Prediction of turnover intentions among employed adults with multiple sclerosis.

The intentions of adults with multiple sclerosis (MS) to voluntarily leave positions of employment (i.e., turnover intentions) are of interest in vocational rehabilitation because they are often the end result of deficiencies in on-the-job interventions by rehabilitation professionals and employers (Gilbride & Hagner, 2005). Calling for a career perspective in rehabilitation, Rumrill and Roessler (1998) described the need for on-going follow-up services to help employed individuals with chronic illnesses identify and reduce or remove barriers to productivity. Allaire, Li, and LaValley (2003) demonstrated that such interventions enabled individuals with rheumatic diseases to maintain their employment, thereby avoiding both an interruption in their work lives and a "sick role" self-definition. Consequently, planning to leave one's job is an early warning sign that immediate rehabilitation interventions are in order if employees with chronic illnesses are to solve job-related problems and maintain their employment (D'Zurilla & Nezu, 1999).

Employed adults with MS are particularly appropriate candidates for workplace interventions targeted at job retention. The rationale for this statement is very simple. Most adults with MS have an employment history (over 90%, LaRocca, Kalb, Schienberg, & Kendall, 1985; LaRocca, 1995; Rumrill, Roessler, & Koch, 1999), and most (approximately 60%, LaRocca, 1995; Rumrill, 1996) were working at the time of their diagnosis. However, retention of their jobs is another matter, given the fact that only 20 to 40 percent continue to work following diagnosis of their MS (Beatty, Blanco, Wilbanks, Paul, & Hames, 1995; LaRocca, 1995; Rumrill et al., 1999). Thus, the purpose of this study was to identify factors predicting "turnover intention" in light of the need to improve job retention success through on-the-job follow-up services that rehabilitation professionals could provide adults with MS.

Research indicates that turnover intention, i.e., thinking about leaving one's job, is the best and most immediate predictor of turnover (Dougherty, Bludorn, & Keon, 1985; Hui, 1988; Martin, 1979; Mobley, Homer, & Hollingsworth, 1978; Steers & Mowday, 1981). Although it is true that some jobs are not worth keeping, it is also accurate to say that much time, effort, and money is invested in ensuring that vocational placements through rehabilitation services are in the best interests of the person. Losing such a position, when it might have been avoided through timely and appropriate interventions, is therefore, a tragedy of no small proportions. Hence, empirical findings regarding factors affecting turnover intention among individuals with disabilities have considerable pertinence for the development and implementation of job retention interventions in rehabilitation.

Since the turn of the century, researchers in the fields of management, human resources, and industrial/organizational psychology have invested considerable effort in developing a model to predict voluntary turnover (Moore, 1998). Models resulting from their work have ranged from simple linear ones generated through regression analyses to more complex, dynamic ones resulting from structural equation modeling (e.g., Feeley, & Barnett, 1997; Flaherty & Pappas, 2002; Mitchell, Wise, & Fireman, 1996; Quinn, Rycraft, & Schoech, 2002). The amount of variance in turnover accounted for in these studies typically ranges from 17% to 30%, with one particular model explaining 63% of the variance in turnover intentions. Predictors of turnover in the research literature include job satisfaction, organizational commitment, supervisor relations, coworker relations, compensation and benefits, internal job mobility, organizational communication, organizational support, job match, and, finally, turnover intention. Turnover intention is central to this investigation because, as previously noted, it is the best and most immediate predictor of voluntary turnover.

It is important to point out that, although many models of employee turnover have been introduced and tested, few models include the symptoms of a disabling or chronic illness condition (Devins & Shnek, 2000; Roessler, 2004) as a precipitant of turnover intentions and, therefore, voluntary turnover. Therefore, in addition to including workplace variables encompassed in current turnover models, the model used in this study also incorporated a disability and illness factor, severity of symptoms, which was used as an independent variable predicting turnover intentions.

The rationale for the study is compelling. If rehabilitation professionals can predict turnover intentions, they can predict turnover and intervene prior to job loss in situations in which the person wishes to retain employment. Increasing the probability of job retention means that more people with disabilities will lead fulfilling lives. Indeed, maintenance of employment is directly related to quality of life (QOL) reported by people with chronic illnesses such as MS (Koch, Rumrill, Roessler, & Fitzgerald, 2001; Rumrill, Roessler, & Fitzgerald, 2004). The reasons for this connection between employment and QOL ratings are obvious in western cultures (Szymanski et al., 2003). Work fulfills survival and psychological needs of individuals. Income generated through employment helps people secure the goods and services they need to exist, and participating in work offers individuals a sense of meaning and identity.

Findings from this study also contribute to our understanding of rehabilitation interventions that enable people with disabilities to maintain their employment. These interventions may range from medical to workplace strategies. The importance of services to decrease turnover among people with disabilities is stressed in discussions of the mission of vocational rehabilitation and the role of the counselor (Hershenson, 1998; Rumrill & Roessler, 1998; Salomone, 1996). Whenever rehabilitation clients become turnover statistics, they constitute a failure of the program to use its resources effectively to achieve one of its most central goals, i.e., long-term competitive employment of people with disabilities. Loss of employment only adds to the stress that individuals experience in dealing with disability (Banks, 1995), a combined negative status that has devastating effects on adults with disabilities and their families.

Focusing on employed adults with MS, this study assumed that most of the workplace correlates of turnover intentions for the general population were pertinent for this group as well. Therefore, the job satisfaction variable from Moore's (1998) model was included in the current model, and the organizational commitment variable was replaced by one labeled "employment support." Employment support in the current model can be equated to organizational support as described by Moore. Employment support was substituted for organizational commitment because it was hypothesized that most people with disabilities would be committed to their place of work. Bromoge (1999) reported that people with disabilities were less likely to be absent from work and much less likely to voluntarily leave their place of employment. Furthermore, Moore hypothesized that the level of organizational support would predict turnover intentions in her original model, but found that it did not in the population she studied. She attributed the lack of a relationship between organizational support and turnover intentions to the fact that all of the organizations in her study (integrated-circuit chip manufacturers in the Silicon Valley) were highly supportive of their employees. It is much less likely for the cross section of employers who have hired people with MS to be uniformly supportive of their employees; therefore the employment support construct was added to the current model (Roessler, Neath, McMahon, & Neath, in press).

Representing the constructs of distress and coping ability, adaptation is the final variable of the model of turnover intention used in this investigation. Evidence for the inclusion of the constructs of distress and coping can be found in Layne, Hohenshil, and Singh (2004). They found that both stress and strain were positively correlated with turnover intentions in a population of rehabilitation counselors and that coping was negatively correlated with turnover intentions. They also reported that stress level was predictive of turnover intentions in their sample. Wallace and Tighe (1994) also reported that coping ability is a predictor of turnover intentions.

Assumptions were also tested as to whether adding disease factors to the model (e.g., the severity of symptoms item) resulted in greater understanding of the phenomenon of turnover intention. Precedent for this speculation exists in that research indicates that people with MS experiencing more severe and numerous symptoms and the presence of cognitive limitations are less likely to be employed (Rao, et al., 1991; Roessler, Fitzgerald, Rumrill, & Koch, 2001; Rumrill, 1996). Gender was included in the model because studies (i.e., Reipert, 2004; Roessler, Turner, Robertson, & Rumrill, 2005; Rumrill, Tabor, Hennessey, & Minton, 2000) have reported differences in the employment experiences of men and women with MS. LaRocca, Kalb, Scheinberg, and Kendall (1985) reported that 80% of women with MS were unemployed as compared to 66% of men with MS.

The current study attempted to show that the addition of symptom characteristics in a model predicting turnover intentions of adults with MS would increase the predictive power of the model. Furthermore, the hypotheses were that one of the primary employment-related variables in Moore's (1998) model (job satisfaction) was applicable to adults with MS; that the employment support variable, equivalent to the organizational support concept from Moore's original model, would predict turnover intentions in the current model; and that the addition of the variable labeled severity of symptoms would increase the predictive power of the current model. In addition it was hypothesized that the addition of adaptability and gender variables would further increase the model's ability to predict turnover intentions.

Methods

Data for exploring a model of turnover intention were drawn from sections of a large national survey regarding the employment concerns of adults with MS (Roessler, Rumrill, & Hennessey, 2002). In this study, turnover intention was defined in terms of employed participants' responses to a survey item asking whether they expected to be working in one year.

Participants

Overall 1,310 adults with MS responded to the employment concerns survey. Only participants who were currently employed at the time of completing the questionnaire were selected for the current study, resulting in a response dataset of 548 members. Of the 548 employed participants, 19% were male and 81% were female. Ninety-seven percent of the employed respondents completed the questionnaire themselves, with 1%o of the respondents' spouses completing the survey, and another 1% of the surveys completed by parents, friends, or another family member. The average age of the employed respondent was 46 years (SD = 9.64). The age range for the employees was 23 to 74 years. Ninety-one percent of the respondents were White, whereas 6% of the employees were African-American, 2% Hispanic, 1% Native American / Alaskan Native, and only 1 of the respondents was Asian / Pacific Islander. The majority of the employees were married (67%).

Only 2% of the respondents were diagnosed with MS prior to becoming an adult, whereas 64% were diagnosed prior to age 40. Eighty-seven percent of the respondents were employed at the time they were diagnosed with MS. The average annual income of the respondents was $40,682 (SD = $48,647). Thirty-eight percent of the respondents reported working 40 hours per week with an average work week of 35.24 hours (SD = 12.86). Of all of the employed respondents, only 37% had requested accommodations in their workplace with 63% indicating that they had not requested accommodations.

Instrumentation 

Five variables were selected to predict the outcome variable of turnover intention or "Certainty of Employment" in one year from completion of the survey. The "certainty of employment" variable was worded "How certain are you that you will be working 1 year from now?" The original five-point scale was collapsed into three categories for the logistic regression analysis using the following coding: 1 = not certain at all or a less than 50% chance, 2 = approximately a 50% probability, and 3 = greater than a 50% chance or completely certain. The predictor variables included symptom severity, level of job satisfaction, current level of employment support, adaptability to stressors, and gender.

Symptom severity was measured on a 5-point scale ranging from I = no current symptoms to 5 = multiple, severe symptoms, significantly limiting daily functioning. Level of job satisfaction was measured using a 3-point classification: 1 = not satisfied, 2 = uncertain, and 3 = satisfied. Current level of employment support was a scale consisting of seven items measuring benefits, workplace accommodations, work environment, and employer attitudes. The items making up the employment support scale can be seen in Table 1. The scale range was from -1 to 1 with negative scores representing a negative perception of employment support and positive values representing a positive perception of employment support. The internal consistency of the scale was .82. Adaptability to stressors consists of ten items measuring emotional response to stressors and perception of one's ability to cope with stressors (Hewitt, Flea, & Mosher, 1992). The scale range was 1 to 5 with lower scores indicative of high perceived levels of stress and low levels of coping effectiveness and high scores representing a lower perceived levels of stress and higher levels of coping effectiveness. One of the items ("felt that things were going your way") was removed from the scale due to its low negative correlation with the remaining items in the scale. The internal consistency of the items was .91.

Data Analysis

Data were analyzed by predicting the categorical variable of turnover intention using logistic regression. Two of the five predictor variables (symptom severity and job satisfaction) were selected for first entry into the model due to the difficulty of influencing these indicators externally. Next, a forward selection model was used to enter the variables of employment support, adaptability to stressors, and gender. If the variables of employment support and adaptability to stressors are predictive, it is hypothesized that interventions might be selected or developed to affect these factors. Gender was included in the model last to determine if it accounts for a significant amount of variability in turnover intention beyond the first four predictor variables. A significant gender factor would indicate a need for considering gender issues in the development or use of interventions. Table 2 present properties of the predictor variables.

Results

Descriptive Statistics 

The majority of the employed participants believed that they would be employed one year from now (64%). Fourteen percent of the respondents believed it is unlikely that they would still be employed one year from now, and 23% of the individuals responded that they had a "50/50" chance of being employed in one year (see Table 3). Severity of symptoms for the employed respondents was not extremely high, as a whole, with 87% rating their symptom severity as 3 or less on a 5 point scale. The majority of employees with no symptoms of MS were very certain that they would still be employed in one year (80%). As severity of symptoms increased, certainty of employment decreased (see Table 3). Even though they reported multiple, severe symptoms that limit everyday life, 22% of the respondents with severe symptoms were still certain that they would be employed in one year. The majority of the respondents (70%) were satisfied with their current job. Of those respondents either uncertain or not satisfied with their job, the certainty of being employed in one year was substantially less (47% and 46%, respectively) than for satisfied employees (71%). There was no difference in certainty of employment in one year for males and females (65% and 63% were certain of employment, respectively). Adaptability to stressor scores increase as one's certainty of employment increases (see Table 4). Additionally, employment support is positively related to certainty of employment.

Logistic Regression

A logistic regression analysis was conducted using participants' certainty of employment ratings to measure turnover intention. Five predictor variables included symptom severity, job satisfaction, current employment support, adaptability to stressors, and gender. The number of participants in the sample with complete data on the six logistic regression variables was 388. Symptom severity and job satisfaction were placed into the model first. Current employment support, adaptability to stress, and gender were entered using forward regression, with inclusion requiring a significant relationship at alpha = .05. Current employment support and adaptability to stress were retained in the model along with symptom severity and job satisfaction (see Table 5). Gender was not a significant predictor of employment certainty for adults with MS ([X.sup.2](1) = 0.53,p = .47).

The proportion of variability in employment certainty predicted using symptom severity, job satisfaction, employment support, and adaptability to stress was 0.26. The percent concordant in classification was 74.6, the percent discordant was 25.0, with the percent of ties being 0.3. The odds ratio for symptom severity was 0.52, indicating that a one point increase on the 5-point severity scale resulted in someone being one half as likely to be in a higher employment certainty category (see Table 5). A one point increase on the 3-point job satisfaction scale increased one's odds by 1.63 of being in a higher employment category. A one-point increase on the employment support and adaptability to stress scales resulted in participants having increased odds of 1.81 and 1.79, respectively, of being in a higher employment certainty category.

Of the three predictor variables other than MS symptom severity (job satisfaction, employment support, and adaptability to stress), the variable with the largest potential impact on certainty of employment in one year is adaptability to stress. The odds ratio is similar in value to job satisfaction and employment support; however the range of values for the adaptability scale is 5 points compared to the 3 point scales of job satisfaction and employment support. A comparison of the probabilities of being classified in low certainty of employment as compared to high certainty of employment categories for adults with different levels of MS symptom severity are provided in Table 5. Results indicate that adults reporting no MS symptoms have a high certainty of being employed in one year if they have moderate to high levels of job satisfaction, employment support, and adaptability to stress. If adults have low levels of job satisfaction, low levels of perceived employment support, and low levels of adaptability to stress, they have a low certainty of being employed in one year regardless of MS symptom severity.

As MS symptom severity increases, certainty of employment decreases significantly. An adult with moderate levels of job satisfaction, employment support, and adaptability to stress with no MS symptoms has a 79% probability of being classified as high certainty of employment in one year; however, an individual with similar predictor variable values with moderate MS symptoms has a 49% probability of being classified as having high certainty of employment. When an adult's symptoms become severe, his or her probability of being classified as having a high certainty of employment in one year is reduced to 20% when job satisfaction, employment support, and adaptability to stress is moderate.

Of interest is whether there are factors that can increase the likelihood that one believes he/she will be employed in one year, even when their MS symptoms are becoming increasingly debilitating. Changes in the three variables of job satisfaction, employment support, and adaptability to stress are compared on their predicted outcomes. With moderate levels on the three predictor variables other than symptom severity, the probability of being in the high certainty of employment category is .20. If adults with MS are satisfied with their jobs, their probability of having high certainty of employment is .30 (see Table 6). Increasing employment support from moderate to high resulted in a .32 probability of high certainty of employment. However, increasing adaptability of stress from moderate to high resulted in an increase to .45 of being in the high certainty of employment category.

In order for adults with severe MS symptoms to increase their probability of being in the high certainty of employment category, they need to be at the higher levels of two of the three predictor variables, with one of the predictor variables being adaptability to stress. If the adult with severe MS symptoms is only moderate in terms of adaptability to stress, the highest probability of being in the high certainty of employment category is .43 (see Table 6). Ideally, if the adult with severe MS symptoms is satisfied with his or her job, has high employment support, and has high levels of adaptability to stress, then the probability of being in the high certainty of employment category is .71, which is substantially higher than adults with no MS symptoms who are not satisfied with their jobs and have low employment support and low adaptability to stress. It is interesting that adults with moderately severe MS symptoms can list themselves as being completely certain of employment in one year if they have high perceptions of employment support and high adaptability to stress, even if their job satisfaction levels are not high (satisfied with job = .82, undecided job satisfaction = .74, unsatisfied with job = .64; see Table 7).

Discussion

Results of the study provide some good news in that nearly two-thirds (64%) of the employed respondents with MS were reasonably certain that they would be working in one year. Optimism regarding continuing to work was even more apparent among those with no current MS symptoms; 80% of those reporting no current symptoms expected to be working in one year. These results underscore the commitment to the work role among adults with MS, a finding consistent with National Organization on Disability survey (2000) that indicated that people with disabilities want to work. The fact that the expectation of employment retention was particularly strong among those with no current symptoms is consistent with Devins and Shnek's (2000) findings regarding the relationship between health status or disability intrusiveness and participation in employment.

Like most Americans (Fritzsche & Parrish, 2005), the majority of the respondents (70%) were satisfied with their current jobs. Consequently, adults with MS would respond positively to on-the-job rehabilitation interventions that would help them maintain valued employment such as identification and provision of job accommodations. Recent research, unfortunately, suggests that accommodation services on the job are lacking. Failure to receive reasonable accommodations was one of the more frequent types of allegations of discrimination filed with the EEOC by adults with MS (Roessler et al., in press).

Results underscore the commitment of this group of adults with MS to employment in general and, even more so, to jobs that they consider satisfying. Their desire to maintain satisfying work provides a strong rationale for the employment protections available through Title I of the Americans with Disabilities Act (ADA) which Blanck (2004) referred to as the "most heavily litigated provision of the ADA" (p. 5). Drafted to counter the bias in society against people with disabilities (Batavia & Schriner, 2000; Miller, 1997), the ADA advances multiple strategies for helping people with disabilities retain employment such as reasonable accommodation, new or modified equipment, job restructuring, or provision of a reader or interpreter. The ADA does not of course ensure that employers will fully subscribe to these practices as evidenced in a recent study by Wooten and James (2005). Obviously rehabilitation counselors must continue to be vigilant regarding the existence of discriminatory treatment in hiring and retention and the need to help business and industry resolve those situations in an appropriate manner.

Interestingly, although incidence of MS in the general population is higher for women (2 to 1; Rumrill & Hennessey, 2001), no differences appeared in comparisons of men and women with MS regarding turnover intention. About two-thirds of both groups expected to be working one year from the time of the survey. This similar outlook regarding job retention is somewhat inconsistent with reality. Research (Beatty et al., 1995; Kornblith, LaRocca, & Baum, 1986) indicates that women with MS are more likely to be unemployed than men with MS. Moreover women with MS report different experiences in the workplace than do men. Specifically, women with MS are more likely to allege discrimination in regard to harassment and intimidation than in regard to discharge. In comparisons with women with MS, men with MS are more likely to allege discrimination regarding discharge than harassment and intimidation (Roessler et al., in press). Rehabilitation counselors may need therefore to involve women in anticipatory coping regarding the types of conditions (e.g., harassment and intimidation) that will negatively affect their employment. For example, they may need assistance in identifying the cues associated with these types of treatment and the appropriate responses when they do occur.

Results of the logistic regression model provide strong support for important principles regarding employment expectations and turnover intentions for individuals with MS. As common sense would dictate, severity of illness does influence one's employment outlook as was clearly indicated in this study. For adults with MS responding to the survey, severity of their MS symptoms was one of the major predictors of turnover intention. Devins and Shnek (2000) and others (Roessler, Rumrill, & Fitzgerald, 2004; Saunders, Leahy, McGlynn, & Estrada-Hernandez, 2006) have documented the relationship between symptom severity and employment expectations and status. Consequently, rehabilitation counselors must help people with MS stay abreast of advances in the medical treatment of MS and in assistive technology and job modifications to improve their abilities to manage their symptoms and avoid their exacerbation at work.

In reviews of his research findings, Devins and others (Devins & Shnek, 2000; Roessler, 2004) have stressed that severity of health conditions interferes with participation or expected participation in adult roles in two ways. First the functional limitations associated with the condition make it more difficult for the person to perform certain essential functions of the position. Hershenson's theory of career development stresses that the effects of mid-career disabilities and chronic illnesses such as MS are first experienced in terms of decrements in worker capacities (Szymanski, Enright, Hershenson, & Ettinger, 2003) which then radiate out to influence work goals and work personality. Therefore findings underscore the importance of the rehabilitation counselor as a source of information regarding a) the medical aspects and treatment of MS, b) the relationship between reasonable accommodation strategies described in the ADA and functional limitations associated with MS, and c) preferred strategies for requesting that one's supervisor or employer review the worker's on-the-job accommodation needs.

According to Devins & Shnek (2000), severity of disability also affects involvement in work by virtue of schedule interruptions associated with treatment of the condition such as frequent appointments with the doctor and side effects of medication. In the case of MS, powerful immunosuppressant drugs are an important part of the treatment which may leave individuals more susceptible to infections of various types. As a result employees with MS may need to carefully monitor their health and seek treatment at the first signs of illness. Again rehabilitation counselors can make their clients with MS aware of this situation and assist them, at least initially, in the management of personal health.

The second principle finding support in this study is the contribution of stress and its management to turnover intention. The level of stress and the person's self-perceived ability to manage stress was the second major predictor of turnover intention. Theory and research combine to make a case for enhancing the adaptability of employees with disabilities. More than 20 years ago, Crites (1982) criticized job/person match or trait/factor career theories such as the Minnesota Theory of Work Adjustment as being too "static" in their conceptions of work adjustment. Workers continually encounter barriers to career motives (e.g., discrimination) on the job that require responses on their part that reduce or remove both the barriers and the tension associated with those barriers. He referred to this type of reaction as an integrative response as opposed to "nonadjustive" responses which affect neither the barrier nor the tension (e.g., distraction or withdrawal) and "adjustive" responses which affect the tension level hut not the barrier (e.g., substance abuse). Crites advocated for support groups led by counselors that would enable employees to identify areas of dissatisfaction at work and to develop integrative solutions to those problems.

Stress management in the form of enhanced on-the-job problem solving skills of workers has received renewed attention in the literature (Roessler, 2002). In recent revisions of job/person theories such as the Person-Environment Congruence theory described by Dawis (2005), on-the-job coping by workers is examined in terms of work adjustment style which includes both the employee's tolerance for discorrespondence or stress and the worker's preference for either active or reactive responding styles to such stress. Rehabilitation counselors can help employed adults with MS improve their problem-solving skills so that they might take a more active and assertive approach to job stressors.

In their description of problem-solving therapy, D'Zurilla and Nezu (1999) called for preventive problem-solving training (PST) programs in business and industry for both employees and managers. They concluded their discussion of the need for training in problem-solving by saying, "Hence, it is possible that preventive PST programs might enable managers and other employees to more effectively resolve the daily work problems that lead to poor job performance, high job stress, absenteeism, accidents, and burnout. If successful, such programs could have major economic and health benefits for society" (p. 217). Rehabilitation counselors would be effective leaders of problem-solving groups in the work setting that would have the potential to improve the job satisfaction and retention of adults with MS.

At the same time, one should comment on the fact that the full model included two additional variables familiar to researchers in rehabilitation and industrial psychology--job satisfaction and employer support. These findings confirm Moore's (1998) research on factors predicting turnover intention, strongly implying that rehabilitation professionals should assist both workers and employers in identifying and addressing factors contributing to job satisfaction and employer support. An acknowledged predictor of the intention to seek a new job, level of job satisfaction is affected both by conditions in the workplace and by characteristics of the employee (Fritzsche & Parrish, 2005). Therefore counselors must use a multifaceted strategy when attempting to influence worker job satisfaction in a positive way. Specifically changes in the environment are called for requiring identification and remediation of conditions resulting in dissatisfaction. Research on job satisfaction suggests that work-related aspects such as nature of job tasks and job reinforcers are points of intervention (Fritzsche & Parrish, 2005; Dawis, 2005).

Counselors may also attempt to positively influence job satisfaction by addressing worker perceptions as well. Research indicates that job satisfaction is often a function of the personal dispositions of workers that influence their expectations and perceptions of the workplace. In such cases counselors may need to counsel with employees to help them better understand the impact of their personality traits on their perceptions of employment. For example, individuals with strong Type A personalities may have unrealistic expectations of supervisors and co-workers and become impatient with the perceived failure of others to respond as expected (Fritzsche & Parrish, 2005). Other individuals may have little experience with the world of work (Szymanski et al., 2003) and therefore have unrealistic ideas as to what is expected of them in terms of attendance, punctuality, and productivity.

The final predictor in the model was employer support which reinforces research findings (Dyck & Jongbloed, 2000; Moore, 1998) regarding factors affecting turnover intention. In this study employer support was defined in terms of items reflecting the respondent's views of the responsiveness of his or her employer to the needs of workers with MS specifically and in comparison with non-disabled workers. This finding provides documentation for the recommendation that counselors develop educational interventions that enable employers to increase their support of their employees with disabilities, e.g., programs addressing disability awareness, reasonable accommodation, and provisions and protections in Title I of the Americans with Disabilities Act (Younes, 2001). A rationale for such intervention at the worksite is extremely easy to formulate at this point in history. Human resource professionals have identified worker retention as the number one priority for business and industry because it is one of the most cost effective ways for them to respond to a severe upcoming labor shortage (Schramm, 2006).

Conclusion

Results of this study confirm a number of principles regarding job retention from the literature in business, psychology, and rehabilitation. Counselors can combine these principles of job retention in their approaches to intervening at the workplace on behalf of employees with disabilities and chronic illnesses such as MS. The first principle is that severity of health condition matters; strategies are needed to help employees deal with health-related factors such as functional limitations and medication side effects. The second principle is that employee adaptability is important; employees who experience higher levels of stress on the job and who feel less able to cope with that stress are more likely to consider leaving their employment. The next principle is that job satisfaction is a good indicator of job search activity and therefore counselors must help employees identify and ameliorate both work-related tasks and activities and personal dispositions that are adversely affecting their job satisfaction. The final principle is that the support of the employer is critical in helping a person with a disability retain employment. Although each of the above principles is a simple matter of common sense, it is important to once again document their validity through empirical research. It is also more important than ever for rehabilitation personnel to intervene with workers with disabilities and their employers and supervisors, given the importance of work to people with disabilities, the impact of a projected labor shortage on national productivity, and cost effectiveness of retaining current employees.

Author's Note

This research was supported by a Healthcare Delivery and Policy Research Contract from the National Multiple Sclerosis Society.

References

Allaire, S., Li, W., & LaValley, M. (2003). Work barriers experienced and job accommodations used by persons with arthritis and other rheumatic diseases. Rehabilitation Counseling Bulletin, 46(3), 147-156.

Banks, M. (1995). Psychological effects of prolonged unemployment: Relevance to models of work re-entry following injury. Journal of Occupational Rehabilitation, 5(1), 37-53.

Batavia, A., & Schriner, K. (2001). The Americans with Disabilities Act as engine of social change: Models of disability and the potential of a civil rights approach. Policy Studies Journal, 29, 690-702.

Beatty, W., Blanco, C., Wilbanks, S., Paul, R., & Hames, K. (1995). Demographic, clinical, and cognitive characteristics of multiple sclerosis patients who continue to work. Journal of Neurological Rehabilitation, 9(3), 167-193.

Blanck, P. (2004). Stories about Americans with disabilities and their civil rights. The Journal of Gender, Race, and Justice, 8, 1-30.

Bromoge, N. (1999). Employing the disabled, it could be you. Management Accounting: Magazine .for Chartered Management Accountants, 77(8), 69.

Crites, J. (1982). Measurement of career development. In B. Bolton & R. Roessler (Eds.), Proceedings of the symposium on applied research methodology (pp. 1-8). Fayetteville, AR: Research and Training Center.

Dawis, R. (2005). The Minnesota theory of work adjustment. In S. Brown and R. Lent (Eds), Career development and counseling (pp. 3-23). New York: Wiley.

Devins, G., &Shnek, Z. (2000). Multiple sclerosis. In R. Frank & T. Elliott (Eds.), Handbook of rehabilitation psychology (pp.163-184). Washington, DC: American Psychological Association.

Dougherty, T., Bluedorn, A., & Keon, T. (1985). Precursors of employee turnover: A multi-sample causal analysis. Journal of Occupational Behavior, 6, 259-271.

Dyck, I., & Jongbloed, L. (2000). Women with Multiple Sclerosis and employment issues: A focus on social and institutional environments. Canadian Journal of Occupational Therapy, 67(5), 337-346.

D'Zurilla, T., & Nezu, A. (1999). Problem-solving therapy. New York: Springer.

Feeley, T. H., & Barnett, G. A. (1997). Predicting employee turnover from communication networks. Human Communication Research, 23(3), 370-387.

Flaherty, K. E., & Pappas, J. M. (2002). Using career stage theory to predict turnover intentions among salespeople. Journal of Marketing Theory and Practice, 10(3), 48-58.

Fritzsche, B., & Parrish, T. (2005). Theories and research on job satisfaction. In S. Brown and R. Lent (Eds), Career development and counseling (pp. 180-202). New York: Wiley.

Gilbride, D., & Hagner, D. (2005). People with disabilities in the workplace. In R. Parker, E. Szymanski, & J. Patterson (Eds.), Rehabilitation counseling (pp. 281-306). Austin, TX: PROED.

Hershenson, D. (1998). Systematic, ecological model for rehabilitation counseling. Rehabilitation Counseling Bulletin, 42(1), 40-51.

Hewitt, P., Flett, G., & Mosher, S. (1992). The perceived stress scale: Factor structure and relation to depression symptoms in a psychiatric sample. Journal of Psychopathology and Behavioral Assessment, 14(3), 247-257.

Hui, C. (1988). Impacts of objective and subjective labor market conditions on employee turnover. Journal of Occupational Psychology, 66, 211-219.

Koch, L., Rumrill, P., Roessler, R., & Fitzgerald, S. (2001). Illness and demographic correlates of quality of life among people with multiple sclerosis. Rehabilitation Psychology, 46(2), 154-164.

Kornblith, A., LaRocca, N., & Baum, H. (1986). Employment in individuals with multiple sclerosis. International Journal of Rehabilitation, 9, 155-163.

LaRocca, N. (1995). Employment and multiple sclerosis. New York: National Multiple Sclerosis Society.

LaRocca, N., Kalb, R., Scheinberg, L., & Kendall, P. (1985). Factors associated with unemployment of patients with multiple sclerosis. Journal of Chronic Disabilities, 38, 203-210.

Layne, C., Hohenshil, T., & Singh, K., (2004). The relationship of occupational stress, psychological strain, and coping resources to the turnover intentions of rehabilitation counselors. Rehabilitation Counseling Bulletin, 48(1), 19-28.

Martin, T. (1979). A contextual model of employee turnover intentions. Academy of Management Journal. 22, 313-324.

Miller, S. (1997). Keeping the promise: The ADA and employment discrimination on the basis of psychiatric disability. California Law Reviews; 85, 701-749.

Mitchell, T., Wise, L., & Fireman, S. (1996). An unfolding model of voluntary employee turnover. Academy of Management Journal, 39(1), 5-37.

Mobley, W., Homer, S., & Hollingsworth, A. (1978). An evaluation of precursors of hospital employee turnover. Journal of Applied Psychology, 63, 408-413.

Moore, C. W. (1998). Understanding voluntary employee turnover within the new workplace paradigm: A test of an integrated model. Unpublished Doctoral Dissertation, Claremont Graduate University.

National Organization on Disability. (2000). N.O.D./Harris Survey. Washington, DC: Author.

Quinn, A., Rycraft, J., & Schoech, D. (2002). Building a model to predict caseworker and supervisor turnover using a neural network and logistic regression. Journal of Technology in Human Services, 19(4), 65-85.

Rao, S., Leo, G., Ellington, L., Nauertz, T., Bernardin, L., & Unverzagt, F. (1991). Cognitive dysfunction in multiple sclerosis. Neurology, 41, 692-696.

Reipert, B., (2004). Multiple sclerosis: A short review of the disease and its differences between men and women. Journal of Men's Health & Gender, 1(4), 334-340.

Roessler, R. (2002). Improving job tenure outcomes for people with disabilities: The "3M" Model. Rehabilitation Counseling Bulletin, 45, 207-212.

Roessler, R. (2004). The illness intrusiveness model: Rehabilitation implications. Journal of Applied Rehabilitation Counseling, 35(3), 22-27.

Roessler, R., Fitzgerald, S., Rumrill, P., & Koch, L. (2001). Determinants of employment status among people with multiple sclerosis. Rehabilitation Counseling Bulletin, 45, 31-39.

Roessler, R., Neath, J., McMahon, B., & Rumrill, P. (in press). The relationship of selected supply and demand side factors to forms of perceived discrimination among adults with multiple sclerosis. Rehabilitation Counseling Bulletin.

Roessler, R., Rumrill, P., & Fitzgerald, S. (2004). Predictors of employment status for people with multiple sclerosis. Rehabilitation Counseling Bulletin, 47(2), 97-103.

Roessler, R. T., Rumrill, P. D., & Hennessey, M. L. (2002). Employment concerns of people with multiple sclerosis: Building a national employment agenda (Report to the National Multiple Sclerosis Society, New York, NY): University of Arkansas and Kent State University, Center for Disability Studies.

Roessler, R., Turner, R., Robertson, J., & Rumrill P. (2005). Gender and perceived illness severity: Differential indicators of employment concerns for adults with multiple sclerosis? Rehabilitation Counseling Bulletin, 48(2), 66-74.

Rumrill, P. (1996). Factors associated with unemployment among persons with multiple sclerosis. Work, 6, 177-184.

Rumrill, P., & Hennessey, P. (2001). Multiple Sclerosis: A guide for rehabilitation and health care professionals. Springfield, IL: Charles C Thomas.

Rumrill, P., & Roessler, R. (1998). Reducing workplace barriers to enhance job satisfaction: An important post-employment service for employees with chronic illnesses. Journal of Vocational Rehabilitation, 10(3), 219-229.

Rumrill, P., & Roessler, R. (1999). New directions in rehabilitation counseling: A career development perspective on closure. Journal of Rehabilitation, 65(1), 26-30.

Rumrill, P., Roessler, R., & Fitzgerald, S. (2004). Predictors of employment status for people with multiple sclerosis. Rehabilitation Counseling Bulletin, 47, 96-103.

Rumrill, P., Roessler, R., & Koch, L. (1999). Surveying the employment concerns of people with multiple sclerosis: A participatory action research approach. Journal of Vocational Rehabilitation, 12, 75-82.

Rumrill, P. D., Tabor, T. L., Hennessey, M. L., & Minton, D. L., (2000). Issues in employment and career development for people with multiple sclerosis: Meeting the needs of an emerging vocational rehabilitation clientele. Journal of Vocational Rehabilitation, 14, 109-117.

Salomone, P. (1996). Career counseling and job placement: Theory and practice. In E. Szymanski, & R. Parker (Eds.), Work and disability (pp. 365-420). Austin, TX: PRO-ED.

Saunders, J., Leahy, M., McGlynn, C., & Estrada-Hernandez, N. (2006). Predictors of employment outcomes for persons with disabilities: An integrative review of potential evidence-based factors. Journal of Applied Rehabilitation Counseling, 3 7(2), 3-20.

Schramm, J. (2006). SHRM workplace forecast. Alexandria, VA: Society for Human Resource Management.

Steers, R. M., & Mowday, R. T. (1981). Employee turnover and post-decision accommodation processes. In L. Cummings & B. Staw (Eds.), Research in Organizational Behavior (pp. 235-281). Greenwich, CT: JAI Press.

Szymanski, E., Enright, M., Hershenson, D., & Ettinger, J. (2003). Career development theories, constructs, and research: Implications for people with disabilities. In E. Szymanski & R. Parker (Eds.), Work and disability (pp. 91-154). Austin, TX: PRO-ED.

Szymanski, E., Parker, R., Ryan, C., Merz, M., Trevino-Espinoza, B., Johnston-Rodriguez, S. (2003). Work and disability: Basic constructs. In E. Szymanski & R. Parker (Eds.), Work and disability (pp. 1-26). Austin, TX: PRO-ED.

Wallace, J., & Tighe, R., (1994). What are the most important factors affecting levels of job stress? Rehabilitation Review, 5(8).

Wooten, L., & James, E. (2005). Challenges of organizational learning: Perpetuation of discrimination against employees with disabilities. Behavioral Sciences and the Law, 23, 123-141.

Younes, N. (2001). Getting corporations ready to recruit workers with disabilities. Journal of Vocational Rehabilitation, 16, 89-91.

Tresvil G. Pack

University of Arkansas

Richard. T. Roessler

University of Arkansas

Ronna Turner

University of Arkansas

Judith Robertson

University of Arkansas

Tresvil Pack, Sassin & Associates, Inc., 7146 S. Braden Ave., Suite 600, Tulsa, OK 74136.

Email: [e-mail]

Table 1
Items in the Employment Support Scale: People with MS ...

Receive the same pay as would a nondisabled person.
Can get help in identifying and designing workplace accommodations.
Are considered for other jobs in the same company if their disabilities
prevent them from going back to their own jobs.
Receive reasonable accommodations in the workplace.
Are given support from employers and supervisors after returning to
work.
Have assistance in coping with stress on the job.
Can work with employers and supervisors who understand the effects of
MS.

Table 2
Properties of Predictor Variables for the Certainty of Employment
Outcome

Variable name                 N      M      SD    Min.   Max.

Symptom Severity             388    2.50   0.97     1     5
Job Satisfaction             388    2.57   0.71     1     3
Current Employment Support   388   -0.18   0.67    -1     1
Adaptability to Stressors    388    2.55   0.71    -1     5

                             Coefficient    Number
Variable name                   alpha      of Items

Symptom Severity                 --            1
Job Satisfaction                 --            1
Current Employment Support      0.82           7
Adaptability to Stressors       0.91          10

Table 3
Certainty of Employment, Severity of Symptoms, Job Satisfaction, and
Gender Frequency Distributions

                      Not Certain of   50% Certainty   Certain of
                        Employment     of Employment   Employment

                        f    (%)         f    (%)       f    (%)

Total                   53   (14)        88   (23)     247   (64)
Symptom Severity
  No Symptoms            0    (0)         9   (20)      37   (80)
  Low Symptoms          17   (10)        23   (13)     131   (77)
  Moderate Symptoms     22   (18)        33   (27)      66   (55)
  Strong Symptoms        8   (25)        15   (47)       9   (28)
  Multiple, Severe       6   (33)         8   (44)       4   (22)
    Symptoms *
Job Satisfaction
  Not Satisfied         15   (31)        11   (22)      23   (47)
  Uncertain             11   (16)        26   (38)      31   (46)
  Satisfied             27   (10)        51   (19)     193   (71)
Gender
  Male                  10   (13)        16   (21)      49   (65)
  Female                43   (14)        72   (23)     198   (63)

* Significantly Limiting Daily Functioning

Table 4
Certainty of Employment, Adaptability to Stress, and Employment
Support Distributions

                     Not Certain of         50% Certainty
                      Employment *          of Employment

                   N      M      (SD)     N      M      (SD)
Adaptability to   106    3.15   (0.82)   176    3.18   (0.65)
  Stressors
Employment        106   -0.51   (0.61)   176   -0.32   (0.63)
  Support

                       Certain of
                       Employment

                   N      M      (SD)
Adaptability to   494    3.61   (0.66)
  Stressors
Employment        494   -0.05   (0.67)
  Support

* Certainty of Employment in one year

Table 5
Logistic Regression Model for Prediction Certainty of Employment in
One Year

Variable                       B      SE    Wald [chi square]

MS Symptom Severity          -0.65   0.12        30.07 *
Job Satisfaction              0.49   0.15        10.24 *
Current Employment Support    0.59   0.18        11.02 *
Adaptability                  0.58   0.17        11.60 *

Variable                       B     OR     95% CI (OR)

MS Symptom Severity           0.35   0.52   0.41, 0.66
Job Satisfaction             -0.19   1.63   1.21, 2.20
Current Employment Support   -0.22   1.81   1.27, 2.56
Adaptability                 -0.23   1.79   1.28, 2.50

* p < .05

Table 6
Predicted Probabilities of Employment Certainty for Levels of Physical,
Psycho-social, and Work-related Variables

Symptom           Job        Employment   Adaptability
Severity      Satisfaction    Support      to Stress

No Symptoms   Unsatisfied    Low          Low
No Symptoms   Undecided      Moderate     Moderate
No Symptoms   Satisfied      High         High
Moderate      Unsatisfied    Low          Low
Moderate      Undecided      Moderate     Moderate
Moderate      Satisfied      Moderate     Moderate
Moderate      Undecided      High         Moderate
Moderate      Undecided      Moderate     High
Moderate      Undecided      High         High
Moderate      Satisfied      High         High
Severe        Unsatisfied    Low          Low
Severe        Undecided      Moderate     Moderate
Severe        Satisfied      Moderate     Moderate
Severe        Undecided      High         Moderate
Severe        Undecided      Moderate     High
Severe        Satisfied      High         Moderate
Severe        Satisfied      Moderate     High
Severe        Undecided      High         High
Severe        Satisfied      High         High

              Certainty of Employment
                Outcome in One Year

Symptom       Not Certain   Completely
Severity        At All       Certain

No Symptoms       .36          .27
No Symptoms       .06          .78
No Symptoms       .01          .97
Moderate          .67          .09
Moderate          .18          .49
Moderate          .12          .61
Moderate          .11          .63
Moderate          .06          .75
Moderate          .04          .85
Moderate          .02          .90
Severe            .88          .03
Severe            .44          .20
Severe            .33          .30
Severe            .31          .32
Severe            .20          .45
Severe            .21          .43
Severe            .13          .57
Severe            .12          .60
Severe            .08          .71

* symptom severity 1 to 5; job satisfaction 1 to 3; employment support
-1 to 1; adaptability 1 to 5

Table 7
Predicted Probabilities of Employment Certainty for Levels of Physical,
Psycho-social, and Work-related Variables

Symptom           Job        Employment   Adaptability
Severity      Satisfaction    Support      to Stress

Mod. Severe   Unsatisfied    Moderate     Moderate
Mod. Severe   Unsatisfied    Moderate     High
Mod. Severe   Unsatisfied    High         Moderate
Mod. Severe   Unsatisfied    High         High
Mod. Severe   Undecided      High         Moderate
Mod. Severe   Undecided      Moderate     High
Mod. Severe   Undecided      High         High
Mod. Severe   Satisfied      High         High

              Certainty of Employment
Symptom         Outcome in One Year
Severity
              Not Certain   Completely
                At All       Certain

Mod. Severe       .40          .23
Mod. Severe       .18          .49
Mod. Severe       .27          .35
Mod. Severe       .11          .64
Mod. Severe       .19          .47
Mod. Severe       .12          .61
Mod. Severe       .07          .74
Mod. Severe       .04          .82

* symptom severity 1 to 5; job satisfaction 1 to 3; employment support
-1 to 1; adaptability 1 to 5