Lynn Kalnbach
Gene Griffin
Upper Great Plains
Transportation Institute
North Dakota State University
Fargo, North
Dakota
December 2002
The Upper Great Plains Transportation Institute, North Dakota State University, and the research and support staff responsible for the development of this report wish to express our sincerest thanks to the University Transportation Centers Program for sponsorship of this research.
The authors would also like to thank the four trucking companies that participated in this project:
The contents of this report reflect the views of the authors, who are responsible for the facts and accuracy of the information presented herein. This document is disseminated under the sponsorship of the Department of Transportation, University Transportation Centers Program, in the interest of information exchange. The U.S. government assumes no liability for the contents or use thereof.
TABLE OF CONTENTS CHAPTER ONE: INTRODUCTION.......................................1 Employee Embeddedness........................................2 The Unfolding Model of Turnover..............................3 Definitions and Clarification of the Unfolding Model.........4 Decision Paths of the Unfolding Model........................5 Results of Studies Conducted to Test the Unfolding Model.....6 Issues to Consider in Predicting Voluntary Turnover..........9 Hypotheses...................................................10 CHAPTER TWO: RESEARCH METHODOLOGY...............................13 Survey Process...............................................13 CHAPTER THREE: ANALYSES AND RESULTS.............................17 General Results..............................................17 Differences Between Drivers Who Stayed and Those Who Left....20 Hypothesized Relationships...................................21 Possible Interventions.......................................24 CHAPTER FOUR: CONCLUSIONS AND RECOMMENDATIONS...................27 Future Research..............................................29 REFERENCES......................................................31
Voluntary turnover rates among truckload carriers are extremely high, ranging from 50 to more than 100 percent annually. These high turnover rates result in elevated costs for carriers in terms of recruitment and training as well as costs associated with reduced productivity and decreased customer satisfaction resulting from inexperienced drivers. Although much research has been conducted to determine the relationship between job satisfaction of drivers and the likelihood of them leaving or intending to leave an organization, research addressing other reasons why drivers leave their organizations has been lacking. As a result, this project was designed to increase understanding of voluntary turnover of drivers.
To accomplish this objective, written surveys were used to gather data from drivers. Surveys were completed by drivers at two different times during the study. The first survey, assessing personal and situational characteristics, was completed by applicants for the job of driver during the hiring process. The second survey was distributed only to those drivers who quit their jobs during the study in order to ascertain how they made their decisions to leave their organizations. Although the actual number of drivers that completed both surveys was very small, analyses were conducted to determine how those that left the organization compared to those who stayed with the organization. Results and recommendations are provided in subsequent sections of this report.
The trucking industry suffers from high levels of voluntary employee turnover. The rates of turnover are often as high as 50 to 100 percent annually. Unfortunately, these high rates can lead to substantial negative consequences for organizations in this industry. Griffeth and Hom, experts in the area of employee retention, concluded that there are many costs associated with employee turnover that are not necessarily considered by the organization, but that can really affect the organization's bottom line. Besides the costs associated with actual separation of the employee and recruiting, hiring, and training of new employees, other costs may not be as easy to measure. These additional costs may result from a reduction in quality of service given to customers, a loss of business to competitors, a tendency for remaining employees to quit as well, and a lack of talented employees who could become future leaders of the organization. As a result, the actual cost of driver turnover may be higher than what top managers in these trucking companies had thought, especially considering the high volume of voluntary turnover most are experiencing.
For organizations in the trucking industry to be competitive, they must have a better understanding of how and why drivers are leaving their jobs. They also would like to know if there are any steps they could take to prevent losses of qualified drivers. As a result, this study will focus on a different model of turnover, which takes into account the fact that many people leave their jobs for reasons other than dissatisfaction with the job. In fact, there are many aspects of a job and organization that encourage employees to stay. The organization's goal is to determine what those factors are and how they influence employees' intentions for staying or leaving the organization.
Mitchell, Holtom, Lee, Sablynski, and Erez (1999a) used the term embeddedness to describe the attachment employees have to the organization and surrounding environment. It includes elements of personal and organizational fit, sacrifice, and links. More specifically, fit refers to an individual's perceived match or comfort with the organization and with the surrounding environment, including the community. Sacrifices are the benefits and perks that would be given up if the individual chooses to leave the organization. They include travel opportunities, rewarding work assignments, and retirement benefits. According to Lee and Mitchell, links refer to the formal and informal ties that an individual has to the job and organization, and to the community. Examples of links include memberships in churches and other community organizations, as well as any work-related groups. Other embeddedness factors not mentioned by Lee and Mitchell include personal and situational characteristics such as family responsibilities, a person's level of openness to experience, conscientiousness, and willingness to take risks. As a result of these factors, people may be encouraged to stay rather than to leave their jobs.
Lee and Mitchell (1994) created the unfolding model of turnover to better define and classify the process people go through in making decisions to leave their organizations. However, independent from their research on this model, Lee and Mitchell and their colleagues also began to research how their concept of embeddedness affects individuals's decisions to leave their organizations. Although Lee and Mitchell have not integrated the two lines of research, evidence indicates that the connections should be made. As a result, one goal of this study is to extend the unfolding model of turnover by including embeddedness factors as predictors of how individuals decide to quit their jobs and organizations.
Lee and Mitchell's (1994) new model of turnover addresses many limitations of other more traditional theories of turnover. In their unfolding model of turnover, Lee and Mitchell proposed several reasons why individuals quit their jobs that are not directly due to dissatisfaction or negative job attitudes, which are often the cornerstone of other voluntary turnover models (Hom & Griffeth, 1991, Price & Mueller, 1986). The unfolding model consists of four main paths. One of the paths (Path 4) is separated into two subpaths, which can be used to classify turnover decisions. These two decision subpaths (Paths 4a and 4b) describe quit decisions that result from feelings of job dissatisfaction. The other three decision paths (Paths 1-3) involve factors such as shocks, image violations, and/or searches for alternatives to the job rather than diffuse feelings of dissatisfaction. Inclusion of these paths for turnover decisions that do not necessarily result from job attitudes or dissatisfaction represents a significant theoretical advance in understanding voluntary turnover.
Rather than proposing that all people go through a deliberate process in making the decision to leave an organization, Lee and Mitchell's (1994) model utilizes ideas from the image theory to suggest that some people may make relatively quick decisions to leave that do not require much cognitive deliberation. In addition to incorporating ideas from image theory, Lee and Mitchell also included in their model many of the same factors upon which more traditional models of turnover were founded. Perceptions and consideration of alternatives to the job are taken into account in both the unfolding model and traditional models, as are job attitudes and withdrawal cognitions. Yet, the focus of the unfolding model may make it more useful than other theories from a practical standpoint as well. It differs from other turnover models in that it can be used for understanding and classifying quit decisions into one of four paths. If we can determine which variables best predict employees most likely to make each type of quit decision, then, organizations may be able to better intervene and retain productive employees. This point will be discussed in greater detail in a later section of the paper.
Before describing each of the paths of the unfolding model, it is necessary to define the main components of the unfolding model. The first component is what Lee and Mitchell referred to as a shock. Lee, Mitchell, Wise, and Fireman (1996: 6) describe a shock as "a particular, jarring event that initiates the psychological analyses involved in quitting." The shock prompts a process of interpretation and must be integrated into the person's system of beliefs, values, and images. Examples of shocks include marriages, job transfers, serious conflicts with coworkers, and unsolicited job offers. Whether events are even perceived to be shocks varies with people's beliefs, values, and frame of reference.
The second component of the model and of several decision paths involves image violations. According to Lee and Mitchell, these violations result from some event that leads an individual to determine that he or she cannot integrate his or her values with the shock. As a result, the individual perceives a lack of fit with the organization or with the job and decides to either change the image or to the leave the organization. Some general dissatisfaction may result from image violations; however, in the unfolding model, these violations are discussed mainly as resulting from some type of shock.
Scripts also are an important component of the unfolding model of turnover. Scripts are cognitive plans for automatic behavioral sequences in well-known situations. The nature of scripts seems to suggest that they are most likely to develop out of past experiences in similar situations. One of the paths of the unfolding model focuses primarily on this scripted behavior.
Another important component of the model is a search for and/or evaluation of alternatives to the job. In the Lee, Mitchell, Holtom, McDaniel, and Hill (1999) paper they specifically recognized work and non-work alternatives. Non-work options may include going to graduate school or deciding not to work outside of the home. Lee et al. (1999) also modified their original hypotheses about search and evaluation. They recognized that the processes could be intertwined or that each could occur independently. As a result, individuals could be faced with an alternative to their job for which they did not search, but to which they gave consideration, and alternatively, they also could search for alternatives but not find any to be evaluated.
The main components and distinguishing features of the decision paths are:
Decision paths 1, 2, and 4a can be distinguished from paths 3 and 4b by a lack of search for or evaluation of alternatives to the job. Although individual decisions classified into paths 1, 2, and 4a may include some consideration of general perceptions of alternatives to the job (e.g., labor market or economic conditions), specific alternatives are not sought or considered. Path 1 decisions are easily distinguishable from those of other paths in that they are the only ones that include a well-defined script or plan of behavior in response to a shock. Paths 4a and 4b also are easily distinguishable from the others because they are the only paths that do not include specific shocks or events that prompt thoughts of quitting but instead focus on gradual dissatisfaction.
Lee and Mitchell, and several of their colleagues, have conducted studies to test parts of their unfolding model of turnover. They have found that, in general, individuals's turnover decisions can be classified into one of the four main paths they specified. They also have found that more people leave jobs and organizations due to shocks (Paths 1-3) than due to dissatisfaction (Paths 4a & 4b).
Lee et al. (1996) conducted the first empirical test of the unfolding model with a sample of nurses who had voluntarily quit their jobs at hospitals. They interviewed the nurses using questions that assessed major components of the unfolding model, including shocks and search for alternatives to the job. They also sent out follow-up surveys to the nurses to assess reliability of the information obtained in the interviews. Responses from the interviews were categorized into one of the decision paths by two of the paper's authors who had not conducted the interviews. The responses from 11 of the 44 nurses did not fit a particular path. Overall, the results of Lee et al. (1996) showed that 20 of the 33 classified turnover decisions were due to shocks rather than to dissatisfaction.
The Lee et al. (1999) study was conducted with a sample of 229 individuals who had quit their jobs at one of the Big 6 public accounting firms. Information about factors that led up to decisions to quit was obtained through a questionnaire that included items assessing major components of the unfolding model. Quit decisions then were categorized into one of the main decision paths by four authors and a volunteer who had no connection to the study. The categorization process was based on predetermined decision rules for the participants' responses to questionnaire items. The results of Lee et al. (1999) were even more skewed than results of the Lee et al. (1996) study, with 149 of the 212 classified decisions resulting from shocks rather than from dissatisfaction. This is not to say that those who left via dissatisfaction paths did not experience any events that could be considered shocks. However to be classified as dissatisfaction paths, reasons for leaving must include some form of dissatisfaction that was not the result of a single, particular event.
Although modifications in the unfolding model did improve the rate of classification of the turnover decisions in the Lee et al. (1999) study when compared with the study done in 1996, 17 individuals in the 1999 study could not be classified. Twelve of these unclassified decisions could have been categorized as path 3 decisions if they had reported image violations. As a result, there seems to be many individuals who report shocks, but do not experience image violations even though they do search for or evaluate alternatives to the job.
In the study conducted by Mitchell, Holtom, and Lee (1999b) with 232 grocery store employees, embeddedness was measured along with job satisfaction, organizational commitment, job involvement, and intent to turnover. In this study, embeddedness was measured with 43 items loaded onto six factors. These six factors represented job-related and non job-related components of fit, links, and sacrifice which Mitchell et al. hypothesized as making up the construct of embeddedness. In this study, embeddedness was considered a global measure and as a composite of the six more specific factors. Mitchell et al. (1999b) found that embeddedness had significant incremental prediction of voluntary turnover over job satisfaction, organizational commitment, and job involvement. More specifically, fit with the organization and sacrifices from leaving the organization factors showed the strongest relationships with people's intention to leave.
Although results of the initial studies support much of Lee and Mitchell's unfolding model, they have yet to conduct a full test of their model. Their studies of classifying turnover decisions into one of the four main paths described previously and studies investigating appropriateness of predicting voluntary turnover with measures of their concept of embeddedness have remained relatively separate. Lee and Mitchell (1994) initially made some informal predictions about what characteristics of the individual may lead to use of the different decision paths, but have not formally hypothesized or tested any of these relationships. Since they have only begun to examine how the different paths may be differentially predicted by various factors including those associated with their concept of embeddedness, age, marital status, and occupation, making these links is the next important step. This will be the focus of the current study.
From a practical standpoint, many organizations would be most interested in assessing constructs that could help predict turnover decisions, and even the path for the turnover decision, during the hiring process or at an early stage of employment. If an organization is able to predict which applicants are more likely to leave the organization via path 1, meaning that the decision to leave is a scripted or planned response to a particular shock, the organization may be able to avoid hiring those applicants. The organization may choose to measure in the selection process the characteristics that predict path 1 turnover decisions. It may then decide not to hire those applicants who, based on those characteristics, would be more likely to quit via path 1 to reduce unpredictable or unavoidable turnover.
Alternatively, the organization may attempt to intervene and try to retain employees who would be most likely to use other decision paths in deciding to leave the organization. For example, if the organization hires individuals who would be predicted to quit via paths including a search for or consideration of alternatives to the job, the organization may be able to communicate a willingness to work with employees who consider leaving the organization to pursue other alternatives. This type of intervention may include discussions about opportunities for development and career advancement, the possibility of flextime or part-time accommodations, or the possibility of an increase in salary or bonuses in the current organization. The current study addressed this issue further by allowing employees who have left their jobs to describe any potential interventions the organizations could have used to prevent them from leaving their organizations.
Knowing which paths applicants or employees would most likely use in making a decision to quit also may give the organization some information about how much time there may be to intervene before the individuals actually leave the organization. The Lee et al. (1999) study with accounting firms demonstrated, as expected, that the elapsed time between the first thoughts of quitting and the ultimate decision to leave was longer for the dissatisfaction paths (4a & 4b) than for any of the other paths (1-3). This finding supports Lee and Mitchell's suggestion that an organization may have more time to intervene with those employees who gradually become dissatisfied and decide to leave the organization than with those who decide to leave in response to a shock.
Past research suggests various personal and situational characteristics will differentially predict decision paths that people will use in deciding to leave an organization. The following hypotheses outline the predicted differences that will be examined in this study.
Hypothesis 1: Individuals low in conscientiousness, high in risk taking, and high in openness to experience will be more likely to leave an organization without searching for or considering other alternatives (paths 1, 2, and 4a) than will individuals who are high in conscientiousness, low in risk taking, and low in openness to experience.
Hypothesis 2: People who report more responsible behavior in the past (e.g., longer tenure at previous jobs, tendency to save rather than spend money, more experience with work at young age) will be more likely to repeat that type of behavior by seeking out and considering alternatives to the job than those who report less responsible behavior (paths 3 and 4b).
Hypothesis 3: Individuals who have voluntarily left many jobs in the past are more likely to use path 1 and a script for leaving the organization than are people who do not have as much experience in quitting previous jobs.
Hypothesis 4: Individuals who perceive there to be many alternatives to their jobs in general will be less likely than those who perceive there to be few alternatives to leave via paths that include search for or consideration of other jobs or opportunities (paths 3 and 4b).
Hypothesis 5: Low perceptions of fit (compatibility of values) will lead to more quit decisions based on dissatisfaction (4a & 4b) than on particular shocks (1-3).
The methodology for this study relied upon written surveys used to gather data from drivers. Surveys were completed by drivers at two different times during the study. The first survey, assessing personal and situational characteristics, was completed by newly hired truck drivers. The second survey was distributed only to those drivers who quit their jobs during the study in order to determine how they made their decisions to leave their organizations. Survey information was entered into a database and analyzed.
As part of their new-hire paperwork, drivers were asked to complete a questionnaire that would not have any influence on their status with the organization, but was to be used for research purposes only. A letter stating these issues accompanied the initial questionnaire. The questionnaire measured personal characteristics such as conscientiousness, openness to experience, and willingness to take risks. It also assessed respondents' recollections of fit with the organization with items from the Mitchell et al. (1999b) embeddedness questionnaire. It included questions about variables such as general demographics, marital/family status, perceived alternatives to the job, and education and credentials. The items for this questionnaire are shown in Appendix A.
As drivers quit the organization, they were asked to answer some questions in a questionnaire assessing their reasons for leaving the organization. These questionnaire items assessed the presence or absence of shocks, scripts, image violations, job satisfaction, search and/or consideration of other alternatives to the job, and possible interventions that the organization could have taken to prevent the quit decision. Questions used for this study were adapted from Lee et al. (1999) to clarify the meaning of some of the items and to ensure that the items clearly assess their designated features. This modified questionnaire is located in Appendix B.
Respondents were assured that the answers they gave to the questions would be kept confidential and would be used only by the researchers. This precaution was taken to help encourage respondents to answer questions honestly rather than in a socially desirable manner.
For approximately six months after distribution of the original questionnaires, turnover data and exit questionnaire information were collected on a monthly basis. Turnover data were gathered from organizational records. Based on the turnover reports, exit questionnaires were sent to individuals who had quit their jobs. The questionnaires were to be returned in self-addressed, stamped envelopes. This mailing process helped to ensure that each respondent's information was kept confidential from the organization. For those former employees who did not return the exit questionnaire (211 individuals), a follow-up phone call was placed to obtain the reasons for the quit decision.
Once the exit questionnaires were received, each turnover decision was classified into one of the four main decision paths. The guidelines and processes used for classifying turnover decisions in this study modeled those used by Lee et al. (1999). In the 1999 study, Lee and his colleagues created a set of decision rules to follow based on answers participants gave to the questions included in the exit survey. Application of these decision rules resulted in the classification of turnover decisions. A defining feature for a particular decision path was determined to be present if the participant responded yes to one or more of the questions assessing the defining feature. However, the participant's other responses must not contradict the prior yes responses in order to made a correct classification. For example, if a participant responded that he or she used a script or action plan in making the decision and did not contradict the path 1 classification by responding that alternatives were searched for or considered, the decision would have been correctly classified as a path 1 decision.
After all the turnover decisions were classified, analyses were conducted to determine whether there was any support for the hypothesized relationships of specific personal or situational characteristics with particular decision paths. Because the number of drivers who returned their exit questionnaires was small (37 returns), many of the findings should be interpreted with caution. Additionally, qualitative analyses were used to summarize information about possible interventions that could have been used to prevent some of the drivers from leaving their organizations.
Initially, 804 drivers returned the first questionnaire that they completed with their new hire paperwork. Throughout the duration of this study, approximately 31 percent or 247 drivers left their organizations. However, of the number of drivers that left, only 26 exit questionnaires could be included in the study. Eleven additional exit questionnaires were returned, but could not be used as a result of missing identification information or determination that the respondent did not voluntarily leave the organization. Unfortunately, follow up phone calls resulted in few additional responses. In many cases, drivers who had left their organizations were unavailable by telephone for a variety of reasons, including disconnected lines.
The initial questionnaire assessed the personal and situational characteristics of new drivers. Several questions were combined to form the conscientiousness, risk taking, openness to experience, perceived organizational fit, and expectations about benefits variables. For each of these variables, responses on individual items were averaged to obtain a mean value based on a 1 to 5 scale, with a value of 1 being very inaccurate and a value of 5 being very accurate. The mean values were then aggregated across all of the study respondents to determine the overall level of these characteristics in the groups of new drivers entering organizations that participated in the study. The mean values are displayed in the table below.
| Characteristic | Mean |
| Conscientiousness | 3.03 |
| Willingness to Take Risks | 2.50 |
| Openness to Experience | 3.44 |
| Perceptions of Organizational Fit | 4.04 |
| Expectations about Benefits | 3.92 |
The means in the table show that new drivers were in the middle on the conscientiousness dimension, less willing to take risks, and more open to experiences. In addition, drivers had favorable perceptions of fit with the organization and good expectations about benefits associated with the job. Overall the levels of these characteristics are appropriate for what an organization would hope to find in new employees. However, according to Barrick and Mount (1991), personality characteristics such as conscientiousness and openness to experience are reliable predictors of turnover. As a result, organizations typically ould want to have employees who showed higher levels of conscientiousness.
Other interesting statistics from the first questionnaire include:
The main purpose of the exit questionnaire was to determine how drivers made their decisions to leave their organizations. Each of the turnover decisions was classified into one of the paths put forth by Lee and Mitchell (1994). Using their guidelines for classification resulted in the following numbers of each decision path:
| Path | Number of Decisions Classified |
| Path 1 | 3 |
| Path 2 | 2 |
| Path 3 | 13 |
| Path 4a | 0 |
| Path 4b | 1 |
| Unclassifiable | 7 |
Out of the total number of drivers that participated in the study (804), 557 of them stayed with their organization, while 247 of them left their organization within approximately six months. The following analyses were conducted to make comparisons between these two groups of drivers on the personal and situational characteristics that were assessed in the first questionnaire of the study.
The first set of analyses involved comparisons of the two groups'means on the following variables:
Results of these analyses showed that two significant differences between those drivers who stayed and those who left their organizations. The groups were significantly different in level of conscientiousness and the number of self-reported work-related skills. More specifically, the findings indicate that those drivers who were less conscientious and those who reported having more work-related skills were more likely to leave their organizations. The resulting relationship between conscientiousness and leaving the organization is consistent with findings of previous studies (Barrick & Mount, 1991). In addition, the result showing that drivers were more likely to leave when they reported having more work-related skills also can be easily interpreted. The likely explanation for this finding is that drivers having more skills were more likely to have other opportunities available to them, allowing them to quit more easily than if they had few skills.
The other set of analyses that were conducted examined the two groups of drivers in relation to categorical variables, including the driver's marital status, whether the driver owns his/her home, whether the driver tends to save or spend money, and the driver's educational level. Results of these analyses showed that there were no significant associations between these categorical variables and driver job status (whether the drivers stayed or left the company).
All of the hypothesized relationships involved the paths of the turnover decisions from drivers that completed the second questionnaires. Due to the small number of respondents, statistical analyses could not legitimately be used. As a result, hypotheses were examined through the comparison of percentages of quit decisions made via the various paths.
Hypothesis 1 stated that certain personal characteristics would be related to a lack of search for other alternatives when deciding to leave the organization. Specifically, the prediction was that those individuals low in conscientiousness, high in risk taking, and high in openness to experience would be more likely to leave the organization via paths 1, 2, and 4a than those individuals who were high in conscientiousness, low in risk taking, and low in openness to experience. Results showed that when each of the personal characteristics was examined in relation to the path of the quit decision, there was slight support for conscientiousness and openness to experience prediction, but not for risk taking. Those individuals who were low in risk taking were more likely to quit without searching for other alternatives than those who were high on that dimension. Although this finding seems counterintuitive, it may only suggest that these individuals did not perceive quitting without looking for another alternative as a risk. This may have been true for many reasons. One example may be that although they did not search for or evaluate specific alternatives to the job they were leaving, they may have had a general perception that another opportunity would come along at some point.
Hypothesis 2 stated that people who reported more responsible behavior in the past would be more likely than those who reported less responsible behavior to quit only after searching for or evaluating other alternatives to the job. When defining responsible behavior as a tendency to save rather than spend money and having more than one year of early work experience, there was support for this hypothesis. Results showed that whether these two variables were examined together or separately, individuals who were high in the responsible behavior were more likely than those who were low in those behaviors to have quit via path 3, which includes a search and/or evaluation of alternatives to the job.
The prediction stated in hypothesis 3 was that people who had left many jobs in the past would be more likely to use a script in making the decision to quit and to follow path 1. Unfortunately, only three people quit via path 1; two that had left five or fewer jobs and one that had left six or more jobs in the past five years. As a result, no meaningful findings can be reported.
Hypothesis 4 stated that individuals who perceive many alternatives to their jobs would be less likely than those who perceive there to be few or no alternatives to leave the organization via paths involving a search or consideration of other alternatives to the job. Comparisons of the numbers of quits showed that those who perceived many alternatives to the job quit less often via paths 3 or 4b than those who were less optimistic about alternatives. This finding lends support to the hypothesized relationship between perceived alternatives and the likelihood of a search before leaving the organization.
The last hypothesis focused on the relationship between individuals' perceptions of fit and decision path for leaving the organization. More specifically, the hypothesis stated that low perceptions of fit would lead to more quit decisions based on dissatisfaction than on specific shocks. Results showed that none of those who had low perceptions of fit quit due to dissatisfaction; instead, the three individuals that had low perceptions of fit quit via paths that included specific shocks. Although this finding is contrary to what was predicted originally, there may be a simple explanation. In this study, individuals' perceptions of fit were assessed right after the hiring process, and at this time, the overwhelming majority of drivers had positive perceptions of fit with the organizations. This early assessment was done to determine the perceived fit before the day-to-day work and environment had an influence. As a result, these perceptions could have easily changed as the drivers had a chance to become more familiar with the organizations and specific aspects of their jobs. An alternative explanation also could be that the duration of the study was too short to capture this relationship between perceived lack of fit and dissatisfaction with the job or the organization.
In addition to having respondents rate various statements in the second questionnaire, an open-ended question was used to obtain ideas about what the organizations could have done to prevent the individual from leaving. Of the 26 completed second questionnaires included in the study, 22 of them indicated that the respondent believed the organization could have done something to prevent the quit. In addition, many of the drivers' descriptions of the shocks that had led them to make the decisions to quit indicated that there were negative situations with which organizations could have dealt relatively easily.
Analysis of the suggestions that could have prevented some of the drivers from leaving their organizations indicated that many of the drivers felt that they had been misled about various aspects associated with the job. Many suggestions centered around the need for the organization to keep its promises. Three areas for which this was specifically mentioned were pay, time at home, and driving routes. Several drivers did not believe that they were being paid what they had been told they would earn, they were not driving the routes they believe they were promised, and they did not feel that they were getting as much time at home as they were led to believe they would get. The largest source of shocks reported by drivers was a lack of time at home. In some cases, the treatment from someone else in the organization (supervisor, dispatcher, etc.) also contributed to the negative experience of not being able to get home as expected. Such situations demonstrate the need for better communication to new employees to reduce the chance of unrealistic expectations.
Drivers also were concerned about the driver training. Four of the suggestions from drivers focused on training. The specific suggestions included: the need for expanded in-truck training, better trainers, a faster response with trainers, and an opportunity to give feedback about the training. In addition, these issues were not raised by drivers for only one company; they were raised by drivers in three of the four companies that participated in the study. Therefore, training seems to be another area in which each of the organizations could make some changes to retain qualified drivers.
Other shocks experienced by drivers and suggestions for changes in the organizations varied widely and were quite specific. However, two other areas mentioned by more than one driver were the poor condition of equipment and the practice of having to wait in truck stops. Other shocks described by drivers reflected a negative experience with some other person in the organization, whether it was a supervisor, trainer, team leader, dispatcher, or other employee. Whether these issues resulted from inadequate communication, training, or some other variable, the organizations participating in this study could make some small changes that would ultimately help improve driver retention.
This study was designed to further develop the understanding of voluntary employee turnover in the trucking industry. Specific objectives of the study were to 1) quantify types of turnover decisions, 2) make predictions about which drivers would leave and how they came to their decisions, and 3) identify interventions that could lead to improved retention. To accomplish these goals, questionnaires were used to assess drivers' personal and situational characteristics as they began new jobs and to assess the reasons behind their decisions after they left the organizations. Although the number of respondents who returned usable second questionnaires was small, some interesting findings were obtained. Responses indicated that drivers left their organizations for many reasons. However, drivers cited specific shocks that influenced their decisions to leave including poor equipment conditions, having to wait in truck stops, and negative experiences with other individuals in the organization.
Relationships between personal and situational characteristics and whether the drivers left or stayed with their companies were examined. Consistent with results from previous studies, there was a significant relationship between conscientiousness and job status. However, the overall level of conscientiousness of new drivers in this study was relatively low. As a result, the organizations participating in this study could potentially benefit from using a measure of conscientiousness when selecting applicants to become drivers.
The biggest area of concern from the study would seem to be that most drivers who left believed their decisions could have been prevented by the organization. In many cases, drivers' expectations were not met. These unmet expectations contributed to rather quick decisions to leave the organization as opposed to the more gradual nature of decisions to leave due to dissatisfaction. The fact that there were many unmet expectations supports the finding that drivers had high perceptions of fit with the organization in the first questionnaire and that most of them reported image violations in the second questionnaire.
To decrease the amount of turnover resulting from these unmet expectations, organizations in this industry must do the following:
This information should include actual pay that drivers will earn, realistic estimates of the amount of time drivers can expect to be at home, routes that the drivers are most likely to have and for what length of time, and what type of equipment they can expect to have.
Ensure that training is available to all in a timely manner and that trainers have the appropriate skills and knowledge to pass on to new employees. In addition, allow trainees to evaluate the training and trainers to determine when improvements should be made.
Many of the shocks experienced by drivers seemed to have resulted from either miscommunications or a lack of communication. The quality of communication between drivers and dispatchers should be examined so that improvements can be made.
It is clear that drivers have specific suggestions for changes that could help to retain employees. Employees should be encouraged to share this information through the use of a well-developed survey assessing all areas important to organizational health, including training, supervision, working conditions, and performance management.
Since this was a first study looking at the relationships between certain personal and situational characteristics and the paths of drivers' decisions to leave their organizations, many changes could be made to improve future research in this area. One change that would be beneficial is to conduct a longitudinal study using similar procedures over several years. A longer term study would allow researchers to determine whether personal characteristics would have more of an effect over time as drivers made their decisions to leave an organization. For example, one hypothesis for this study was that low perceptions of fit with the organization would lead to more quits due to dissatisfaction rather than to specific shocks. Although the finding for this hypothesis in the study was different than predicted, the relationship may have been supported over time, as drivers may have become gradually dissatisfied with the job and/or the organization.
Additional modifications made to the questionnaire assessing the reasons for the quit decisions also would improve future research in this area. Although the questionnaire was piloted with a group of new drivers to ensure that questions were clear and well-written, there still seemed to be confusion about some of the items for a small group of drivers that responded. Future studies using this method for determining paths of turnover decisions should include extensive testing of the questionnaire before it is utilized with actual participants.
Finally, future research in this area must include a larger number of participants. The return rate of the second questionnaire was extremely low in this study, even with the use of follow-up telephone calls. As a result, the use of statistical analyses was limited. It may be helpful to include some type of reward for returning the second questionnaire in future studies. A chance for a cash reward or a small prize could be enough to encourage some people to respond. However, clarification and specific modifications to items in the second questionnaire also may help to increase the return rate in future studies.
Barrick, M. R., &Mount, M. K. (1991). The big five personality dimensions and job performance: a meta-analysis. Personnel Psychology, 44, 1-26.
Griffeth, R. & Hom, P. (2001). Retaining Valued Employees. Thousand Oaks, CA: Sage Publications.
Hom, P., & Griffeth, R. (1991). Structural equations modeling test of a turnover theory: cross-sectional and longitudinal analyses. Journal of Applied Psychology, 76, 350-366.
Lee, T. W., & Mitchell, T. R. (1994). An alternative approach: the unfolding model of voluntary employee turnover. Academy of Management Review,19, 51-89.
Lee, T. W., Mitchell, T. R., Wise, L., & Fireman (1996). An unfolding model of voluntary employee turnover. Academy of Management Journal,39, 5-36.
Lee, T. W., Mitchell, T. R., Holtom, B. C., McDaniel, L., & Hill, J. W. (1999). The unfolding model of voluntary turnover: a replication and extension. Academy of Management Journal, 42, 450-462.
Mitchell, T., Holtom, B., & Lee, T. (1999b). Decisions regarding organizational attachment: why people leave and why they stay. Paper presented at 5th international conference of the Decisions Sciences Institute, Athens.
Mitchell, T., Holtom, B., Lee, T., Sablynski, C., & Erez, M. (1999a). The retention of employees: the role of organizational embeddedness. Paper presented at the annual meeting of the Academy of Management, Chicago.
Price, J. L., & Mueller, C. W. (1986). Absenteeism and turnover of hospital employees. Greenwich, CT: JAI Press.