University of Southern California Dissertations and Theses

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

(USC Thesis Other)

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

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content INFORMATION TO USERS This manuscript has been reproduced from the microfilm master. UMI films the text directly from the original or copy submitted. Thus, some thesis and dissertation copies are in typewriter face, while others may be from any type of computer printer. The quality of this reproduction is dependent upon the quality of the copy submitted. Broken or indistinct print, colored or poor quality illustrations and photographs, print bleedthrough, substandard margins, and improper alignment can adversely affect reproduction. In the unlikely event that the author did not send UMI a complete manuscript and there are missing pages, these will be noted. Also, if unauthorized copyright material had to be removed, a note will indicate the deletion. Oversize materials (e.g., maps, drawings, charts) are reproduced by sectioning the original, beginning at the upper left-hand corner and continuing from left to right in equal sections with small overlaps. ProQuest Information and Learning 300 North Zeeb Road, Ann Arbor, Ml 48106-1346 USA 800-521-0600 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. BEYOND SKILL DEVELOPMENT: THE EFFECTS OF TRAINING AND DEVELOPMENT ON THE ATTITUDES AND RETENTION OF EMPLOYEES by George Sinclair Benson A Dissertation Presented to the FACULTY OF THE GRADUATE SCHOOL UNIVERSITY OF SOUTHERN CALIFORNIA in Partial Fulfillment o f the Requirements for the Degree DOCTOR OF PHILOSOPHY (BUSINESS ADMINISTRATION) May 2002 Copyright 2002 George Sinclair Benson Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. UMI Number: 3073745 ___ ® UMI UMI Microform3073745 Copyright 2003 by ProQuest Information and Learning Company. All rights reserved. This microform edition is protected against unauthorized copying under Title 17, United States Code. ProQuest Information and Learning Company 300 North Zeeb Road P.O. Box 1346 Ann Arbor, Ml 48106-1346 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. UNIVERSITY OF SOUTHERN CALIFORNIA The Graduate School U niversity Park LOS ANGELES, CALIFORNIA 90089-1695 This dissertation , w ritten b y George Sinclair Benson U nder th e direction o f Ai.s.. D issertation C om m ittee, an d approved b y a ll its m em bers, has been p resen ted to a n d accepted 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 ian o f Graduate Studies D ate .May IQ.y 2GQ2- Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. ii ACKNOWLEDGEMENTS This dissertation would not have been possible without help from a number o f people. Most important has been the support o f 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 o f time and effort of behalf of 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 o f Jack Leary, Susan Rennie, Raquel Marero, Gail Heiss, Edward Duffy, Sandra Martinez, Jonathan Brink, Kathy Biederman, Nancy Coleman, Henry Mercolino, and many others. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 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 M easures.....................................................................................................................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 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. IV APPENDIX 9 ...........................................................................................................140 APPENDIX 10.........................................................................................................146 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. V LIST OF TABLES Table 1 Proposed Dimensions of a Development Construct.............................. 15 Table 2 Survey Response Rates by U nit................................................................33 Table 3 Distribution o f Job G rades........................................................................ 37 Table 4 Employees who Received Merit Pay and Promotion............................43 Table 5 Turnover 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 D ata............................................................................................................. 68 Table 10 Earning a Degree, Timing of Rewards and Intention to Turnover... 70 Table 11 Rates o f Voluntary Turnover Across the Samples.................................73 Table 12 Cox Regression R esults............................................................................. 80 Table 13 Timing o f Promotion Relative to Degree and Quit Rates....................84 Table 14 Timing o f Merit Pay Relative to Degree and Quit Rates.....................84 Table 15 Time-Dependent Cox Regression R esults.............................................. 86 Table 16 Summary of Hypotheses and Findings....................................................96 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. LIST OF FIGURES Figure 1 Employee Ratings o f 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 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. vii 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. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 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 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 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 o f firms (Cappelli, 2000; Gardner, Moynihan, Park, & Wright, 2000; Huselid, 1995; Koch, & McGrath, 1996; Lawler, Mohrman, & Benson, 2001). The management o f 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 o f 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 o f 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 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 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 o f company support in turn contribute to an employee’s organizational commitment and willingness to remain with the finn (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 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 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 o f 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 o f training and development have historically been evaluated in terms o f skill acquisition and performance. There Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 5 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 o f employees, such as organizational commitment and job satisfaction, should also be examined (Feldman 1989; Noe et ah, 1997; Tannenbaum et al., 1991). Only recently have these other potential attitudinal outcomes been incorporated into empirical studies of training and development (Birdi et ah, 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 ah, 1997; Noe et ah, 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 o f 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 o f a larger set of HR practices is Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 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 o f four practices included in a scale of “employer investment.” Tsui et al. (1997) investigated the balance o f 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 o f 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 o f 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 o f Birdi et al. (1997) support the social exchange model o f development in which participation (at least during paid work hours) leads to increased organizational commitment. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 7 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 o f 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 o f Becker (1965) and others, human capital theory predicts that training which serves to increase the skills o f 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 o f 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 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 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 o f 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 o f 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 o f 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 o f work. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 9 Taken together these two sets of literature from management and labor economics suggest two important conclusions regarding the outcomes o f 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 o f 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 o f 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 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 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 o f 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 o f existing theory on organizational commitment and employee retention. Organizational commitment. Commitment is the strength o f 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 o f the Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 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 o f 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 o f continuance commitment. They argued that the development of company-specific skills and knowledge is one o f the “side bets” or investments that employees make in their employers. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 12 While employee development has been suggested as an antecedent o f both affective and continuance commitment, different outcomes are associated with these forms o f 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 o f 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, Goffm, & 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 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 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 o f 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 o f movement to predict intentions to turnover and eventually turnover behavior (Gerhart, 1990; Trevor, 2001). The causal ordering of commitment, intention to turnover, and external Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 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 o f 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 o f human capital and ultimately the productivity o f 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. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 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 o f development outcomes. The dimensions presented by each o f 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 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 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. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 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. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 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 o f 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 o f 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 o f 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 o f participation is the nature of the skills Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 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, Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 0 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 o f 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 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 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 o f tuition-reimbursement is also supported by economic theories o f “signaling” which suggest that firms use formal education in hiring decisions as a sign o f the future productivity o f the applicant (Spence, 1974; Strober, 1990). The quality o f 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 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 22 provides easily comparable information about the nature o f 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 o f 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 o f 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 o f 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 o f providing tuition-reimbursement to employees. General skills Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 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 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 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 o f 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, & Tamili, 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 o f 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 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 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 o f 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 o f his or her inputs compared to a similar person (Adams, 1965). When a worker’s ratio of inputs to outputs diverges from that o f a comparison person, negative and uncomfortable feelings are aroused that manifest themselves in the form o f 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 o f 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 o f any type in Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 2 6 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 o f 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 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 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 o f 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 o f 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. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 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. H12: 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 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 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 o f 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. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 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 o f 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 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 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 fimi 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 o f 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. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 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 o f Southern California. This larger study was focused on technical employees, and units of the company were selected for survey because o f 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. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 33 Table 2 Survey Response Rates by Unit 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 o f the company in January and June o f 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 o f 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 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 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 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 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 o f 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.3 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 o f 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. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 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 o f commitment and turnover. First of all, research has shown that participation in employee development often differs by the level o f 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. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 3 / Table 3 Distribution of Job Grades A rchival 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 o f 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. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 38 Fifty-eight percent o f 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 o f 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 o f 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 o f closing in an effort to cut Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 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 o f 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 o f 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 o f 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 o f the outcomes of employee development based on social exchange, which imply an increasing (or Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 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 o f company- sponsored training and that several of the business units maintained goals of up to 20 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 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 o f 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 o f the analyses of past participation in tuition-reimbursement, separate variables Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 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 o f 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. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 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 o f an individual’s salary, and the receipt o f discretionary rewards on organizational commitment (Mathieu, & Zajac, 1990). Conversely, research has also found that failing to receive a promotion Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 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 o f 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 o f these two variables to provide multiple perspectives on the same phenomenon o f rewarding employees when they gain new general skills through tuition-reimbursement. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 45 Dummy variables indicating receipt o f 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. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 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 o f 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 o f 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 o f 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 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 47 relationship between development participation and attitudes. The survey sample reported a reliable scale with Chronbach’s alpha o f .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 o f 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 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 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 o f 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% o f 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 o f 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 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 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 o f the archival data on participation in tuition-reimbursement, turnover is measured over the period o f 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. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 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 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 51 attitudes measured in the survey. Both of these analyses o f 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 o f development participation. The first analysis tests Hypotheses 1-4, 7, and 8 using survey data and hierarchical OLS regression to predict the influence o f 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. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 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 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 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 o f 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 o f 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 o f 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 o f duration permits the estimation of the conditional probability o f leaving Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 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 o f 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 o f 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 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 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 o f 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. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 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. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 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 o f 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 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 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 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 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 o f Development Usefulness.* 4.50 3.83 4.00 - S 3.50 - 0) T O 3.02 3.02 3.00 - 5 2.50 - 3 L L 2.00 Company training On-the-job training Tuition- reimbursement classes *Em ployees were asked to evaluate developm ent activities in terms o f “Their usefulness to yo u r fu tu re marketability. " Responses were given on a five-p o in t scale fro m “ N ot useful at all ” to “ Extrem ely useful. " Tuition-reimbursement differs significantly from company classes and on- the-job training in terms o f its usefulness to future marketability (ANOVA, p < .001). See Appendix 7 for the complete distributions o f employee assessments of the usefulness to future marketability of the three development activities. This finding Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 6 0 provides some evidence from the present sample o f the validity o f 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 o f 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 o f 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 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 61 organizational commitment. Regression results for organizational commitment are presented in Table 7. Table 7 Results of Regression Analyses for Organizational Commitment Step 1 Beta0 t Step 2 Beta0 t Tenure0 -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 .6 6 + 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<A0 *P<. 05 * * P < .0 1 ***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. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 6 2 Table 8 Results of Regression Analysis for Intention to Turnover Step 1 Step 2 Beta3 t 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 R~ {adjusted R2 ) .108 (.1 0 0) .141 (.130) F 12.77 *** 12.87 *** D f 955 955 A R : .032 A F 11.85 *** N 973 973 “ standardized + P<A0 *P<. 05 **P<. 01 ***P< .001 The addition of the three development variables significantly increased variance explained in the model o f 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 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 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 o f 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 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 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 o f 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 o f 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. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 65 Figure 2 Participation in Tuition-Reimbursement and Intention to Turnover 3 I- O 4.50 - I ------ ® 4 .2 5 ------- | 4.00 - I 3.75 - o 3 .5 0 ------- f — o o r - 3.27 3.47 3.93 3.33 2.50 Not Participated (N=480) Took Classes Earned a Earned Degree then (N=395) Degree (N=95) Degree, No Reward Reward After (N=72) (N=23) *Receiving a rew ard after earning a degree is defined as a prom otion or m erit aw ard in addition to the $10,000 bonus received by all 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 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 6 6 that the differences between each of first three groups presented in Figure 2 are significant (t, p < .001). The effects o f 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 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 67 are estimated for both organizational commitment and intention to turnover using either promotions or merit pay to moderate the effects o f 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 o f 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. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 6 8 Table 9 Hierarchical Regression Results for Intention to Turnover: Archival Data Step 1 Beta t Step 2a Beta t Step2b Beta t Tenure2 -0.22 -6.13 *** -0 .2 1 -5.87 *** -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 1.2 2 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 6 ^ *** 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 .1 1 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 R2 (adjusted R2 ) .134 (.125) .145 (.133) .143 (.132) F 14.9 *** 12.4 *** 1 o 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 o f 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 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 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 o f 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 o f receiving the merit award or the promotion. The multivariate results presented in Table 9 demonstrate that the interaction o f 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 n also has an impact on intention to turnover of employees. 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 of receiving a promotion relative to earning a degree is examined. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 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- Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 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. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 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 o f 2000 for the 979 survey respondents. Since there were two waves o f 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 o f 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 employees were used in a categorical analysis of voluntary turnover. The 9,243 employees who were active at the beginning of 1996 were used in the Cox regression analyses to avoid including employees who were hired during the study period and may have joined the firm to take advantage of the generous tuition-reimbursement program. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 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 o f 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 o f 12,360 employees and the 9,243 employees who were active at the beginning o f 1996. A total o f 8.4% o f the employees who were active at the beginning o f 1995 quit the company over the 54-month study period, compared to 9.5% o f all the employees who were active at any time between January 1996 through June 2000. This difference is significant (x~, p < .001) given the large sample size, and indicates that Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 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 o f 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 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 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. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 76 Figure 3 Hazard Estimates for Earning a Degree vs. Not Earning a Degree .06 .05 .04 .03 •O N 02 C D X 0 3 1 0 1 D E o 0.00 Earned Degree Earned Degree D No Degree 60 10 20 50 0 30 40 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 o f 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 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 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 o f 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% o f those employees who received a either a promotion or merit award after Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 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 o f 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 § 20% ■ 18% §5 16% 17.7% 11.9 % 10.6% Not Participated (N=6446) Took Classes Earned Degree Earned (N=4825) (N=1089) Degree, No Reward After* (N=327) Degree then Reward* (N=762) *Receiving a reward after earning a degree is defined as a prom otion or m erit award in addition to the $10,000 bonus received by all employees who earned a degree through tuition-reimbursement. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 79 The archival measures of tuition-reimbursement are also analyzed using Cox regression to test Hypotheses 9 and 12. Using time-to-tumover as the dependent variable, Cox regressions are estimated using dummy variables for whether or not the employee has taken classes without earning a degree, and whether the employee has earned a degree through tuition-reimbursement. The Cox regressions include control variables for job grade and years of tenure, and interactions between the two tuition-reimbursement variables, promotions, and merit pay to test whether these rewards moderate the effects of tuition-reimbursement on voluntary turnover. The basic Cox proportional hazards regression model is specified in Equation 1. H, is the baseline hazard function with X ’ s as the predictors and interaction terms. (1) H(t:x) — H(t) exp [f$i (Xc o n tro ls ) + P ifX to o k classes) P i (Xdegree) PXX eward) Ps(XdegreeXrew aid) P i(lo o k classesXrewnrd)] Cox regressions were estimated using the 9,243 employees who were active in January o f 1996 when the company’s new tuition-reimbursement program was introduced. Results of the first Cox regression models are presented in Table 12. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 80 Table 12 Cox Regression Results Step 1 Step 2 a Step 2 b B Exp(B) Wald B Exp(B) Wald B Exp(B) Wald Tenurea -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 .2 1 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 *** - 1.10 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 7 5 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 < . 0 1 ***P< .001 Step 1 of the regression includes control variables for tenure and job grade along with the two tuition-reimbursement dummy variables. The moderating effects o f receiving a promotion or merit award are then examined in steps 2a and 2b. In Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 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 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 8 2 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 o f 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 ress X .51 p ro m o tio n X 1 ■ " 7 to o k ciassesX p ro m o tio n ) for taking classes together with a promotion, and .69 (2.19d e g ree x • 51 p ro m o tio n x .63(jegreeX prom 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 c lasses X .23meritpay X 5.22took classesX m erit pay) for taking classes without earning a degree and .16 (5.27deg rC C x .11 m erit p ay x .27degrecX m critpay) for earning a degree. This means that employees who took classes and received a merit award 9 The coefficients are multiplied because they are exponentiated. The same result can be achieved by summing the raw coefficients and taking the exponent of the result. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 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 o f 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 o f 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 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 84 merit awards relative to earning a degree through tuition-reimbursement is a potentially important predictor of voluntary turnover. Table 13 Timing o f Promotion Relative to Earning a Degree and Quit Rates Quit Total Earned Degree without Promotion 11.8% 356 Promoted before Degree only 17.0% 111 Promoted after Degree only 8.6% 197 Promoted before and after Degree 3.7% 159 yt,P< ■001 Table 14 Timing o f 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 y?,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 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 85 elapsed time between earning a degree and receiving a promotion or merit award instead o f 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 o f 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 o f 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. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 8 6 Table 15 Time-Dependent Cox Regression Results Step 1 Step 2a Step 2b B Exp(B) Wald B Exp(B) Wald B Exp(B) Wald Tenure3 -1.57 0.21 589.56 *** -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 0.68 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 0.20 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 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 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 o f 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 o f receiving a promotion or merit award. Exponentiated coefficients o f 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 o f 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 o f 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 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 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 e a m e d d e g re e X . 3 9 p ro n lo tio n X ( 1 ) 1 . 0 4 m o n th s to p r o m o tio n ) 3S likely to quit 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 . 1 0 ( . 5 4 e a m e d d e g re e x . 3 9 p r 0 m o tio n x (1 0 ) 1. 0 4 niOn th s to p ro m o tio n ) 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 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 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 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 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. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 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 o f on-the-job training, development is Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 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 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 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 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 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 o f 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 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 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 o f 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 o f 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 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 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 Development Activity Hypothesized Relationship Predicted Outcome Findings 1 OLS Co. Classes Positive Commitment Not Supported 2 OLS OJT Positive Commitment Supported 3 OLS Co. Classes Negative Intention To Turnover Not Supported 4 OLS OJT Negative Intention To Turnover Supported 5 Logistic Co. Classes Negative Turnover Not Supported 6 Logistic OJT Negative Turnover Not Supported 7 OLS Tuition- reimbursement Negative Commitment Not Supported 8 OLS Tuition- reimbursement Positive Intention To Turnover Supported 9 Cox Tuition- reimbursement Positive Turnover Supported 10 OLS Tuition- reimbursement Moderated by Rewards Commitment Not Supported 11 OLS Tuition- reimbursement Moderated by Rewards Intention To Turnover Supported 12 Cox Tuition- reimbursement Moderated by Rewards Turnover Supported Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 97 Theoretical Contribution The effects o f development on employee attitudes and retention depend on the nature o f 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 o f 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 o f 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 o f development on organizational commitment with a notable exception. Birdi et al. (1997) found effects for company Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 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 o f the outcomes o f 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 o f skills gained. However, these dimensions do not adequately address whether the skills gained through development are useful and marketable to Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 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 o f the training and development literature already argues for viewing development and development outcomes in the larger context o f the employee’s work history and enviromnent. 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 o f development on commitment and retention, what happens to the employee following his or her participation should be considered. In terms o f 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 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 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 o f 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 o f part-time study) that it takes an employee to complete a degree. A second reason is that providing development may include Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 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 o f 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 o f 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 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 102 promotion or merit award entails additional costs to the employer, it also reduces the risk o f turnover and increases the chances o f 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 o f employee development onto these models of voluntary turnover illustrates how the social and economic effects o f 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 o f 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. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 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 o f 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. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 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 o f organizational commitment. Although organizational commitment and intention to turnover formed independent scales, the fact that three items were dropped from a shortened version o f Porter et al.’s (1974) scale to eliminate cross-loading with intention to turnover suggests that the measure contains element o f 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 o f 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 o f the company and the employee sample studied. Most importantly, the company experienced substantial downsizing in recent history, Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 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 o f job security may have heightened the employee’s awareness of the benefits o f 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 o f 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 o f 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% o f 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 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 106 12% o f 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 o f 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 o f manager and peer support that are significant predictors o f 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 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 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 o f 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. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 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 o f 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 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 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 during 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, & Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 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 o f 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 o f the skills developed. The findings also substantiate the role of rewards in determining the Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. I l l retention effects o f providing general skills. These new ways o f 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. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 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 o f Vocational Behavior, 49, 252-276. __________________ . 1990. The measurement and antecedents of affective, continuance, and normative commitment to an organization. Journal o f 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 o f Labor Economics, 13(3), 401-426. Bartol, K. 1979. Professionalism as a predictor o f organizational commitment, role stress, and turnover: A multidimensional approach. Academy o f Management Journal, 22 (4), 815-837. Bateman, T., & Strasser, S. 1984. A longitudinal analysis o f the antecedents of organizational commitment. Academy o f Management Journal, 27, 95-112. Becker, G. 1965. Human Capital. Chicago, II: University o f Chicago Press. Begley, T., & Czajka, J. 1993. Panel analysis o f the moderating effects of commitment on job satisfaction, intent to quit, and health following organizational change. Journal o f Applied Psychology, 78(4), 552-556. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 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 o f Applied Psychology, 82(6), 845- 857. Bishop, J. 1997. What we know about employer-provided training: A review of the literaturz. 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 o f Vocational Behavior, 12, 88-103. Blau, G., & Boal, K. 1987. Conceptualizing how job involvement and organizational commitment affect turnover and absenteeism. Academy o f Management Review, 12, 288-300. Bluedom, A. 1982. A unified model of turnover from organizations. Human Relations, 55(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, 75(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 o f Management Journal, 56(5), 1140-1157. Cotton, J., & Tuttle, J. 1986. Employee turnover: A meta-analysis and review with implications for research. Academy o f Management Review, 77(1), 55-70. Davy, J., Kinicki, A., & Scheck, C. 1997. A test of job security’s direct and mediated effects of withdrawal cognitions. Journal o f Organizational Behavior, 18, 323-349. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 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 o f 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 o f 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 o f adult students: Factor stability and construct validity o f the educational participation scale. Adult Education Quarterly, 47(1), 29-40. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 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 o f 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 o f Applied Psychology, 73, 606-613. Griffeth, R., Horn, P., & Gaertner, K. 2000. A meta-analysis o f antecedents and correlates of employee turnover: Update, moderator tests, and research implications for the next millennium. Journal o f 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, 52(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 o f Management Journal, 30(3), 542-552. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 116 Horn, P., Caranikas-Walker, F., Prussia, G., & Griffeth, R. 1992. A meta-analytical structural equations analysis of a model o f employee turnover. Journal o f Applied Psychology, 77(6), 890-909. Horn, P., & Griffeth, R. 1991. Structural equations modeling test o f a turnover theory: Cross-sectional and longitudinal analyses, Journal o f 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 o f 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 o f Management Journal, 38(3), 635-672. Huselid, M., & Day, N. 1991. Organizational commitment, job involvement, and turnover: A substantive and methodological analysis. Journal o f Applied Psychology, 74(3), 381-391. Jackofsky, E. 1984. Turnover and job performance: An integrated process model. Academy o f Management Review, 9, 606-611. Jackofsky, E., Ferris, K., & Breckenridge, B. 1986. Evidence o f a curvilinear relationship between job performance and turnover. Journal o f 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 o f Management, 79(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 o f Organizations, New York, NY: Wiley. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 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 o f learning outcomes to new methods of training. Journal o f Applied Psychology, 78(2), 311-329. Krueger, A., & Rouse, C. 1998. The effects of workplace education on earnings, turnover, and job performance. Journal o f Labor Economics, 16(1), 61-94. Lawler, E., Mohrman, S., & Benson, G. 2001. Organizing fo r 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 o f Managerial Issues, 11(4), 493-513. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 118 Lee, T., & Mitchell, T. 1994. An alternative approach: The unfolding model of voluntary employee turnover. Academy o f Management Journal, 19(1), 51- 89. Lee, T., Mitchell, T., Holton, B., McDaniel, L., Hill, J. 1999. The unfolding model of voluntary turnover. Academy o f 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 o f turnover. Academy o f 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 o f Management Journal, 24(1), 31-48. Lengermann, P. 1996. How long do the benefits o f training last? Evidence o f 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 o f Human Resources, 34(4), 710-733. __________________ . 1998. Dividing the costs and returns to general training. Journal o f 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 o f 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 o f Management Annual Meeting. Toronto, Canada, August. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 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 o f perceived environment, perceived outcomes, and person variables in relationship to voluntary development activities by employees. Journal o f Applied Psychology, 79(1), 3-14. Meister, J. 1998. The quest for lifetime employability. Journal o f 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' o f organizational commitment: Some methodological considerations. Journal o f 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 o f the commitment that counts. Journal o f Applied Psychology, 74, 152-156. Meyer, J., & Smith, C. 2000. HRM Practices and organizational commitment: Test of a mediation model. Canadian Journal o f Administrative Sciences, 17(4), 319-331. Mobley, W., Griffeth, R., Hand, H., & Meglino, B. 1979. Review and conceptual analysis o f 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 o f Management Journal, 56(6), 1430-1464. ___________________. 1989. Introducing survival analysis to organizational researchers: A selected application to turnover research. Journal o f Applied Psychology, 74(2), 280-292. Mowday, R., Porter, L., & Steers, R. 1982. Employee-Organization Linkages. New York, NY: Academic Press. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 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 o f 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 o f Applied Psychology, 71, 492-499. Porter, L., Steers, R., Mowday, R., & Boulian, P. 1974. Organizational commitment, job satisfaction, and turnover among psychiatric technicians. Journal o f Applied Psychology, 59, 603-609. Recruiting and Retaining Employees: Using Training and Education in the War fo r Talent [American Society for Training and Development]. 1999. Alexandria, VA: American Society for Training and Development. Romzek, 1989. Personal consequences of employee commitment. Academy o f 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(4), 377-390. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 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 o f Vocational Behavior, 38, 237-287. Schwarzwald, J., Koslowsky, M., & Shalit, B. 1992. A field study of employees’ attitudes and behaviors following promotion decisions. Journal o f Applied Psychology, 77^,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 o f 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 o f Occupational and Organizational Psychology, 69, 315-326. __________________ . 1995. Organizational commitment, turnover and absenteeism: An examination of direct and interaction effects. Journal o f 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. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 122 Tannenbaum, S., Mathieu, J., Salas, E., & Cannon-Bowers, J. 1991. Meeting trainees' expectations: The influence o f training fulfillment on the development of commitment, self-efficacy and motivation. Journal o f Applied Psychology, 76, 759-769. Tannenbaum, S., & Yukl, G. 1992. Training and development in work organizations. Annual Review o f 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 o f 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 o f 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 o f 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 o f 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 o f 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 o f Management Journal, 40: 82-111. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Wexley K. & Latham, G. 2002. Developing Training and Human Resources Organizations, Upper Saddle River, N.J.: Prentice Hall. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 124 APPENDIX 1 Summary of Hypotheses HI -H 6 H 7-H 9 H 10-H 12 T uition-reimbursement Company-sponsored training On-the-job training Org. Commitment Intent to Turnover Actual turnover Rewards: Promotion or Merit Pay Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 125 APPENDIX 2 Survey Measures During the past year, please indicate the number o f 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 classes or programs .......... 2. For-credit college courses............... ........... 3. Company seminars, conferences or knowledge-sharing networks .......... 4. External conferences and knowledge-sharing..................................... 5. Structured on-the-job training .......... 6. Special assignments....................... ........... 7. Visiting with customers, suppliers, and partner companies................... ........... Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 126 APPENDIX 3 Rotated Component Matrix8 Components I am proud to tell others that I am part of this organization. 1 .768 2 I find that my values and the organization’s values are very similar. .711 I talk up this organization to my friends as a great place to work. .815 I am extremely glad to have chosen this organization .772 For me this is the best of all organizations for which to work. .700 This organization inspires the very best job performance. .774 I would be willing to change companies for career advancement now. .911 I will look for a new job outside my organization in the next year. .781 aFactor loadings less than .4 are suppressed. Extraction Method: Principal Component Analysis. Rotation Method: Varimax with Kaiser Normalization. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 127 APPENDIX 4 Descriptive Statistics: Survey Sample Variable N Minimum Maximum Mean Std. Deviation Organizational commitment 963 1 5 3.16 0.77 Intention to turnover 970 1 5 3.09 1.10 Days tuition-reimbursement classes 979 0 200 9.04 24.17 Days on-the-job training 979 0 300 3.60 16.27 Days company training classes 979 0 66 6.87 7.59 Earned a degree 979 0 1 0.10 0.30 Took classes 979 0 1 0.41 0.49 Number of promotions 979 0 4 0.53 0.68 Number of 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 0.21 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 0.21 0.41 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 128 Descriptive Statistics: Archival Sample-All Employees Variable N Minimum Maximum Mean Std. Deviation All quits 12360 0 1 0.10 0.30 All turnover any reason 12360 0 1 0.30 0.46 Received stock award 12360 0 1 0.09 0.28 Received promotion 12360 0 1 0.37 0.48 Received 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 0.10 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 0.10 0.30 Grade 6 (Manager 3) 12360 0 1 0.06 0.24 Business Unit 1 12360 0 1 0.12 0.33 Business Unit 2 12360 0 1 0.04 0.20 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Descriptive Statistics: Archival Sample-1996 Actives 129 Variable N Minimum Maximum Mean Std. Deviation All quits 9243 0 1 0.08 0.28 All turnover any reason 9243 0 1 0.33 0.47 Received stock award 9243 0 1 0.10 0.29 Received promotion 9243 0 1 0.37 0.48 Received 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 0.21 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 0.11 0.32 Grade 6 (Manager 3) 9243 0 1 0.07 0.26 Business Unit 1 9243 0 1 0.10 0.30 Business Unit 2 9243 0 1 0.04 0.19 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission. APPENDIX 5 Correlations 1996 - 2000 Archival Sample N = 12360 I. 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 1.0 0 5. Merit pay 1996-2000 -0.15 0 .1 0 0 .1 2 0.15 1.00 6 . Promotion after degree -0 .0 2 1.0 0 -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 1.0 0 9. Grade 1 0.13 0 .0 0 0 .0 0 0 .0 2 -0.27 0 .0 2 -0.01 -0.25 1.0 0 1 0. Grade 2 0 .0 2 0 .0 2 0.05 0.1 1 -0.19 0 .0 2 0 .0 0 -0.28 -0.07 1.00 1 1. Grade 3 0.08 0 .0 2 -0.03 -0 .1 0 -0 .0 2 0 .0 0 0.01 -0.04 -0 .1 2 -0.19 1.00 1 2. 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 1.0 0 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 1.0 0 U ) O Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission. Correlations 1996 Archival Sample N = 9243 1. 2 . 3. 4. 5. 6 . 7. 8 . 9. 1 0. 11. 1 2. 13. 14. 1. Quit 1.0 0 2 . Earned degree 0.04 1 .0 0 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 .0 0 6 . Promotion after degree -0 .0 2 1.0 0 -0.15 0.25 0.04 1.0 0 7. Merit pay after degree 0 .0 0 0.80 -0 .2 0 0.17 0.07 0.89 1.0 0 8 . Seniority in years -0.27 -0.21 -0.14 -0.34 -0 .0 2 -0.16 -0.16 1.0 0 9. Grade 1 0 .0 2 0.05 0 .0 0 0.07 -0.05 0 .1 0 0.05 -0.07 1.0 0 1 0. Grade 2 0.06 0.07 0 .0 2 0.08 -0.06 0.06 0.05 -0.09 -0 .0 2 1.00 11. Grade 3 0.11 0.03 -0.04 -0 .1 0 -0.07 0 .0 0 0 .0 2 -0.05 -0.05 -0 .1 2 1.00 1 2. Grade 4 -0.06 -0.05 0.01 -0.05 0.07 -0.05 -0.04 0.07 -0.08 -0.18 -0.45 1.0 0 13. Grade 5 -0.04 0.01 0 .0 0 0.17 0.04 0.05 0 .0 2 0 .0 0 -0.03 O O oo o k > o -0.29 1 .0 0 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. 10. 1. O rg. commitment 1.00 2. Intent to turnover -0.49 1.00 3. Quit -0.13 0.17 1.00 4. #days tuit’n reimburse’t -0.02 0.12 0.01 1.00 5. #days structured O JT 0.10 -0.11 0.07 0.11 1.00 6. #days company classes 0.06 -0.03 0.01 0.00 0.18 1.00 7. Earned degree 0.01 0.11 0.02 0.11 -0.01 -0.05 1.00 S . Took classes -0.06 0.13 0.02 0.17 -0.03 0.08 -0.27 1.00 9. Promoted 1996-2000 0.01 0.16 0.04 0.11 0.04 -0.01 0.18 0.08 1.00 10. M erit pay 1996-2000 -0.19 0.08 0.03 -0.03 -0.11 -0.09 0.11 0.08 0.20 1.00 11. Promotion after degree 0.03 0.04 -0.02 0.05 0.04 0.00 0.61 -0.17 0.23 0.03 12. M erit pay after degree 0.01 0.07 0.00 0.01 -0.04 -0.05 0.76 -0.21 0.12 0.14 13. Seniority in years -0.15 -0.15 -0.15 -0.13 -0.09 -0.07 0.02 -0.12 -0.07 0.49 14. G rade 1 0.07 0.00 0.30 0.05 0.10 0.17 -0.07 0.02 -0.05 -0.32 15. Grade 2 0.12 0.00 0.00 0.06 0.04 0.10 0.02 0.08 0.13 -0.28 16. Grade 3 -0.01 0.01 0.02 0.07 -0.03 -0.13 -0.01 0.05 -0.11 -0.05 17. Grade 4 -0.17 -0.02 -0.08 -0.08 -0.01 -0.02 0.00 -0.07 -0.14 0.23 18. Grade 5 0.03 0.01 -0.09 -0.03 -0.03 -0.04 0.08 -0.03 0.20 0.17 19. Grade 6 0.05 -0.01 -0.07 -0.04 -0.05 -0.02 -0.04 -0.04 0.04 0.14 20. Bus. Unit 1 0.14 -0.20 0.02 -0.04 0.01 0.15 -0.05 0.04 -0.09 -0.15 21. Bus. Unit 2 -0.17 0.23 0.12 -0.01 -0.01 0.10 -0.01 0.05 0.04 0.10 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Correlations 2000 Survey Sample N = 979 Cont’d 133 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 1 . 2 . 3. 4. 5. 6 . 7. 8 . 9. 10. 11. 1.00 12. 0.44 1.00 13. 0.00 0.03 1.00 14. -0.06 -0.07 -0.29 1.00 15. 0.07 -0.04 -0.35 -0.12 1.00 16. -0.02 0.01 -0.13 -0.15 -0.21 1.00 17. -0.03 -0.01 0.30 -0.21 -0.31 -0.38 1.00 18. 0.03 0.10 0.17 -0.11 -0.16 -0.20 -0.29 1.00 19. 0.02 0.00 0.19 -0.09 -0.13 -0.15 -0.22 -0.12 1.00 20. -0.05 -0.03 -0.15 0.07 0.13 -0.07 0.03 -0.06 -0.13 1.00 21. 0.00 -0.05 0.03 0.06 -0.02 0.00 0.01 -0.06 0.02 -0.51 1.00 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 134 APPENDIX 6 P-P Plots for Training Time Transformations On-the job training 1.00 .75 .50 .25 1.00 .75 .50 .25 0.00 Observed Cum Prob 1.00 ln(On-the-job training) L U 0.00 Observed Cum Prob Transform s: natural log 1.00 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 135 Company-training classes 1 .0 0 .75 .50 s CL I .25 O " O Q ) o 0 ) Q. uj o.oo .50 .25 .75 1.00 0 .0 0 Observed Cum Prob ln(Company-training classes) 1.00 .75 .50 .25 ■ O LU 0.00 .50 0 .0 0 .25 1.00 .75 Observed Cum Prob Transform s: natural log Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 136 Tuition-reimbursement classes .75 .50 .O O Q . I .25 O * o 0 ) o CD Q. UJ 0.00 1.00 .50 0.00 .25 .75 Observed Cum Prob ln(Tuition-reimbursement classes) 1.00 .75 .50 .25 ■ o CL x L U 0.00 .25 .50 1.00 0.00 .75 Observed Cum Prob Transform s: natural log Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 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 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 138 Usefulness to Future Marketability: On-the-job Training 200 ----------------------------------------------------------------------------------------------------------- Usefulness to Future Marketability Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 139 APPENDIX 8 Results of Hierarchical Regression for Organizational Commitment: Archival Data Step 1 Step 2a Step2b Beta t Beta t Beta t Tenurea -0.11 -2.84 ** -0.11 -2.88 ** -0.07 -1.72 + Grade 1 -0.17 -2.67 ** -0.17 -2.67 ** -0.13 -1.91 + Grade 2 0.01 0.16 0.01 0.20 0.02 0.45 Grade 4 -0.08 -1.53 -0.08 -1.53 -0.06 -0.98 Grade 5 -0.02 -0.34 -0.01 -0.26 0.02 0.33 Grade 6 0.02 0.43 0.02 0.43 0.05 1.02 Business Unit 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 0.00 -0.13 -0.02 -0.41 0.01 0.17 Took classes -0.08 -2.50 * -0.06 -1.53 -0.10 -1.49 Promoted 1996-2000 0.00 -0.09 Progress X promoted -0.03 -0.53 Degree then promoted 0.02 0.53 Merit pay 1996-2000 -0.12 -2.63 ** Progress X merit pay 0.05 0.66 Degree then merit pay 0.01 0.15 R2 (adjusted R2 ) .134 (.125) .145 (.133) . 143 (. 132) F 149 *** 12 4 *** 12.2 *** D f 969 969 969 N 985 985 985 " standardized + P < . 10 * P < .05 **P < .01 ***P < .001 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 140 APPENDIX 9 Hazard Function Graphed for Each Level of Covariates Tenure C D N C D X 0 > T O D E d o .o Tenure Median Split • High ° Low 50 30 40 60 0 10 20 Months Grade 1 .06 .05 .04 .03 ro ro .02 X < D > ra .01 3 E o O . o o Grade 1 Yes No 20 30 40 0 10 50 60 Months Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 141 Grade 2 .07 .06 .05 .04 *o -03 a 0 3 X .02 d ) > Grade 2 r o 1 3 E 3 o .01 1 Yes 0.00 20 60 10 30 40 50 0 Months Grade 3 .08 .07 .06 .05 .04 S - 0 3 < U .02 > Grade 3 ■ * — * ro 3 E 3 o .01 : Yes 0.00 20 0 10 30 40 50 60 Months Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 142 Grade 4 .06 .05 .04 .03 0 £ '°2 C D > ro -01 3 E o 0.00 Grade 4 Yes 60 30 50 0 10 20 40 Months Grade 5 .06 .05 .04 .03 ■a .02 Grade 5 .01 Yes ° No O 0.00 60 50 0 20 30 40 10 Months Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 143 Grade 6 .07 .06 .05 .04 • O 0 3 a ro X .02 < D > Grade 6 1 Yes 0.00 0 10 20 30 40 50 60 Months Promotion .07 .06 .05 .04 .03 .0 2 - Promotion Promoted o 0.00 n No FVo motion 0 10 20 30 40 5 0 6 0 Months Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 144 Merit Award •o L — a 1 d ) > Merit Award T O 3 E o O .o Merit ° No Merit 50 60 40 20 30 0 10 Months Degree Progress .07 .06 ■ .05 .04 .03 .02 Degree Progress .01 Progress , j— ° No R o g re ss O 0.00 60 20 30 40 50 0 10 Months Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 145 Earned Degree .06 .05 .04 .03 ~ o .0 2 - ( D > Earned Degree « * — * T O 3 E 3 o .01 Earned Degree No D egree 0.00 60 40 50 30 10 20 0 Months Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 146 APPENDIX 10 Logistic Regression Results Step 1 B Exp(B) Tenure3 -0.35 0.71 Grade 1 2.02 7.52 Grade 2 0.17 1.19 Grade 4 0.62 1.87 Grade 5 -6.23 0.00 Grade 6 -6.19 0.00 Business Unit 1 1.48 4.40 Business Unit 2 2.25 9.52 In (Tuit’n-reimbursement) In (On-the-job training) In (Company classes) X s D f N Wald Sig Step 2 B Exp(B) Wald 1.43 -0.35 0.70 1.44 13.35 ** 2.14 8.51 14.62 *** 0.08 0.28 1.32 0.21 1.53 0.57 1.76 1.25 0.20 -7.21 0.00 0.10 0.12 -7.19 0.00 0.06 3.82 * 1.72 5.59 4.94 * 8.71 ** 2.49 12.10 10.13 ** 0.02 1.02 0.05 -0.13 0.88 0.67 -0.22 0.80 1.85 82.95 *** 86.12 *** 8 1 1 979 979 "standardized + P < . 10 * P < .05 * * P < .01 ***P< .001 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

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

PDF

Information systems capabilities and indirect relational value: Sustaining networked organizations

PDF

Does adding nonfinancial value drivers to a summary financial measure improve the learning and performance of managers?

PDF

Inventory and pricing models for perishable products

PDF

Demographic diversity, team process, and team performance: Assessing moderator effects of cognitive conflict management practices and task interdependence

PDF

Effects of redundancy in media coverage on nonprofessional investors' earnings forecasts

PDF

An investigation of repeat visit behavior on the Internet: Models and applications

PDF

Androgyny and transformational leadership: Effects of gender and sex -role identity in the collectivistic context of Taiwan, R.O.C.

PDF

Information diffusion and the boundary of market efficiency

PDF

Essays on technology innovation and new product development strategy

PDF

Board involvement in monitoring and strategy making: Antecedents and consequences

PDF

Essays on technological evolution and financial returns to innovation

PDF

Corporate ethical orientation and crisis management before and after September 11, 2001

PDF

Developing business communication skills: Leveraging stage versus global processes of change in skills improvement approaches

PDF

Customer satisfaction, customer bargaining power, and financial performance

PDF

Essays in supply chain management

PDF

A typology of corporate environemental strategy and its driving factors in jultinational corporations

PDF

Defining the advice -seeking request in electronic solicitations in boundary spanning environments

PDF

"Pay no attention to those men behind the curtains": An ethical examination of Los Angeles charter reform activities 1996--1999 by use of crisis management

PDF

Biomechanical and neuromuscular aspects of non-contact ACL injuries: The influence of gender, experience and training

Asset Metadata

Creator Benson, George Sinclair (author)

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

School Graduate School

Degree Doctor of Philosophy

Degree Program Business Administration

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

Tag business administration, management,OAI-PMH Harvest

Language English

Contributor Digitized by ProQuest (provenance)

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-c16-211982

Unique identifier UC11339211

Identifier 3073745.pdf (filename),usctheses-c16-211982 (legacy record id)

Legacy Identifier 3073745.pdf

Dmrecord 211982

Document Type Dissertation

Rights Benson, George Sinclair

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