Work values and expectancies in occupational rehabilitation: the role of cognitive variables in the return-to-work process.

It is a commonplace to note that the amount of time to return to employment, following an injury or other compensable condition incurred through employment, cannot be fully predicted from knowledge of the type and extent of the injury or condition. Empirical research (see reviews by Hood & Downs 1985; Hester, Decelles & Keepper 1989; Tate 1992) has identified a range of demographic, individual, social and occupational factors related to vocational outcome. However simply listing such variables tells us little about the process of return to work, or how, for example, personal characteristics might interact with organizational variables such as type of rehabilitation service or employer behavior, or even how characteristics of the person such as age or sex, affect behaviors or decisions of key parties that in turn lead to certain employment status outcomes.

Recently, concepts drawn from cognitive theories of motivation have been suggested as appropriate constructs for understanding individual differences in return-to-work motivation and associated behaviors following work injury. Roessler (1989) has proposed that instrumental theory of motivation as described by Vroom, Landy and others (Landy, 1985), could provide an appropriate conceptual framework for both understanding the processes involved in return to employment following occupational injury and for guiding rehabilitation effort. In similar vein, Mitchell, Brodwin and Benoit (1990) argued that Bandura's (1982, 1989) concepts of self-efficacy and outcome expectations have wide application in explaining return-to-work behavior.

The purpose of this paper is twofold: (a) to provide an analysis of the potential utility of such cognitive concepts in explaining the observed variance in return to work rates among individuals with a compensable work injury and (b) report on an initial empirical investigation of the hypothesis that measures of work valence and return-to-work outcome within a sample of workers' compensation clients.

Cognitive motivation theory

Cognitive theories of motivation locate the genesis of motivational factors in the individual's anticipation of future events (based on experience with past and current interactions with the environment) and emphasize the individual's beliefs or subjective interpretation of the relationship between actions and environmental consequences as important determinants of purposeful behavior. As such, they provide a potentially useful set of concepts for understanding the various factors influencing those actions related to an individual's return to employment following the suffering of a work injury. Vroom's (1964) expectancy-valence theory of motivation was originally developed in the context of understanding vocational choice and general vocational behavior. In his model, Vroom restricted the concept of motivation to voluntary actions and conceptualized motivation as the explanatory construct for understanding the choices a person makes among alternate course of action. These choices are hypothesized to depend on the relative strength of the force (motivation) associated with each possible action. In turn motivation to engage in a particular action is hypothesized to be a function of the product of the valences of outcomes associated with the action and the expectancy that given outcomes will result from given actions.

By 'valence' Vroom was referring to the anticipated satisfaction that would be obtained from an outcome, and noted that at least two components were involved in establishing valency: the intrinsic 'value' of the outcome to the individual and the 'cognised instrumentality' of the outcome. By 'cognised instrumentality' or simply, 'instrumentality,' Vroom was referring to the extent to which the individual believes that the given outcome will lead to further desired (or undesired) outcomes, i.e. whether achieving one outcome (such as obtaining a promotion) would be instrumental in achieving other outcomes of positive or negative value (such as increased social status, greater work pressure etc.). In this theory, expectancy is defined as the person's subjective estimate of his or her belief concerning the likelihood that a particular action will achieve the desired outcome. Hence in valence-expectancy terms, individuals demonstrate motivated behavior when they anticipate that they have a reasonable chance of achieving outcomes with positive valences.

Clearly this is a cognitive theory of motivation and one that gives a central role to the individual's subjective understandings of the disadvantages and benefits of alternative outcomes which may result from their actions and of the relationship between their own actions and given outcomes. Within this framework, actions indicative of an intention to resume work or not to resume work are determined by perceptions about likely consequences of these actions. Moreover, environmental events influence behavior to the extent that they modify beliefs or the meaning of the situation to the individual.

A number of authors (e.g. McDaniel, 1975; Lawler, 1983) have developed and extended expectancy-valence theory since Vroom's original formulation. Lawler (1983) made the point that outcomes could in fact be actions or particular levels of performance and drew the useful distinction between expectancies that people hold about the likelihood that actions or effort on their part will lead to the intended performance (effort-performance expectancy) and the expectancies that people hold about the likelihood that given outcomes will follow their achievement of a certain level of behavior or performance (performance-outcome expectancy). Expectancy theory as developed by Lawler is essentially the same as Bandura's Social Cognitive Theory (Bandura 1982; 1989) with respect to the role given to expectations. According to Bandura's model, when applied to the paradigm of the individual engaging in a behavior that will have consequent outcome, motivation will not only be influenced by outcome expectations (i.e. beliefs about the outcomes that will result from one's actions) but most importantly such motivation will be influenced by expectations about one's ability to engage in, or execute, the behavior, i.e. perceived self-efficacy. In general, Bandura argued that self-efficacy expectations are the more potent influences on motivation and that outcome expectations contribute to motivation independently of self-efficacy beliefs when other factors besides quality of performance affect outcomes. Given the multitude of factors besides the individual's own efforts that can influence an outcome like return to employment, Bandura's position would suggest that outcome expectations are likely to be important determinants of motivation in this context.

Return to work and cognitive motivation concepts

Cognitive expectancy theories can be seen to provide a framework for investigating the role of prominent work-related variables in the process of return to work following work injury. Within this framework, an individual's actions in remaining off work or approaching work could be influenced by the person's perception of both positive and negative outcomes associated with return to work and those associated with competing alternatives. The extent to which return to employment has instrumental value for the individual is related to the individual's beliefs that work will lead to other preferred outcomes (such as status), and to the possible non-desired outcomes that the individual believes will result from attaining a return to work (such as unpleasant social interaction). In turn, these beliefs are likely to be influenced by the extent of past work satisfaction and anticipation of future work situations. Collectively these outcome-outcome expectations or instrumentalities contribute to the valence term in the valence-expectancy equation. Similarly, the individual's beliefs about the availability of work and his or her capacity to attain work fitness would appear to be relevant to the expectancy component.

Although beliefs about the likely outcomes of return to work are likely to be based on objective aspects of the individual's situation, expectancy theory emphasizes the subjective nature of these beliefs. Thus the person's subjective estimate of, for example, the likelihood of re-injury if back at work is the factor related to motivation in this model, not necessarily the objective probability of this event. Experiences such as interactions with key medical advisors, conversations with similarly injured employees, communications (or lack of) from representatives of the employer etc. can reinforce or challenge some of the above values and beliefs and thus play a central role in influencing return-to-work motivation.

In summary expectancy theories of motivation view the individual as a 'conscious agent' who actively interprets environmental demands and options in terms of perceived needs and acts in accordance with these perceptions. This model of human functioning is highly compatible with rehabilitation philosophy (see e.g. Wright 1980) which similarly conceptualizes individuals as self-determining agents whose actions are meaningful in terms of needs and evaluations of the social and physical environment.

Although valence-expectancy theory would appear to be consistent with rehabilitation philosophy and a useful way of conceptualizing the issues involved in return to work and guiding return-to-work interventions, little empirical work that utilizes concepts from this theoretical model has been reported in the rehabilitation literature, even though Locke and Henne (1986), in a review of work motivation theories, rated it of moderate to high validity with respect to accounting for or predicting job behavior.

Sandstrom & Esbjornsson (1986) in a prospective study of patients with chronic back pain reported that the patients own predictions of vocational outcome at initial assessment were significant predictors of actual vocational outcome one year, and four years, later. Catchlove & Cohen (1982) conducted a two-group study to confirm a hypothesis that incorporating a positive expectation for work return into a multi-modal treatment program for worker's compensation clients with chronic pain problems would result in greater numbers returning to work. However these are isolated findings, and systematic investigation of the role of valence and expectancy concepts in return-to-work research has yet to be reported. This lack of systematic rehabilitation research explicitly derived from expectancy theory is in no small part due to the difficulty in measuring or quantifying concepts of the theory such as 'valence' and 'expectancy'.

The pilot study reported here was designed to establish a methodology for operationalizing the concepts of valence, expectancy and motivation, and to test the hypothesis that these variables would predict return-to-work outcome in a sample of workers' compensation clients. In line with the definition of motivation as a function of both expectation and valence, it was additionally hypothesized that motivation would be more strongly associated with return-to-work outcome than would either valence or expectancy measures alone.

Method

Subjects

Subjects in the study were 32 public sector-employees who were off work and receiving compensation for a work related condition under the Comcare workers' compensation scheme This scheme which covers all federal employees in Australia is a no fault scheme which provides, to those with an accepted claim, salary replacement benefits of 100% of preinjury salary for the first 45 weeks off work, reducing to 75% of preinjury salary for subsequent time off work. Under Comcare legislation, employers have a legal obligation to make available suitable employment and injured workers are required to participate in occupational rehabilitation with the prime emphasis on return to employment.

Clients who met the criteria of being off work for at least four months but less than six months and who had previously undertaken an employer organized occupational rehabilitation program - typically of three months duration - were selected for participation in the study. Subjects were recruited from consecutive referrals over a two-month period to a private rehabilitation agency which was authorized to provide occupational rehabilitation to workers covered by Comcare. All subjects were receiving 100% of pre-injury salary at the time of testing. Time since injury ranged from 17 to 28 weeks with an average of 23.5 (SD = 3.3). The average age of the group was 35.9 years (SD = 5.1). The sample was comprised of 12 males and 20 females. The compensable condition breakdown was: back injury (n = 3), upper limb or neck soft tissue injury (13), stress, anxiety or depression (13); other disease or injury (3).

Procedures

Operationalising the valence of work:

Warr (1987, 1994) has identified nine features of work environments which are related to affective well being. These aspects of environments include: opportunity for interpersonal contact; opportunity for control; opportunity for skill use; externally generated goals; variety; environmental clarity; financial rewards; safety; and, valued social position. Warr's dimensions have been judged by Feather (1990, p38) as providing a 'useful' and 'comprehensive' set of categories for describing (work) environments. Their validity for this purpose is suggested by their compatibility with Jahoda's (1982) functional approach to the analysis of employment. These dimensions were taken as a common set of instrumental or second order outcomes associated with return to work. Whilst acknowledging that individuals may have additional personalized outcome expectations associated with a return to work, a common set of such possible outcomes was chosen for this investigation in order to simplify the assessment of differences in work valence between individuals. The decision to use such a set of general work features in the measurement of work valence was based, additionally, on the results of Schmitt and Son's (1981) predictive study of post-school behavior of graduating students. The results in that study suggested that outcomes used in expectancy theory research should probably involve general categories rather than specific outcomes.

All subjects were interviewed as part of an initial rehabilitation assessment by one or other of the authors. Interviews lasted, approximately, one hour. Warr's dimensions above relating to sources of satisfaction potentially available in work environments were employed as a framework for the semi-structured interviews during which clients were encouraged to explore and discuss their perceptions of the anticipated outcomes they associated with a return to full-time employment. Subsequently the interviewer assigned a rating for each dimension representing the client's expectation that a particular outcome would be associated with a return to work. (Because two of Warr's dimensions, 'externally generated goals' and 'environmental clarity' subsumed such a wide range of job and work role characteristics they were replaced with one single dimension, 'clarity of job requirements,' resulting in a total of eight dimensions.) These ratings were on a five point scale anchored in numerical probability values. For example if the client was judged to be certain that positive social interactions would be a feature of their work environment on return, this dimension was rated 1.0; if they expressed views that suggested they were sure that this would not be the case, a rating of 0.0 was given, and so on for various degrees of certainty.

The instrumental value associated with each of the eight dimensions was therefore defined as the individual's perceived probability of gaining outcomes related to these dimensions on return to work. The valence of work for each individual was calculated by summing the instrumental values associated with each of the above second order outcomes.

Operationalising expectations regarding returning to work:

The individual's confidence in returning to full-time work within six months was assessed using a five-point scale, again anchored in probability values. For example, a client who expressed the views that they were sure they would not be back at work full-time within six months received a rating of 0.0. This way of measuring expectations does not permit a distinction to be made between self-efficacy expectations (e.g. 'I am capable of doing what is necessary to get back to work') and expectations based on other factors (e.g. 'I know there is no suitable job available') but it is assumed that the subjects' confidence in returning to work reflects both these aspects. Using Vroom's formula, M = V x E, motivation was calculated by multiplying the values obtained for valence and expectancy. As a reliability check on interviewer ratings, nine subjects were re-interviewed a week later by the interviewer not involved in the first interview and who was unaware of the original ratings.

The return-to-work status of each of the clients who had undergone the initial interview was reviewed six months later, by which time benefits for all subjects would have been reduced to 75% of preinjury salary, and clients were classified as either full-time at work, part-time at work, or not returned to work.

Results and Discussion

Spearman's correlation coefficients for ranked data were calculated for the ratings of the work dimensions and expectations of return to work for those subjects for whom two sets of ratings were obtained; coefficients ranged from +0.67 to +0.92, suggesting at least moderate to good inter-rater agreement on ratings of subjects' expectations regarding work characteristics and return to work.

The mean scores for valence, return to work expectations and motivation for the three groups defined by their work status six months after assessment (i.e. at work part-time, at work full-time, not at work) are presented in Table 1.

[TABULAR DATA FOR TABLE 1 OMITTED]

Post-hoc comparisons between pairs of means using the Scheffe test showed that both mean expectation scores and mean motivation scores were significantly different (p[less than].05) between the Not working and Working full-time groups, and between the Working part-time and Working full-time groups but not between the Not working and Working part-time groups. Similar post-hoc analysis showed that mean valence scores were significantly different between Not working and both Working groups, but did not significantly discriminate between the two working groups. Contingency tables (Tables 2 and 3) provide a more graphic presentation of the relationships between the operationalised concepts of valence, expectancy and return to work status. In Table 2 the total client group is categorized as high or low on valence score (i.e. above or below the median value for valence in the sample) and, in terms of return-to-work status at six months, is classified as either at work full-time, part-time, or not back at work.

Table 2

The relationship between valence scores and work status six months
later defined as: working full-time, working part-time or not
working (N=32).

                                  Valence
                               High     Low

               Full-time       10        1
Employment     Part-time        6        5
               Not working      0       10

Note: Chi-square = 17.45, df = 2, p [less than] .002
Table 3

The relationship between motivation and work status six months
later
defined as working full-time or not working full-time (N=32).

                                  Motivation
                                 High     Low

               Full-time         10        1
Employment     Not full-time      5       10

Note: Chi-square = 10.496, df = 1, p [less than] .002

In Table 3 the total client group is categorized as high or low on motivation as defined by V x E product (i.e. above or below the median) and in terms of return to work status at six months. Because only expectations regarding the likelihood of return to full-time work were assessed, the return to work status is categorized as back at work full-time or not back at work full-time in this table.

Clearly, from these basic analysis there is evidence of a strong relationship between measures of valence, expectancy, and return-to-work outcome as would be predicted from expectancy theory. However a comparison of the data in Tables 2 and 3 provide no support for the postulated interactive contribution of valence and expectancy to motivation. Valence scores correctly identified the same proportion of subjects who returned to full-time employment as did motivation scores.

Although the findings of this pilot study provide strong support for the hypothesis that the concepts of valence and expectancy predict return-to-work outcome within a sample of workers' compensation clients, the results need to be interpreted with some caution given that the design precluded strict control of injury and other client variables that could potentially influence return-to-work outcomes. Notwithstanding this, the results do suggest the importance of assessing individuals' work related beliefs as part of any comprehensive rehabilitation assessment.

For rehabilitation researchers these findings suggest that concepts from expectancy theory provide appropriate theoretical tools by which we might better understand return-to-work behavior. In this study "general" outcome expectations associated with any workplace were used and no attempt was made to assess individual outcome expectations for particular clients. It was found that even such general outcomes were strongly associated with return-to-work status six months after assessment, which suggests that time spent with the client in understanding their evaluations of key features of their work environment is a valuable rehabilitation activity but one that needs to be informed by knowledge of workplace characteristics. Unfortunately, results from a wide variety of studies of rehabilitation knowledge and competencies (see example Herbert and Wright, 1985; Murphy and Foreman, 1993) indicate that while clinical and counselling domains are both popular with practitioners (see for example the results of Wright, Leahy & Shapson, 1987). Expectancy-valence approaches to motivation however imply that the individual's evaluation of the work-related outcomes need to be taken into account if one is attempting to understand the return-to-work achievements of injured employees who have spent time off work. Moreover, to the extent that the rehabilitation practitioner is required to advise on work re-design that will facilitate the injured worker's re-entry to the workplace, the rehabilitation practitioner needs not only to assess the client's work related beliefs but also the objective characteristics of the work environments and to understand the relationship of these characteristics to the client's values and expectations.

References

Bandura, A. (1983). Self-efficacy mechanism in human agency. American Psychologist, 37, 122-147.

Bandura, A. (1989). Human agency in social cognitive theory. American Psychologist, 44, 1175-1184.

Catchlove, R., & Cohen, K. (1982). Effects of a directive return to work approach in the treatment of workman's compensation patients with chronic pain. Pain, 14, 181-191.

Feather, N. (1990). The psychological impact of unemployment. New York: Springer.

Herbert, J. and Wright, G. (1985). Attitudes of state vocational rehabilitation agencies towards counselor education. Rehabilitation Counselling Bulletin 29, 155-160.

Hester, E. J., Decelles, P. G., & Keepper, K. L. (1989). A Comprehensive Analysis of Private Sector Rehabilitation Services and Outcomes for Workers' Compensation Claimants. Kansas: The Menninger Foundation.

Hood, L. E., & Downs, J. D. (1985). Return to work: A literature review. Kansas: The Menninger Foundation.

Jahoda, M. (1982). Employment and unemployment: a social psychological analysis. Cambridge: Cambridge University Press.

Lawler, E. E. (1983). Expectancy theory. In R. Steers & L. Porter. Motivation and work behavior (3rd ed.). New York: McGraw-Hill.

Landy, F. (1985). Psychology of work behavior (3rd ed.). Homewood, Ill: Dorsey.

Locke, E., & Henne, D. (1986). Work motivation theories. In C. Cooper and I. Robertson (Eds.), International Review of industrial and organizational psychology (pp. 1-35). Chichester: John Wiley and Sons.

McDaniel, J. (1976). Physical disability and human behavior. New York: Pergamon.

Mitchell, L. K., Brodwin, M. G., & Benoit, R. B. (1990). Strengthening the workers' compensation system by increasing client efficacy. Journal of Applied Rehabilitation Counseling, 21 (4), 22-26.

Murphy, C.G. and Foreman, P. E. (1993). General patterns of managerial approaches to work motivation: implications for rehabilitation professional working in occupational rehabilitation. Journal of Occupational Rehabilitation, 3, (1) 51-62.

Roessler, R. T. (1989). Motivational factors influencing return to work. Journal of Applied Rehabilitation Counseling, 20 (2), 14-17.

Sandstrom, J., & Esbjornsson, E. (1986). Return to work after rehabilitation. Scandinavian Journal of Rehabilitation Medicine, 18, 29-33.

Schmitt, N. and Son, L. (1981). An evaluation of valence models of motivation to pursue various post -school alternatives. Organizational behavior and human performance, 27, 135-150.

Tare, D.G. (1992). Factors influencing injured employees' return to work. Journal of Applied Rehabilitation Counseling, 23 (2), 11-20.

Vroom, V. H. (1964). Work and motivation. New York: John Wiley & Sons.

Warr, P. (1987). Work, un-employment and mental health. Oxford: Clarendon Press.

Wart, P. (1994). A conceptual framework for the study of work and mental health. Work and Stress, 8, 84-87.

Wright, G.N. (1980). Total rehabilitation. Boston: Little & Brown.

Wright, G., Leahy, M. & Shapson, P. (1987). Rehabilitation skills inventory: Importance of counselor competencies. Rehabilitation Counselling Bulletin 31 (2), 107-108.

Peter Foreman, School of Behavioural Health Sciences, La Trobe University, Bundoora, 3083 Australia.