee
who had not earned a degree or a promotion. Using the same method, similar results
were found for employees who earned a degree through tuition-reimbursement and
subsequently received merit pay. An employee who earned a degree was . 17 as
likely to quit as an employee who did not earn a degree or receive a merit award if
the merit payment was within one month after earning the degree. The same
employee who waits ten months after earning a degree to receive a merit award is
1.70 as likely to quit the company as employees who did not earn a degree or receive
a merit award. These findings indicate that earning a degree through tuition-
reimbursement initially has a negative effect on voluntary turnover, which eventually
turns positive and continues to increase as the months elapse and the employee does
not receive either a promotion or merit award.
In summary, two of the four hypotheses relating employee development to
voluntary turnover were supported. Analysis of voluntary turnover using the survey
self-reported training time data yielded no significant findings, leaving Hypotheses 5
and 6 predicting negative relationships between company-classes and on-the-job
training with voluntary turnover unsupported. These results might be explained in
part by the circumstances under which the survey data were collected and turnover
measured. First, a large number of the survey respondents worked at a facility that
was in the process of closing. These employees may have been more likely to quit
the company since many employees were looking for work outside the company to
avoid relocation or layoff. Second, turnover was measured less than a year later,
which is much shorter than the 54-month study period used in the Cox regressions.
This might indicate that seven to eleven months is likely too short of a time period to
see the retention effects of worker participation in employee development. This is
particularly true in firms such as the one studied with relatively low rates of
voluntary turnover.
Analysis using the archival tuition-reimbursement data generally supports
both Hypothesis 9 and Hypothesis 12. In the case where an employee has earned a
degree through tuition-reimbursement, participation has a positive effect on
voluntary turnover, and receiving either a promotion or merit award following the
degree moderates this positive effect. Employees who earned a degree through
tuition-reimbursement and received a subsequent promotion or a merit award were
the least likely to quit the company over the 54-month study period. The results
show that taking classes without earning a degree has a negative effect on voluntary
turnover. When an employee graduates and earns a degree, however, the likelihood
that they will quit the company begins to increase steadily every month until the
90
employee receives either a promotion or a merit award. In fact, the employees who
earned a degree through tuition-reimbursement, but failed to received a promotion or
merit award afterwards, were by far the most likely to quit the company. These
findings for the importance of receiving a promotion or merit award after the degree
are particularly robust given the fact that all employees who earn degrees through
tuition-reimbursement have already been rewarded with a $10,000 bonus.
91
V. DISCUSSION
Analysis of employees from this one company provides strong evidence for a
relationship between employee development, attitudes, and voluntary turnover. The
results show that development activities play an important role in how employees
regard their employers and ultimately their decisions to remain with the company.
Most importantly, this study demonstrates that training and development can have
both positive and negative effects on employee attitudes and retention depending on
the type of activity and whether the employee is rewarded afterwards through a
promotion or merit award. Participation in on-the-job training was related to higher
organizational commitment and lower intention to turnover. On the other hand,
taking classes through tuition-reimbursement was related to higher intention to
turnover, and higher voluntary turnover for those employees who earned degrees.
Finally, employees who earned degrees and then received a promotion or merit
award were the least likely employees to quit the company.
These findings suggest that company-sponsored development plays a unique
role in the employment relationship by acting simultaneously as an element of both
social and economic exchange. In the case of on-the-job training, development is
92
clearly a benefit like other “employee-friendly” HR practices and has positive effects
on attitudes as employees reciprocate through social exchange. Tuition-
reimbursement, on the other hand, has negative affects on employee attitudes and
retention. This does not mean, however, that tuition-reimbursement is not a benefit
enjoyed by employees. Providing tuition-reimbursement to employees may translate
directly into positive attitudes and retention because tuition-reimbursement is not
only a benefit and investment in employees, it also creates economic incentives by
acting on employees’ perceptions of their career opportunities and what they should
receive in return for their skills.
This dissertation suggests that the degree to which the forces of social
exchange and economic exchange predominate over attitudes and retention of
employees is based on the general versus specific nature of the skills developed.
More firm-specific development, such as on-the-job training, creates fewer external
job opportunities. This means that only the social exchange effects of providing
development come into play. Employees therefore respond with higher commitment
and lower intention to turnover. General skills development, such as tuition-
reimbursement, is likely to increase perceptions of external job opportunities.
External job opportunities lead to withdrawal and turnover (Bateman, & Strasser,
1984; Gerhart, 1990; Kirschenbaum, & Mano-Negrin, 1999; O’Reilly, & Caldwell,
1981). The results indicate that the economic incentives created by new general
93
skills tend to outweigh any potential social exchange benefits of providing tuition-
reimbursement to employees.
Evidence of the economic incentives created by participating in tuition-
reimbursement comes from the differences in voluntary turnover between those
employees who took classes without earning a degree and those employees who
earned a degree through tuition-reimbursement. Economic signaling suggests that a
degree provides more information to potential employers regarding an employee’s
future productivity than simply taking classes (Spence, 1974). Employees who took
tuition-reimbursement classes reported a significantly higher intention to turnover
than employees who did not take classes, but actual voluntary turnover did not
increase until employees earned a degree. This suggests that the marketability of the
skills gained through tuition-reimbursement increases significantly when an
employee graduates.
If tuition-reimbursement increases intention to turnover and leads to
increased voluntary turnover when an employee earns a degree, the question
becomes why would companies want to pay to facilitate the exit of these employees?
First, the retention effect when the employee is taking classes might be enough alone
to justify company expenditure on tuition-reimbursement. Even though willingness
to turnover increases with taking classes, tuition-reimbursement only increases
voluntary turnover after graduation. Since employees who are actually taking
classes are significantly less likely to leave the company, tuition-reimbursement
94
programs appear to be an effective strategy to keep employees for the time that is
takes an employee to complete a degree. This retention effect may be augmented in
this case since the $10,000 bonus provides a strong incentive to stay until the degree
is completed.
Second, this dissertation finds that developing general skills through tuition-
reimbursement does not necessarily have negative effects on employee retention. A
company can retain employees who participate in tuition-reimbursement by
rewarding them with a promotion or merit award. The results show that the effects
of tuition-reimbursement on intention to turnover and actual voluntary turnover are
moderated in such a way that participating in tuition-reimbursement and receiving a
promotion or merit award increases employee retention. This was true both for
employees who took classes without earning a degree and those who earned a degree
through tuition-reimbursement.
The findings regarding the moderating effects of receiving either a promotion
or merit award after participating in tuition-reimbursement suggest that receiving
rewards has multiple effects. To begin, a promotion or merit award helps to
neutralize economic incentives by reducing the attractiveness of other potential jobs.
This suggests that companies can compete with alternative employers to keep their
trained employees by rewarding them through promotion or merit pay. In addition,
receiving a promotion or merit award also acts as a positive signal to employees that
the company values them. The interaction between participating in general skills
95
development and then receiving a promotion or merit award is likely to generate
feelings of support from the company, and translate into increased employee
retention. In fact, the results show that employees who earned a degree through
tuition-reimbursement and then received a merit raise or promotion (in addition to
the $10,000 bonus) were the least likely to quit the company. These employees were
less likely to voluntarily leave the company than employees who never participated
in the program (including those who received either a promotion or merit award
without participating in tuition-reimbursement). This suggests that additional social
exchange benefits accrued when earning a degree is followed by a promotion or
merit award.
The outcomes of the various hypotheses tests are summarized in Table 16.
The individual findings (e.g., support for Hypothesis 2 and Hypothesis 9) from this
dissertation generally fit with previous research from both psychology (Birdi et al.,
1997) and human capital economics (Lynch, 1991) regarding the outcomes of on-
the-job training and tuition-reimbursement on the attitudes and retention of
employees. More importantly, this dissertation suggests that the main perspectives
on the likely outcomes of providing development to employees can be integrated into
a single theoretical model. Findings for Hypotheses 11 and 12 support this
integrative model by demonstrating that receiving a promotion or merit award
moderates the relationships between participation in tuition-reimbursement, intention
to turnover, and quitting the company. Although there are several potential
96
limitations to this study, these results have important implications for multiple
literatures on training and development, as well as company practice.
Table 16
Summary of Hypotheses and Findings
Hypos Method Developm ent
A ctivity
Hypothesized
Relationship
Predicted
Outcome
Findings
1 OLS Co. Classes Positive Commitment N ot
Supported
2 OLS OJT Positive Commitment Supported
3 OLS Co. Classes Negative Intention
To Turnover
N ot
Supported
4 OLS OJT N egative Intention
T o T urnover
Supported
5 Logistic Co. Classes N egative Turnover N ot
Supported
6 Logistic OJT Negative Turnover N ot
Supported
7 OLS Tuition-
reimbursement
N egative Commitment N ot
Supported
8 OLS Tuition-
reimbursement
Positive Intention
T o Turnover
Supported
9 Cox Tuition-
reimbursement
Positive Turnover Supported
10 OLS Tuition-
reimbursement
Moderated by
Rewards
Commitment N ot
Supported
1 1 OLS Tuition-
reimbursement
Moderated by
Rewards
Intention
To Turnover
Supported
1 2 Cox Tuition-
reimbursement
Moderated by
Rewards
Turnover Supported
97
Theoretical Contribution
The effects of development on employee attitudes and retention depend on
the nature of the skills developed and whether or not the employee receives a merit
award or promotion after he or she takes classes or earns a degree through tuition-
reimbursement. The results suggest that contemporary models of employee training
and development from management and psychology should be better equipped to
handle outcomes that go beyond learning and transfer of skills to the job. To address
the effects of development on employee attitudes and turnover, models of T&D
should include two key dimensions for future research. First, employee development
activities should be examined in terms of the skill that they provide and the
marketability of those skills to other companies. Second, future theory and research
need to include the effects of rewarding or not rewarding employees after they
participate in general skills development.
Development models in the management literature such as Birdi et al. (1997)
argue that the level of benefit that a development activity provides affects social
exchange and whether the employee responds with positive attitudes towards the
company. Birdi et al. (1997) attributed differences among development activities’
effects on employee attitudes to the amount of work time versus personal time
expended by the employee. This dissertation generally agrees with the previous
findings of Birdi et al. (1997) regarding the effects of development on organizational
commitment with a notable exception. Birdi et al. (1997) found effects for company
98
classes taking place on work time, but not on-the-job training. Findings from this
dissertation included a significant relationship between organizational commitment
and on-the-job training, but not company classes. Both studies suggest that firm-
specific training that takes place during work hours has positive effects on employee
attitudes through social exchange, but the exact nature of the differences between on-
the-job training and company-based classes merits additional future research and
theoretical development.
When participation in tuition-reimbursement (which develops more general
skills) is examined, however, it is also clear that development can have negative
effects on employee retention. Both earning a degree and taking courses through
tuition-reimbursement without earning a degree are positively related to intention to
turnover. Earning a degree is also positively related to actual voluntary turnover.
Until now, management and psychology models of the outcomes of training and
development have not adequately acknowledged the potential for employees to quit
the company for better jobs when they develop new skills.
Developing skills that are in demand at other companies may have negative
effects on retention, because they increase employees’ perceptions of their external
job opportunities. Dimensions such as “current vs. future oriented” (Noe et al.,
1997) and “job vs. non-job focus” (Birdi et al., 1997) differentiate development
based on the type of skills gained. However, these dimensions do not adequately
address whether the skills gained through development are useful and marketable to
99
other companies. This dissertation implies that the general versus specific nature of
the skills needs to be included in future models of the outcomes of training and
development.
This dissertation also suggests that models of training and development
should include whether or not employees are rewarded when they develop new
general skills. Much of the training and development literature already argues for
viewing development and development outcomes in the larger context of the
employee’s work history and environment. For example, models of training
effectiveness that include the transfer of learned skills to the job recognize that skills
transfer is dependent on the work environment (Facteau, Dobbins, Russell, Ladd, &
Kudisch, 1995; Ford, Quinones, Sego, & Sorra, 1992; Rouiller, & Goldstein, 1993).
Similarly, when predicting the outcomes of development on commitment and
retention, what happens to the employee following his or her participation should be
considered.
In terms of the economics literature, the results validate several important
predictions from human capital theory. Most importantly, general skills
development in the form of earning a degree through tuition-reimbursement was
found to significantly increase voluntary employee turnover. This study of detailed
employment histories in one company confirms the finding from large-scale
individual survey data that tuition-reimbursement, or “off-the-job” training, has a
positive relationship with employee turnover (Lynch, 1991). The differences in
100
voluntary turnover between those employees who earned a degree versus those who
took classes without earning a degree may also demonstrate the importance of
economic signaling theory (Spence, 1974). The ability of any particular
development activity to increase the productivity of an employee is difficult for other
companies to discern. A college degree is therefore a stronger signal since it
provides some indication of future potential. The voluntary turnover results from
this study suggest that the marketability of skills gained through tuition-
reimbursement is likely to be accrued upon graduation.
The findings of this dissertation also inform a central tension in the
economics literature on training and development. This is the long-standing
discrepancy between the human capital theory and company practice regarding the
prediction that employers will not provide general skills training unless wages are
decreased. Employers regularly provide large-scale, company-financed general
skills development such as tuition-reimbursement programs without lowering trainee
wages (Bishop, 1997; Frazis, Herz, & Horrigan, 1995). Findings from this
dissertation offer some potential explanation as to why companies might offer
general skill development to employees. One reason is that participation in tuition-
reimbursement lowers voluntary turnover while the employee is in school. This
means that tuition-reimbursement provides an incentive to stay with a firm for the
period of time (typically several years of part-time study) that it takes an employee to
complete a degree. A second reason is that providing development may include
101
benefits derived through social exchange that are not accounted for in human capital
models. Employees who gained general skills through tuition-reimbursement and
then received either promotion or merit award afterward were the least likely to quit
the company. Moreover, they were less likely to quit the company than employees
who received a promotion or merit award without participating in tuition-
reimbursement. This may be because the interaction of participating in tuition-
reimbursement (with or without earning a degree) and receiving either a promotion
or merit award creates positive attitudes by acting as a signal to employees that they
are valued. When the company signals that it is interested in an employee’s
development and career, it fosters perceptions of organizational support and the
employee responds in kind with additional loyalty to the company.
As Bishop (1997, 51) summarizes the evidence on employer-provided
training: “If employers are paying some of the costs of general training, they are not
doing it for altruistic reasons. They are comparing the training costs incurred to the
expected productivity benefits the firm will receive from the workers who stay at the
firm.” Employers may be counting on trained employees staying with the company
because they are satisfied and committed to the organization that provided them with
the skills. It may be that companies are able to provide general skills training
without decreasing wages because social exchange translates into lower intention to
turnover and higher employee retention if employees receive either a promotion or
merit award after they graduate. Although rewarding employees through a
102
promotion or merit award entails additional costs to the employer, it also reduces the
risk of turnover and increases the chances of recouping the investment in the
development of an employee.
Taken together, the findings also fit well with the existing literature on
employee turnover. Two main theoretical frameworks have dominated the turnover
literature. First, the more common models of employee turnover use job satisfaction
and perceived ease of movement to predict voluntary turnover (Bluedom, 1982;
March, & Simon, 1958; Mobley et al., 1979). Mapping the effects of employee
development onto these models of voluntary turnover illustrates how the social and
economic effects of providing skill development might work simultaneously. From
the social exchange perspective, development is likely to have a negative effect on
turnover by increasing organizational commitment and job satisfaction. For the
economic exchange perspective, some employee development activities are likely to
increase voluntary turnover by enhancing perceived ease of movement.
The second type of turnover theory views turnover as a process that takes
place or “unfolds” over time (Lee, & Mitchell, 1994). In the unfolding model, an
event or “shock” to employees causes them to reflect on their employment situation.
As employees react to the shock, they tend to follow distinct “paths” towards
voluntary turnover. One of these paths includes initiating a search process.
Conceptualizing turnover as a process allows the likelihood that an employee will
quit the company to vary over time depending on what happens to the employee.
103
Similarly, the effects of participating in general skills development on attitudes and
retention can be conceptualized as variable depending on certain critical events or
“shocks.” These critical events might include earning a degree, or receiving a
reward such as a promotion or merit award. When an employee earns a degree
through tuition-reimbursement, the immediate effect on employee retention appears
to be positive. As time goes on, however, the likelihood that an employee will quit
the company after earning a degree through tuition-reimbursement increases every
month that the employee does not receive either a promotion or merit award.
Limitations of the Research
Several limitations to this study should inform the interpretation of the
findings. First, there are limitations from the research design and measures used to
test the hypotheses. A longitudinal study with multiple measurements of
organizational commitment and intention to turnover would provide a more accurate
picture of the effects of development participation on employee attitudes. The
literature on training generally follows a model in which organizational commitment,
self-efficacy, and training motivation are treated as both antecedents and outcomes
of participation in development (Gist, 1987; Tannenbaum et al., 1991; Tannenbaum,
& Yukl, 1992; Wexley, & Latham, 2001). A one-time measure of organizational
commitment and intention to turnover precludes any measurement of change in those
variables directly attributed to the development experience.
104
In addition to measuring the changes in commitment over time, future
research should also address several potential problems with measurement error in
the variables studied. Measures of employee attitudes might include a job
satisfaction scale and a more complete measure of organizational commitment.
Although organizational commitment and intention to turnover formed independent
scales, the fact that three items were dropped from a shortened version of Porter et
al.’s (1974) scale to eliminate cross-loading with intention to turnover suggests that
the measure contains element of both affective commitment and intention to remain .
Additional measures could also capture the different dimensions of organizational
commitment (affective, continuance, and normative) and ease the comparison of
findings with a wider array of commitment research (Meyer, & Allen, 1997; Meyer
& Smith, 2000). Another potential problem with measurement error comes from the
use of self-reported training time. Self-reports suffer from the difficulty employees
have in recalling all development experiences over the past twelve months.
Although tuition-reimbursement was measured using both archival and self-report
data, only self-report data was available for on-the-job training and company-classes.
A more reliable archival measure of participation in these activities may yield more
significant relationships with attitudes and voluntary turnover.
Second, there are potential limitations to the generalizability of the findings
due to the nature of the company and the employee sample studied. Most
importantly, the company experienced substantial downsizing in recent history,
105
which may have affected how employees responded to training and development.
Studying the company during downsizing raises questions about how the conditions
affect the relationship between development, attitudes and retention. When
employees are concerned about being laid off, things other than skill development
may drive their attitudes and decision to quit. On the other hand, the loss of job
security may have heightened the employee’s awareness of the benefits of increasing
skills through voluntary development activity.
Downsizing also meant there were large numbers of layoffs and early
retirements, followed by new hires and rehires. This has several potential effects on
the findings. First, many of the layoffs were associated with the facility closure that
took place during 2000, which is likely to have residual effects on attitudes and
voluntary turnover of the survivors of this downsizing. A site variable was used in
the analysis of survey data to control for some of these effects, but no such control
was available for the archival data analysis. Second, the early retirements and
“voluntary separations” that took place during this time were not included in the
analysis as voluntary turnovers. While these employees are clearly different than
those quitting the company, they still represent voluntary departures from the
company and could be explored with further analysis of these data. These other
voluntary leavers represented 10% of the turnover that occurred January 1996
through June 2000. Third, there were a large number of employees that were
temporarily laid-off and rehired during the study period. They represented nearly
106
12% of the salaried employees who were active at some point during 1996 - June
2000, and 5% of employees active in 1996. These employees were excluded from
the analyses because the experience of the layoff would likely have far greater
effects on their attitudes and decisions to remain with the company than their
development experiences.
Potential limitations in generalizability of the findings also arise from the
sample of employees that were studied. The employees studied were full-time
salaried engineers and technical managers. Nearly all were college educated with a
relatively large number of employees having advanced graduate degrees. This
suggests that these employees are likely to place a greater value on keeping their
skills current and therefore participate in voluntary development activities more
often than non-technical or non-salaried employees (Dubin, 1990; Farr, &
Middlebrooks, 1990; Kozlowski, & Farr, 1988; Kozlowski, & Hults, 1987).
Additional research needs to be done to understand whether the effects found in this
dissertation are widely generalizable, especially to lower-skilled employees.
Finally, replicating these findings across companies depends on company
policies and factors in the work environment, such as perceptions of manager and
peer support that are significant predictors of participating in development activities
(Maurer, & Tarulli, 1994; Noe, & Wilk, 1993). The company studied in this
dissertation places an unusually high priority on offering development opportunities
and encouraging employees to participate in company training. The company has an
107
generous tuition-reimbursement program, which is widely seen as a generous benefit
in the company. Although nearly all companies provide some type of training to
employees and some 75% offer some type of tuition -reimbursement benefit, there is
a wide variety in the administration of employee development and the level of
support that is offered to employees (Shapiro, 2001).
Questions for Future Research
This dissertation sheds new light on the effects of training and development
on the commitment and retention of employees, but also raises several questions.
Some of the findings suggest that additional research is required to understand
exactly how employees perceive development and how they respond afterward. One
unexpected finding was the difference between company-sponsored training classes
and on-the-job training in predicting employee attitudes. On-the-job training showed
significant relationships with both organizational commitment and intention to
turnover while participation in company classes did not. This finding raises
theoretical questions because these development activities both take place
completely on company time and deliver skills that are similar in terms of their
future marketability. This means that another aspect of these two training activities
needs to be considered in predicting how employees perceive development as a
benefit from companies.
108
One potential explanation is that employees are more likely to respond
positively to company practices that are discretionary towards individual workers
(Eisenberger et al., 1989). On-the-job training might be more discretionary from the
perspective of the employee in that the training program is targeted to the individual,
and is not available to all employees. Research also suggests that the outcomes of
training are affected by whether the activity is voluntary (Hicks, & Klimoski, 1987)
and whether it is effective (Baldwin, Magjurka, & Loher, 1991). Additional research
is needed to understand whether the differences between company classes and on-
the-job training follow from any of these factors or other unmeasured variables.
A second unexpected finding was the difference between the effects of merit
pay and promotion as moderators. Receiving a promotion showed more consistent
and larger effects on intention to turnover and quitting than receiving merit pay. The
stronger effects of receiving a promotion might be explained in that a promotion is
associated with a larger wage increase than a merit award. Previous research has
suggested that the relationship between promotion and turnover is driven primarily
by the wage increases that customarily accompany promotions (Trevor, Gerhart, &
Boudreau, 1997). However, the fact that promotion and merit awards were
significant even after an employee received the $10,000 bonus for earning a degree
suggests that the effects of these rewards may be driven by more than money. The
findings raise the question as to whether career advancement might be the critical
factor in the moderating effects of receiving a merit award or promotion after earning
109
a degree. Additional research is needed to understand whether career advancement
is a more important moderator of employee outcomes than wages. This issue is
critical given that employers cannot promote all employees who earn degrees
through tuition-reimbursement and need to find ways to reward employees for their
new skills.
Finally, the findings show that a promotion or merit award that actually
follows a degree has greater effects on attitudes and retention than a promotion or
merit award received by the employee before or dining participation in tuition-
reimbursement. This raises questions as to how employees perceive the rewards that
they receive. This research suggests that employees need to perceive that they are
being rewarded “for” new skills in order for them to respond positively with
commitment and willingness to remain, but more research is needed to understand
how that differs from receiving rewards in general that may or may not be related to
their development experience.
Conclusion
This dissertation is particularly relevant given that the time that employees
spend in formal development has increased steadily over the past ten years (Van
Buren, 2001). More companies are emphasizing self-development for employees,
and tying T&D into annual reviews is becoming more common (Douglas, &
110
McCauley, 1999; Meister, 1998). This dissertation suggests that companies need to
re-examine these current employee development programs in terms of their potential
effects on attitudes and retention. Prior to this study, research on development as a
retention tool consisted primarily of a small number of case studies concluding that
development has a universally positive effect on employee commitment and
retention (“Recruiting, ’’ 1999). This dissertation demonstrates that there are several
important conditions that affect how an employee reacts to development, and that the
reaction may not necessarily be positive. If companies wish to keep employees by
providing them with skills development, they must also follow-up by recognizing
and rewarding new skills.
Twenty years ago Bartol (1979) found that computer technicians who were
more motivated to work on their careers tended to turnover less. In contrast,
Cappelli (2000) suggested that in today’s labor market computer technicians might
be forming the vanguard of a new breed of employee who develops his or her career
by moving frequently among companies and making employment decisions based on
the developmental potential of the work. This dissertation recognizes the evolving
and growing role that employee development plays in today’s business organizations
and presents a practical model for academic researchers, employees and companies
to cope with these changes. The findings provide strong backing for differences
among development activities based the general versus specific nature of the skills
developed. The findings also substantiate the role of rewards in determining the
Ill
retention effects of providing general skills. These new ways of thinking about the
relationship between employee development, attitudes and retention merit future
research and recognition within current models of the outcomes of company-
sponsored training and development.
112
BIBLIOGRAPHY
Acemoglu, D., & Pischke. 1999. Beyond Becker: Training in imperfect labour
markets. Economic Journal, 109, 12-42.
Adams, J. 1965. Inequity in social exchange, In L. Berkowitz (Ed.), Advances in
Experimental Social Psychology, New York, NY: Academic Press.
Allen N., & Meyer, J. 1996. Affective, continuance, and normative commitment to
the organization: An examination of construct validity. Journal of Vocational
Behavior, 49, 252-276.
_________________ . 1990. The measurement and antecedents of affective,
continuance, and normative commitment to an organization. Journal of
Occupational Psychology, 63, 1-18.
Baldwin, T., Magjurka, R., & Loher, B. 1991. The perils of participation: Effects of
choice of training on employee motivation and learning. Personnel
Psychology, 44, 51-65.
Bartel, A. 1989. Formal employees training programs and their impact on labor
productivity: Evidence from a human resource survey [NBER Working
Paper] (3026). Cambridge, MA.
_________________ . 1995. Training, wage growth, and job performance: Evidence
from a company database. Journal of Labor Economics, 13(3), 401-426.
Bartol, K. 1979. Professionalism as a predictor of organizational commitment, role
stress, and turnover: A multidimensional approach. Academy of Management
Journal, 22 (4), 815-837.
Bateman, T., & Strasser, S. 1984. A longitudinal analysis of the antecedents of
organizational commitment. Academy of Management Journal, 27, 95-112.
Becker, G. 1965. Human Capital. Chicago, II: University of Chicago Press.
Begley, T., & Czajka, J. 1993. Panel analysis of the moderating effects of
commitment on job satisfaction, intent to quit, and health following
organizational change. Journal of Applied Psychology, 78(4), 552-556.
113
Benkhoff, B. 1997. Ignoring commitment is costly: New approaches establish the
missing link between commitment and performance. Human Relations, 50(6),
701-726.
Birdi, K., Allan, C., & Warr, P. 1997. Correlates and perceived outcomes of four
types of development activity. Journal of Applied Psychology, 82(6), 845-
857.
Bishop, J. 1997. What we know about employer-provided training: A review of the
literature. Research in Labor Economics, 16, 19-87.
Blau, G. 1964. Exchange and Power in Social Life. New York, NY: Wiley.
_________________ . Testing the generalizabiltiy of career commitment and its
impact on employee turnover. Journal of Vocational Behavior, 12, 88-103.
Blau, G., & Boal, K. 1987. Conceptualizing how job involvement and organizational
commitment affect turnover and absenteeism. Academy of Management
Review, 12, 288-300.
Bluedom, A. 1982. A unified model of turnover from organizations. Human
Relations, 35(2), 135-153.
Bowen, B., & Schneider, D. 1999. Understanding customer delight and outrage.
Sloan Management Review, 35-45.
Cappelli, P. 2000. A market-driven approach to retaining talent. Harvard Business
Review, 78(1), 103-111.
Camevale, A., Gainer, L., & Villet, J. 1990. Training in America. San Francisco,
CA: Jossey-Bass.
Cohen, A. 1993. Organizational commitment and turnover: A meta-analysis.
Academy of Management Journal, 36(5), 1140-1157.
Cotton, J., & Tuttle, J. 1986. Employee turnover: A meta-analysis and review with
implications for research. Academy of Management Review, 11(1), 55-70.
Davy, J., Kinicki, A., & Scheck, C. 1997. A test of job security’s direct and mediated
effects of withdrawal cognitions. Journal of Organizational Behavior, 18,
323-349.
114
Douglas, C., & McCauley, C. 1999. Formal developmental relationships: A survey
of organizational practices. Human Resource Development Quarterly, 10(3),
Fall, 20-796.
Dubin, S. 1990. Maintaining competence through updating. In S. Willis, & S. Dubin
(Eds.), Maintaining Professional Competence (pp. 9-43). San Francisco:
Jossey-Bass.
Eisenberger, R., Huntington, R., Hutchison, S., & Sowa, D. 1986. Perceived
organizational support. Journal of Applied Psychology, 71, 500-507.
Facteau, J., Dobbins, G., Russell, J., Ladd, R., & Kudisch, J. 1995. The influence of
general perceptions of the training environment on pretraining motivation and
perceived training transfer. Journal of Management, 21(1), 1-25.
Farr, J., & Middlebrooks, C. 1990. Enhancing motivation to participate in
professional development. In S. Willis, & S. Dubin (Eds.), Maintaining
Professional Competence (pp. 195-213). San Francisco: Jossey-Bass.
Farrell, D., & Rusbult, C. 1981. Exchange variables as predictors of job satisfaction,
job commitment and turnover: The impact of costs, alternatives, and
investments. Organizational Behavior and Human Performance, 27, 78-95.
Feldman, D. 1989. Socialization, resocialization, and training: Reframing the
research agenda. In I. Goldstein (Ed.), Training and Development in
Organization. San Francisco: Jossey-Bass.
Fishbein, M., & Ajzen, I. 1975. Belief attitudes, intention and behavior. Reading,
MA: Addison-Wesley.
Ford, J., Quinones, M., Sego, D., & Sorra, J. 1992. Factors affecting the opportunity
to perform trained tasks on the job. Personnel Psychology, 45, 511-527.
Frazis, H., Herz, D., & Horrigan, M. 1995. Employer-provided training: Results
from anew survey. Monthly Labor Review, 118(5), May, 3-17.
Fujita-Starck, P. 1996. Motivations and characteristics of adult students: Factor
stability and construct validity of the educational participation scale. Adult
Education Quarterly, 47(1), 29-40.
115
Gaertner, K., & Nollen, S. 1989. Career experience, perceptions of employment
practices, and psychological commitment to the organization. Human
Relations, 42(11), 975-991.
Gardner, T., Moynihan, L., Park, H., & Wright, P. 2000. Unlocking the black box:
Examining the processes through which human resource practices impact
business performance. Academy of Management Annual Meeting. Toronto,
Canada, August.
Gerhart, B. 1990. Voluntary turnover and alternative job opportunities. Journal of
Applied Psychology, 75(5), 467-476.
Gist, M. 1987. The influence of training method on self-efficacy and idea generation
among managers. Personnel Psychology, 787-805.
Goldstein, I. 1989. Critical training issues: Past, present, and future. In I. Goldstein
(Ed.), Training and Development in Organizations. San Francisco: Jossey-
Bass.
Gouldner, A. 1960. The norm of reciprocity: A preliminary statement. American
Sociological Review, 25, 161-178.
Greenberg, J. 1988. Equity and workplace status: A field experiment. Journal of
Applied Psychology, 73, 606-613.
Griffeth, R., Horn, P., & Gaertner, K. 2000. A meta-analysis of antecedents and
correlates of employee turnover: Update, moderator tests, and research
implications for the next millennium. Journal of Management, 26, 463-488.
Gupta, N., & Jenkins, G. 1991. Rethinking dysfunctional employee behaviors.
Human Resource Management Review, 1, 39-59.
Guzzo, R., & Noonan, K. 1994. Human resource practices as communications and
the psychological contract. Human Resource Management, 33(3), 447-462.
Hashimoto, M. 1981. Firm-specific human capital as a shared investment. The
American Economic Review, 71(3), 475-482.
Hicks, W., & Klimoski, R. 1987. Entry into training programs and its effects on
training outcomes: A field experiment. Academy of Management Journal,
30(3), 542-552.
116
Horn, P., Caranikas-Walker, F., Prussia, G., & Griffeth, R. 1992. A meta-analytical
structural equations analysis of a model of employee turnover. Journal of
Applied Psychology, 77(6), 890-909.
Horn, P., & Griffeth, R. 1991. Structural equations modeling test of a turnover
theory: Cross-sectional and longitudinal analyses, Journal of Applied
Psychology, 76(3), 350-377.
Horn, P., Griffeth, R., & Sellaro, L. 1984. The validity of Mobley’s (1977) model of
employee turnover. Organizational Behavior and Human Performance, 34,
141-174.
Hosemer, S., & Lemeshow, 1999. Applied Survival Analysis: Regression Modeling
of Time To Event Data. New York, NY: John Wiley, & Sons.
Huselid, M. 1995. The impact of human resource management practices on turnover,
productivity, and corporate performance. Academy of Management Journal,
38(3), 635-672.
Huselid, M., & Day, N. 1991. Organizational commitment, job involvement, and
turnover: A substantive and methodological analysis. Journal of Applied
Psychology, 74(3), 381-391.
Jackofsky, E. 1984. Turnover and job performance: An integrated process model.
Academy of Management Review, 9, 606-611.
Jackofsky, E., Ferris, K., & Breckenridge, B. 1986. Evidence of a curvilinear
relationship between job performance and turnover. Journal of Management,
12,105-111.
Johnston, M., Griffeth, R., Burton, S., & Carson, P. 1993. An exploratory
investigation into the relationships between promotion and turnover: A quasi-
experimental longitudinal study. Journal of Management, 19(1), 33-49.
Kallenberg, A., & Reve, T. 1992. Contracts and commitment: Economic and
sociological perspectives on employment relations. Human Relations, 46(9),
1103-1133.
Katz, D., & Kahn, R. 1978. The Social Psychology of Organizations, New York,
NY: Wiley.
117
Kinicki, A., Carson, K., & Bohlander, G. 1992. Relationship between an
organization's actual human resource efforts and employee attitudes. Group,
& Organization Management, 17(2), June, 135-152.
Kirkpatrick, D. 1976. Evaluation of training. In R. Craig (Ed.), Training and
Development Handbook. New York, NY: McGraw-Hill.
Kirschenbaum, A., & Mano-Negrin, R. 1999. Underlying labor market dimensions
of'opportunities': The case of employee turnover. Human Relations, 52(10),
1233-1255.
Koch, M., & McGrath, R. 1996. Improving labor productivity: Human resource
management policies do matter. Strategic Management Journal, 17, 335-354.
Koys, D. 1988. Human resource management and a culture of respect: Effects on
employees' organizational commitment. Employee Rights and
Responsibilities Journal, 1, 57-67.
Kozlowski, S., & Farr, J. 1988. An integrative model of updating and performance.
Human Performance, 1, 5-29.
Kozlowski, S., & Hults, B. 1987. An exploration of climates for technical updating
and performance. Personnel Psychology, 40, 539-561.
Kraiger, K., Ford, K., & Salas, E. 1993. Application of cognitive, skill-based, and
affective theories of learning outcomes to new methods of training. Journal
of Applied Psychology, 78(2), 311-329.
Krueger, A., & Rouse, C. 1998. The effects of workplace education on earnings,
turnover, and job performance. Journal of Labor Economics, 16( 1), 61-94.
Lawler, E., Mohrman, S., & Benson, G. 2001. Organizing for high performance: The
CEO Report on Employee Involvement, TQM, Reengineering, and
Knowledge Management in Fortune 1000 Companies. San Francisco, CA:
Jossey-Bass.
Lee, T., & Maurer, S. 1999. The effects of family structure on organizational
commitment, intention to leave, and voluntary turnover. Journal of
Managerial Issues, 11(4), 493-513.
118
Lee, T., & Mitchell, T. 1994. An alternative approach: The unfolding model of
voluntary employee turnover. Academy of Management Journal, 19(1), 51-
89.
Lee, T., Mitchell, T., Holton, B., McDaniel, L., Hill, J. 1999. The unfolding model of
voluntary turnover. Academy of Management Journal, 42(4), 450-462.
Lee, T., & Mowday, R. 1987. Voluntarily leaving and organization: An empirical
investigation of Steer's and Mowday's model of turnover. Academy of
Management Journal, 30(4), 721-743.
Lepak, D., & Snell, S. 1999. The human resource architecture: Toward a theory of
human capital allocation and development. Academy of Management
Journal, 24(1), 31-48.
Lengermann, P. 1996. How long do the benefits of training last? Evidence of long
term effects across current and previous employers, education levels, test
scores and occupations. [CAHRS Working Paper] (96-18). Ithaca, N.Y.
Loewenstein, M., & Spletzer, J. 1999.General and specific training: Evidence and
implications. Journal of Human Resources, 34(4), 710-733.
_________________ . 1998. Dividing the costs and returns to general training.
Journal of Labor Economics, 16(1), 142-171.
_________________ . 1997. Delayed formal on-the-job training. Industrial and
Labor Relations Review, 51(1), 82-99.
Lynch, L. 1992. Private-sector training and the earnings of young workers. American
Economic Review, 52(1), 299-312.
_________________ . 1991. The role of off-the-job vs. on-the-job training for the
mobility of women workers. American Economic Review, 81(2), 299-312.
Lynch, L., & Black, S. 1998. Beyond the incidence of employer-provided training.
Industrial, & Labor Relations Review, 52(1), October, 64-81.
March, J., & Simon, H. 1958. Organizations. New York, NY: Wiley.
Martin, G., Pate, J., & Beaumont, P. 2000. Company-based continuous education:
What's the pay-off for employees? Academy of Management Annual
Meeting. Toronto, Canada, August.
119
Mathieu, J., & Zajac, D. 1990. A review and meta-analysis of the antecedents,
correlates, and consequences of organizational commitment. Psychological
Bulletin, 108,171-194.
Maurer, T., & Tarulli, B. 1994. Investigation of perceived environment, perceived
outcomes, and person variables in relationship to voluntary development
activities by employees. Journal of Applied Psychology, 79(1), 3-14.
Meister, J. 1998. The quest for lifetime employability. Journal of Business Strategy,
19(3), 25-28.
Meyer, J., & Allen, N. 1997. Commitment in the workplace: Theory, research, and
application. Thousand Oaks: Sage.
_________________ . 1984. Testing the 'side-bet theory' of organizational
commitment: Some methodological considerations. Journal of Applied
Psychology, 68, 372-378.
Meyer, J., Paunonen, S., Gellatly, I., Goffin, R., & Jackson, D. 1989. Organizational
commitment and job performance: It's the nature of the commitment that
counts. Journal of Applied Psychology, 74, 152-156.
Meyer, J., & Smith, C. 2000. HRM Practices and organizational commitment: Test
of a mediation model. Canadian Journal of Administrative Sciences, 17(4),
319-331.
Mobley, W., Griffeth, R., Hand, H., & Meglino, B. 1979. Review and conceptual
analysis of the employee turnover process. Psychological Bulletin, 86,493-
522.
Morita, J., Lee, T., & Mowday, R. 1993. The regression-analog to survival analysis:
A selected application to turnover research. Academy of Management
Journal, 35(6), 1430-1464.
_________________ . 1989. Introducing survival analysis to organizational
researchers: A selected application to turnover research. Journal of Applied
Psychology, 74(2), 280-292.
Mowday, R., Porter, L., & Steers, R. 1982. Employee-Organization Linkages. New
York, NY: Academic Press.
120
Noe, R. 1999. Employee Training and Development, Boston, MA: McGraw-Hill.
Noe, R., & Wilk, S. 1993. Investigation of the factors that influence employees'
participation in development activities. Journal of Applied Psychology, 78(2),
291-302.
Noe, R., Wilk, S., Mullen, E., & Wanek, J. 1997. Employee development: Issues in
construct definition and investigation of antecedents. In J. Ford (Ed.),
Improving Training Effectiveness in Work Organizations (pp. 153-189).
Mahwah, NJ: Lawrence Earlbaum Associates.
Nordhaug, O. 1989. Reward functions of personnel training. Human Relations, 42,
373-388.
Oakland, J., & Oakland, S. 1998. The links between people management, customer
satisfaction and business results. Total Quality Management, 9(4&5), 185-
190.
O’Reilly, C., & Caldwell, D. 1981. The commitment and job tenure of new
employees: Some evidence of postdecisional justification. Administrative
Science Quarterly, 24(4), 597-617.
O’Reilly, C., & Chatman, J. 1986. Organizational commitment and psychological
attachment: The effects of compliance, identification, and internalization on
prosocial behavior. Journal of Applied Psychology, 71, 492-499.
Porter, L., Steers, R., Mowday, R., & Boulian, P. 1974. Organizational commitment,
job satisfaction, and turnover among psychiatric technicians. Journal of
Applied Psychology, 59, 603-609.
Recruiting and Retaining Employees: Using Training and Education in the War for
Talent [American Society for Training and Development]. 1999. Alexandria,
VA: American Society for Training and Development.
Romzek, 1989. Personal consequences of employee commitment. Academy of
Management Journal, 32(3), 649-662.
Rouiller, J., & Goldstein, I. 1993. The relationship between organizational transfer
climate and positive transfer of training. Human Resource Development
Quarterly, 4(A), 377-390.
121
Rousseau, D. 1995. Psychological Contracts in Organizations: Understanding
Written and Unwritten Agreements. Thousand Oaks, CA: Sage Publications.
Russell, J. 1991. Career development interventions in organizations. Journal of
Vocational Behavior, 38, 237-287.
Schwarzwald, J., Koslowsky, M., & Shalit, B. 1992. A field study of employees’
attitudes and behaviors following promotion decisions. Journal of Applied
Psychology, 77(4), 511-514.
Shapiro, D. 2001. EQW National Employer Survey Frequencies from the NES1994
Public Use File. Available at http://www.irhe.upenn.edu/cgi-bin/cat.pl.
Shaw, J., Delery, J., Jenkins, G., & Gupta, N. 1998. An organization-level analysis of
voluntary and involuntary turnover. Academy of Management Journal, 41(5),
511-525.
Shore, L., & Shore, T. 1995. Perceived organizational support and organizational
justice. In R. Cropanzano, & K. Kacmar (Eds.), Organizational Politicism
Justice, and Support: Managing Social Climate at Work. Westport, CT:
Quorum Press.
Somers, M. 1996. Modeling employee withdrawal behavior over time: A study of
turnover using survival analysis. Journal of Occupational and Organizational
Psychology, 69, 315-326.
_________________ . 1995. Organizational commitment, turnover and absenteeism:
An examination of direct and interaction effects. Journal of Organizational
Behavior, 16,49-58.
Spence, M. 1974. Market Signaling: Information Transfer in Hiring and Related
Screening Processes, Cambridge, MA: Harvard University Press.
Stevens, M. 1996. Transferable skills and poaching externalities. In A. L. Booth, &
D. Snower (Eds.), Acquiring Skills. Cambridge: University Press.
Strober, M. 1990. Human capital theory: Implications for HR managers. Industrial
Relations, 29, 214-239.
122
Tannenbaum, S., Mathieu, J., Salas, E., & Cannon-Bowers, J. 1991. Meeting
trainees' expectations: The influence of training fulfillment on the
development of commitment, self-efficacy and motivation. Journal of
Applied Psychology, 76, 759-769.
Tannenbaum, S., & Yukl, G. 1992. Training and development in work organizations.
Annual Review of Psychology, 43, 399-441.
Tett, R,, & Meyer, J. 1993. Job satisfaction, organizational commitment, turnover
intention, and turnover: Path analyses based on meta-analytic findings.
Personnel Psychology, 46, 259-293.
Tharenou, P. 1997. Determinants of participation in training and development. In C.
Cooper, & D. Rousseau (Eds.), Trends in Organizational Behavior. New
York: Wiley, & Sons.
Tracey, J., Tannenbaum, S., & Kavanagh, M. 1995. Applying trained skills on the
job: The importance of the work environment. Journal of Applied
Psychology, 80(2), 239-252.
Trevor, C. 2001. Interactions among actual ease-of-movement determinants and job
satisfaction in prediction of voluntary turnover. Academy of Management
Journal, 44(6), 621-638.
Trevor, C., Gerhart, B., & Boudreau, J. (1997). Voluntary turnover and job
performance: Curvilinearity and the moderating influences of salary growth
and promotions. Journal of Applied Psychology, 82(1), 44-61.
Tsui, A., Pearce, J., Porter, L., & Tripoli, A. 1997. Alternative approaches to the
employee-organization relationship: Does investment in employees pay off?
Academy of Management Journal, 40(5), 1089-1121.
Ulrich, D. 1998. Intellectual capital = Competence X Commitment. Sloan
Management Review, Winter, 15-26.
Van Buren, M. 2001, March, 26. ASTD releases its 2001 State of the Industry
Report. Available: http://www.astd.org.
Wayne, S., Shore, L., & Liden, R. 1997. Perceived organizational support and
leader-member exchange: A social exchange perspective. Academy of
Management Journal, 40: 82-111.
Wexley K. & Latham, G. 2002. Developing Training and Human Resources
Organizations, Upper Saddle River, N.J.: Prentice Hall.
124
APPENDIX 1
Summary of Hypotheses
H1-H6
H7-H9
H10-H12
T uition-reimbursement
Company-sponsored
training
On-the-job training
Org. Commitment
Intent to Turnover
Actual turnover
Rewards:
Promotion or Merit Pay
125
APPENDIX 2
Survey Measures
During the past year, please indicate the number of days you spent in the following
activities and evaluate their usefulness to (A) your current job performance and to
(B) your future marketability?
Number # ofDays
1. Formal company
training cla sses or program s .............
2. For-credit college co u rses .............
3. Company seminars, conferences
or knowledge-sharing netw orks .............
4. External conferences and
knowledge-sharing................................................
5. Structured on-the-job training .............
6. Special assign m en ts............................................
7. Visiting with customers, suppliers,
and partner com p an ies.......................................
126
APPENDIX 3
Rotated Component Matrix2
Components
I am proud to tell others that I am part o f this organization.
1
.768
2
I find that m y values and the organization’s values are very similar. .711
I talk up this organization to m y friends as a great place to work. .815
I am extrem ely glad to have chosen this organization .772
For m e this is the best o f all organizations for which to work. .700
This organization inspires the very best job performance. .774
I would be w illing to change companies for career advancement now. .911
I w ill look for a new job outside m y organization in the next year. .781
“ Factor loadings less than .4 are suppressed.
Extraction Method: Principal Component Analysis.
Rotation Method: Varimax with Kaiser Normalization.
127
APPENDIX 4
Descriptive Statistics: Survey Sample
Variable N M inim um M aximum M ean Std.
Deviation
Organizational commitment 963 1 5 3.16 0.77
Intention to turnover 970 1 5 3.09 1 .1 0
Days tuition-reimbursement classes 979 0 2 0 0 9.04 24.17
Days on-the-job training 979 0 300 3.60 16.27
Days com pany training classes 979 0 6 6 6.87 7.59
Earned a degree 979 0 1 0 .1 0 0.30
Took classes 979 0 1 0.41 0.49
Number o f promotions 979 0 4 0.53 0 .6 8
Number o f merit payments 979 0 5 2.19 1.56
Degree then promotion 979 0 1 0.04 0.19
Degree then merit award 979 0 1 0.06 0.24
Tenure 979 0.36 48.53 13.92 11.41
Grade 1 (Engineer 1) 979 0 1 0.08 0.26
Grade 2 (Engineer 2) 979 0 1 0.15 0.36
Grade 3 (Engineer 3) 979 0 1 0 .2 1 0.40
Grade 4 (Manager 1) 979 0 1 0.35 0.48
Grade 5 (Manager 2) 979 0 1 0.13 0.34
Grade 6 (Manager 3) 979 0 1 0.08 0.28
Business Unit 1 979 0 1 0.49 0.50
Business Unit 2 979 0 1 0 .2 1 0.41
128
Descriptive Statistics: Archival Sample-All Employees
Variable N M inimum M aximum M ean Std.
Deviation
A ll quits
12360 0 1 0 .1 0 0.30
A ll turnover any reason
12360 0 1 0.30 0.46
R eceived stock award
12360 0 1 0.09 0.28
R eceived promotion
12360 0 1 0.37 0.48
R eceived merit pay
12360 0 1 0.83 0.38
Degree then promotion
12360 0 1 0.03 0.17
Degree then merit pay
12360 0 1 0.06 0.23
Tenure
12360 0.29 54.21 18.14 11.77
Grade 1 (Engineer 1) 12360 0 1 0.15 0.36
Grade 2 (Engineer 2) 12360 0 1 0 .1 0 0.30
Grade 3 (Engineer 3) 12360 0 1 0.24 0.43
Grade 4 (Manager 1) 12360 0 1 0.34 0.47
Grade 5 (Manager 2) 12360 0 1 0 .1 0 0.30
Grade 6 (Manager 3) 12360 0 1 0.06 0.24
Business Unit 1 12360 0 1 0 .1 2 0.33
Business Unit 2 12360 0 1 0.04 0 . 2 0
Descriptive Statistics: Archival Sample-1996 Actives
129
Variable N M inim um M aximum M ean Std.
Deviation
A ll quits
9243 0 1 0.08 0.28
A ll turnover any reason
9243 0 1 0.33 0.47
Received stock award
9243 0 1 0 .1 0 0.29
R eceived promotion
9243 0 1 0.37 0.48
R eceived merit pay
9243 0 1 0.92 0.26
Degree then promotion
9243 0 1 0.03 0.17
Degree then merit pay
9243 0 1 0.06 0.24
Tenure
9243 0.72 54.21 22.71 9.40
Grade 1 (Engineer 1) 9243 0 1 0.13 0.34
Grade 2 (Engineer 2) 9243 0 1 0.05 0 .2 1
Grade 3 (Engineer 3) 9243 0 1 0.24 0.43
Grade 4 (Manager 1) 9243 0 1 0.40 0.49
Grade 5 (Manager 2) 9243 0 1 0 .1 1 0.32
Grade 6 (Manager 3) 9243 0 1 0.07 0.26
Business Unit 1 9243 0 1 0 .1 0 0.30
Business Unit 2 9243 0 1 0.04 0.19
APPENDIX 5
Correlations 1996 - 2000 Archival Sample N = 12360
1. 2 . 3. 4. 5. 6 . 7. 8 . 9. 10. 11. 12. 13. 14.
1. Quit 1.00
2 . Earned degree 0 .0 2 1.00
3. Took classes -0.04 -0.25 1.00
4. Promoted 1996-2000 -0 .0 2 0 .2 0 0.13 1.00
5. Merit pay 1996-2000 -0.15 0 .1 0 0 .1 2 0.15 1.00
6 . Promotion after degree -0 .0 2 1.00 -0.15 0.24 0.07 1.00
7. Merit pay after degree -0.01 0.78 -0.19 0.16 0.11 0.87 1.00
8. Seniority in years -0.21 -0.09 -0.11 -0 .2 0 0.35 -0.08 -0.06 1.00
9. Grade 1 0.13 0 .0 0 0 .0 0 0 .0 2 -0.27 0 .0 2 -0.01 -0.25 1.00
10. Grade 2 0 .0 2 0 .0 2 0.05 0.11 -0.19 0 .0 2 0 .0 0 -0.28 -0.07 1.00
11. Grade 3 0.08 0 .0 2 -0.03 -0 .1 0 -0 .0 2 0 .0 0 0.01 -0.04 -0.12 -0.19 1.00
12. Grade 4 -0.07 -0.03 -0.01 -0.06 0.15 -0.04 -0 .0 2 0.19 -0.15 -0.24 -0.41 1.00
13. Grade 5 -0.05 0 .0 2 -0.01 0.14 0.09 0.04 0.03 0.08 -0.07 -0.11 -0.19 -0.24 1 .00
14. Grade 6 -0.03 -0.01 -0.01 0 .0 2 0.06 0 .0 0 0 .0 0 0 .1 2 -0.06 -0.09 -0.15 -0.19 -0.09 1.00
Correlations 1996 Archival Sample N = 9243
1. 2 . 3. 4. 5. 6 . 7. 8 . 9. 10. 11. 12. 13. 14.
1. Quit 1.00
2 . Earned degree 0.04 1.00
3. Took classes -0.06 -0.25 1.00
4. Promoted 1996-2000 -0 .0 2 0.19 0 .1 0 1.00
5. Merit pay 1996-2000 -0 .2 0 0.07 0.14 0 .1 2 1.00
6 . Promotion after degree -0 .0 2 1.00 -0.15 0.25 0.04 1.00
7. Merit pay after degree 0 .0 0 0.80 -0 .2 0 0.17 0.07 0.89 1.00
8. Seniority in years -0.27 -0.21 -0.14 -0.34 -0 .0 2 -0.16 -0.16 1.00
9. Grade 1 0 .0 2 0.05 0 .0 0 0.07 -0.05 0 .1 0 0.05 -0.07 1.00
10. Grade 2 0.06 0.07 0 .0 2 0.08 -0.06 0.06 0.05 -0.09 -0 .0 2 1.00
1 1 . Grade 3 0.11 0.03 -0.04 -0 .1 0 -0.07 0 .0 0 0 .0 2 -0.05 -0.05 -0.12 1.00
12. Grade 4 -0.06 -0.05 0.01 -0.05 0.07 -0.05 -0.04 0.07 -0.08 -0.18 -0.45 1.00
13. Grade 5 -0.04 0.01 0 .0 0 0.17 0.04 0.05 0 .0 2 0 .0 0 -0.03 -0.08 -0.20 -0.29 1 .00
14. Grade 6 -0.03 -0 .0 2 0 .0 0 0.03 0 .0 0 0 .0 0 -0.01 0.08 -0.03 -0.06 -0.16 -0.23 -0 .1 0 1.0 0
132
Correlations 2000 Survey Sample N = 979
1. 2 . 3. 4. 5. 6 . 7. 8 . 9. 1 0.
1. Org. commitment
1 .0 0
2 . Intent to turnover -0.49 1 .0 0
3. Quit -0.13 0.17 1 .0 0
4. #days tuit’n reimburse’t -0 .0 2 0 .1 2 0 .0 1 1 .0 0
5. #days structured OJT 0 .1 0 -0 .1 1 0.07 0 .1 1 1 .0 0
6 . #days company classes
0.06 -0.03 0 .0 1 0 .0 0 0.18 1 .0 0
7. Earned degree
0 .0 1 0 .1 1 0 .0 2 0 .1 1 -0 .0 1 -0.05 1 .0 0
8 . Took classes -0.06 0.13 0 .0 2 0.17 -0.03 0.08 -0.27 1 .0 0
9. Promoted 1996-2000 0 .0 1 0.16 0.04 0 .1 1 0.04 -0 .0 1 0.18 0.08 1 .0 0
1 0. Merit pay 1996-2000 -0.19 0.08 0.03 -0.03 -0 .1 1 -0.09 0 .1 1 0.08 0 .2 0 1 .0 0
1 1. Promotion after degree 0.03 0.04 -0 .0 2 0.05 0.04 0 .0 0 0.61 -0.17 0.23 0.03
1 2. Merit pay after degree 0 .0 1 0.07 0 .0 0 0 .0 1 -0.04 -0.05 0.76 -0 .2 1 0 .1 2 0.14
13. Seniority in years
-0.15 -0.15 -0.15 -0.13 -0.09 -0.07 0 .0 2 -0 .1 2 -0.07 0.49
14. Grade 1
0.07 0 .0 0 0.30 0.05 0 .1 0 0.17 -0.07 0 .0 2 -0.05 -0.32
15. Grade 2 0 .1 2 0 .0 0 0 .0 0 0.06 0.04 0 .1 0 0 .0 2 0.08 0.13 -0.28
16. Grade 3
-0 .0 1 0 .0 1 0 .0 2 0.07 -0.03 -0.13 -0 .0 1 0.05 -0 .1 1 -0.05
17. Grade 4 -0.17 -0 .0 2 -0.08 -0.08 -0 .0 1 -0 .0 2 0 .0 0 -0.07 -0.14 0.23
18. Grade 5
0.03 0 .0 1 -0.09 -0.03 -0.03 -0.04 0.08 -0.03 0 .2 0 0.17
19. Grade 6 0.05 -0 .0 1 -0.07 -0.04 -0.05 -0 .0 2 -0.04 -0.04 0.04 0,14
2 0 . Bus. Unit 1 0.14 -0 .2 0 0 .0 2 -0.04 0 .0 1 0.15 -0.05 0.04 -0.09 -0.15
2 1 . Bus. Unit 2 -0.17 0.23 0 .1 2 -0 .0 1 -0 .0 1 0 .1 0 -0 .0 1 0.05 0.04 0 .1 0
Correlations 2000 Survey Sample N = 979
Cont’d
133
11. 12. 13.
1 .
2.
3.
4.
5.
6 .
7.
8 .
9.
1 0.
11. 1.00
1 2. 0.44 1.00
13. 0 .0 0 0.03 1.0 0
14. -0.06 -0.07 -0.29
15. 0.07 -0.04 -0.35
16. -0 .0 2 0.01 -0.13
17. -0.03 -0.01 0.30
18. 0.03 0 .1 0 0.17
19. 0 .0 2 0 .0 0 0.19
2 0 . -0.05 -0.03 -0.15
2 1 . 0 .0 0 -0.05 0.03
14. 15. 16. 17.
1.00
-0 .1 2 1.00
-0.15 -0.21 1.00
-0.21 -0.31 -0.38 1.00
-0.11 -0.16 -0 .2 0 -0.29
-0.09 -0.13 -0.15 -0 .2 2
0.07 0.13 -0.07 0.03
0.06 -0 .0 2 0 .0 0 0.01
18. 19. 20. 21.
1.00
- 0.12 1.00
-0.06 -0.13 1.00
-0.06 0.02 -0.51 1.00
134
APPENDIX 6
P-P Plots for Training Time Transformations
On-the job training
1.00
.75
.50
n
2
C L
i , 5
O
o
CD
Q.
l3 0.00
0.00 .50 .75 1.00 .25
Observed Cum Prob
ln(On-the-job training)
.00
.75
.50
8
a
E
o 25
■ O
d )
o
0 )
Q.
X
UJ
o . o o .50 .25 1.00 .75
Observed Cum Prob
Transforms: natural log
Expected C u m P rob Expected C u m Prob
Company-training classes
1.00
.75
.50
.25
0.00
0.00 .25 .50 .75 1.00
Observed Cum Prob
ln(Company-training classes)
1.00
.75
.50
.25
0.00
0.00 .25 .50 1.00 .75
Observed Cum Prob
Transform s: natural log
136
Tuition-reimbursement classes
1.00
.75
.50
.25
Q.
U J 0 . 0 0
1.00 0.00 .25 .50 .75
Observed Cum Prob
ln(T uition-reimbursement classes)
1.00
.75
.50
.25
Q.
UJ 0.00
1.00 .25 .50 0.00 .75
Observed Cum Prob
Transforms: natural log
137
APPENDIX 7
Usefulness to Future Marketability: Tuition-Reimbursement Classes
200
Usefulness to Future Marketability
Usefulness to Future Marketability: Company Classes
400 ------------------------------------------------------------------------------------------------------
Usefulness to Future Marketability
138
Usefulness to Future Marketability: On-the-job Training
Usefulness to Future Marketability
139
APPENDIX 8
Results of Hierarchical Regression for Organizational Commitment:
Archival Data
Step 1
Beta t
Step 2a
Beta t
Step2b
Beta t
Tenure3 -0 .1 1 -2.84 ** -0 .1 1 -2 .8 8 ** -0.07 -1.72 +
Grade 1 -0.17 -2.67 ** -0.17 -2.67 ** -0.13 -1.91 +
Grade 2 0 .0 1 0.16 0 .0 1 0 .2 0 0 .0 2 0.45
Grade 4 -0.08 -1.53 -0.08 -1.53 -0.06 -0.98
Grade 5 -0 .0 2 -0.34
I
o
©
-0.26 0 .0 2 0.33
Grade 6 0 .0 2 0.43 0 .0 2 0,43 0.05 1 .0 2
Business U nit 1 0.05 1.39 0.05 1.30 0.05 1.38
Business Unit 2 -0.14 -3 78 *** -0.14 -3 78 *** -0.13 -3 53 ***
Earned degree 0 .0 0 -0.13 -0 .0 2 -0.41 0 .0 1 0.17
Took classes -0.08 -2.50 * -0.06 -1.53 -0 .1 0 -1.49
Promoted 1996-2000 0 .0 0 -0.09
Progress X promoted -0.03 -0.53
Degree then promoted 0 .0 2 0.53
Merit pay 1996-2000 -0 .1 2 -2.63 **
Progress X merit pay 0.05 0 .6 6
Degree then merit pay 0 .0 1 0.15
Rz (adjusted R2 ) .134 (.125) .145 (.133) .143 (.132)
F
14 9 ***
12 4 *** 1 2 .2 ***
D f 969 969 969
N 985 985 985
“ stan dardized
+ P < .1 0
* P < .05
**p< oi
***p < ooi
140
APPENDIX 9
Hazard Function Graphed for Each Level of Covariates
Tenure
2
1
Tenure Median Split
High
0.0 ° Low
0 10 20 30 40 50 60
Months
Grade 1
.06
.05-
.04
.03
.02
Grade 1
.01
Yes
O 0.00 No
20 0 10 30 40 50 60
Months
Cumulative Hazard
Grade 3
Cumulative Hazard
O
o
o
o
O )
o
G
□ Q
CL
Q
— t
0)
Q.
CD
N >
ro
4^
Cumulative Hazard
O
0 0 0 0 0 0 0
o -k ro co at o >
o
o
o
co
o
§
tn
o
s
o □ a
CL
cn
G rade 5
Cumulative Hazard
O
0 0 0 0 0 0 0
O -*• N> CO 4^ t n O )
o
o
o
O )
0
a □
C l
K >
143
Grade 6
.07
.06
.05
.04
T3 03
fl
C O
X .02
0 )
>
G rade 6
0 3
= ■
E
Z J
O
.01
a Yes
° No 0.00
60 20 30 40 50 0 10
Months
Promotion
.07
.0 5 '
.03-
.02
Prom otion
.01
D FYomoted
D No Promotion O 0.00
0 20 30 40 50 60 10
Months
144
Merit Award
.3
.2
1
Merit Award
° Merit
0.0 ° No Merit
0 10 20 30 40 50 60
Months
Degree Progress
.07
.06'
.05
.04
.03
.02
Degree Progress
0 Progress
D No FVogress
.01
6 0.00
0 20 40 10 30 50 60
Months
145
Earned Degree
.06
.05
.04
.03-
.02
Earned Degree
° Earned Degree
D No Degree
.01
Q 0.00
0 20 40 10 30 50 60
Months
146
APPENDIX 10
Logistic Regression Results
Step
1
B Exp(B) Wald
Sig
Step
2
B Exp(B) W ald Sig
Tenure3 -0.35 0.71 1.43 -0.35 0.70 1.44
Grade 1 2 .0 2 7.52 13.35
**
2.14 8.51 14.62
Grade 2 0.17 1.19 0.08 0.28 1.32 0 .2 1
Grade 4 0.62 1.87 1.53 0.57 1.76 1.25
Grade 5 -6.23 0 .0 0 0 .2 0 -7.21 0 .0 0 0 .1 0
Grade 6 -6.19 0 .0 0 0 .1 2 -7.19 0 .0 0 0.06
Business Unit 1 1.48 4.40 3.82
*
1.72 5.59 4.94
*
Business U nit 2 2.25 9.52 8.71
**
2.49 1 2 .1 0 10.13
**
In (Tuit’n-reimbursement) 0 .0 2 1 .0 2 0.05
In (On-the-job training) -0.13 0 .8 8 0.67
In (Company classes) -0 .2 2 0.80 1.85
X2 82.95
***
8 6 .1 2
***
D f
8 11
N 979 979
“ stan dardized
+ P < .10
* P < . 05
**P< .01
***p< .0 0 1

Asset Metadata

Creator Benson, George (author)

Core Title Beyond skill development: The effects of training and development on the attitudes and retention of employees

Contributor Digitized by ProQuest (provenance)

Degree Doctor of Philosophy

Degree Program Psychology

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag OAI-PMH Harvest,psychology, industrial

Language English

Advisor Lawler, Edward (committee chair), Finegold, David (committee member), Maxwell, William (committee member), Mohrman, Susan (committee member)

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c17-67861

Unique identifier UC11349028

Identifier DP22713.pdf (filename),usctheses-c17-67861 (legacy record id)

Legacy Identifier DP22713.pdf

Dmrecord 67861

Document Type Dissertation

Rights Benson, George

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the au...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus, Los Angeles, California 90089, USA