Center for Demography and Ecology University of Wisconsin-Madison Contributions of Job Strain and Workaholism to Work-Family Conflict

Work and family influence the health and health behaviors of employed adults, especially when these domains have conflicting or competing demands. The extent to which work-family conflict is a function of extrinsic factors, such as the organization of work vs. an individual’s work habits, remains unclear. We conducted a mail survey of randomly selected working men (n=655) and women (n=825), ages 30-65, enrolled in a large managed care organization (MCO) to investigate the relationships among workplace characteristics, work habits and work-family conflict. Standard scale items assessed job demands, job control (skill discretion, decision authority) and social support. Using cluster analysis to assess multiple dimensions of work behavior, we classified workers into one of three categories of work habits: workaholic (n=186), unengaged (n=558), and average workers (n=642). We measured work to family conflict by asking respondents how frequently work responsibilities intruded into their family life and the level of distress caused by these intrusions. In multivariate regression models adjusted for age, sex, marital status, major occupational group, and educational attainment, we observed significant and independent associations between measures of workplace characteristics, work habits, and work-family conflict. More specifically we observed significant associations between work-family conflict and social support (b=-0.30, p<.01), decision authority (b=-0.51, p<.01), and a significant interaction effect between job demands and skill discretion (b=0.27, p<.01). Workaholics reported significantly higher work-family conflict (b=1.45, p<.01), while unengaged workers reported significantly lower work-family conflict (b=-0.80, p<.01). We found no interaction between work environment and work habits. The issue of workaholism needs to be included in discussions of policies to reduce work-family conflict. Job strain, excessive work and work-family conflict 3 Introduction Over the last two decades dramatic changes in the labor market, especially the increasing number of women and dual-career couples in the workforce, have hastened the need for equally dramatic changes in policies that take into consideration the competing demands of work and family. Despite a growing interest from employers, labor advocates, and the popular press and government-sponsored initiatives for research to examine the interface between work and family, questions still remain about the causes and consequences of work-family conflict. Much of the recent research has concentrated not on the antecedents that lead to conflict, but on the deleterious consequences that can arise from the competing demands from both domains, leaving a gap in the evidence base that is needed to inform future policy and interventions to prevent or reduce work to family conflict. Questions remain as to whether there are specific job conditions, work characteristics, or working behaviors that lead to greater conflict between domains. In this study we attempt to address this gap by examining whether work-related behavior and psychosocial working conditions are independent predictors of work to family conflict in a sample of salaried workers. In addition to theory from occupational health psychology on job stress, we address in detail the emerging theories and conceptualizations from psychology and organizational sociology of workaholism and, using an integration of these perspectives, test our hypothesis. Specifically we hypothesize that work-related behavior, such as excessive or workaholic behavior patterns, predict work-family conflict independently of job conditions, conceptualized as job demands, decision latitude, and social support. Background Work family interface Job strain, excessive work and work-family conflict 4 The work-family interface is a term that most often describes the intersections of the social roles that are associated with these two domains of life. Within this literature researchers have concentrated much of their effort on work-family conflict, or the conflict that arises when one’s effort to perform tasks in one domain interferes or is incompatible with the demands imposed by the other (Greenhaus & Beutell, 1985). In recent studies, researchers have recognized that different stressors may affect the direction of conflict, and as a result, they have encouraged others to specify and measure whether the conflict originates from work interfering with familyrelated responsibilities (work to family conflict or WFC) or from family-related responsibilities interfering with work (family to work conflict or FWC). This distinction is considered essential for developing a more precise understanding of the antecedents and consequences of work-family conflict that can be used to inform future policy. Our data did not extend into possible family stressors and involvement, but instead converged on working behaviors and the work enviornment. We concentrate our efforts, therefore, solely on WFC. While more recent reviews of the literature have summarized the deleterious consequences of work to family conflict on life and job satisfaction (Kossek & Ozeki, 1998), turnover, absenteeism, and mental and physical health (Allen, Herst, Bruck, & Sutton, 2000), fewer studies have examined its antecedents. In a 1985 review of 18 studies, Greenhaus and Beutell (Greenhaus & Beutell, 1985) identified important links between time constraints, role conflict and ambiguity, lack of autonomy, lack of support, and psychological demands to work-family conflict. They also recognized some evidence linking work-family conflict to behavior, and suggest that excessive work involvement or commitment to work that are often linked to career success may also lead to WFC. (p.85) Frone (Frone, Russell, & Cooper, 1992) built on this work and work from others to develop a larger conceptual model of the work-family interface, of which the antecedents of WFC Job strain, excessive work and work-family conflict 5 were both job stressors and job involvement. While Frone’s larger model presents a more comprehensive representation of the possible direct and indirect relationships among these antecedents, WFC, FWC and potential consequences, we concentrate our hypotheses on one part of this model: the relationship between job stressors, patterns of excessive job involvement and WFC. We suggest that job stress and workaholism independently lead to work-family conflict, even after adjusting for differences in age, gender, family composition, and occupational status. Job stressors In occupational health psychology, the dominant model used to conceptualize how job stressors lead to poor health outcomes has been Karasek’s job strain or demand/control model (Karasek & Theorell, 1990; Karasek, 1979). This model has played a pivotal part in the process of conceptualizing job stress, and is the most recognized and widely-used model to describe the psychosocial work environment’s effect on health. The model posits that psychosocial job conditions, namely the psychological demands and level of worker control, are differentially distributed across occupational groups and that workers with jobs that are demanding, yet lack a sufficient degree of control over the work process are at the greatest risk for psychological strain that, in turn, can lead to psychological distress and poor physical health outcomes. Job demands are most frequently conceptualized as the pace, effort, and volume of work, while control is typically examined in terms of “decision latitude.” Decision latitude, more specifically, is considered to be comprised of decision authority, or the authority to make decisions concerning the work, and skill discretion, or the ability to use one’s skills at work. By testing the possible effects of social support on the demand-control relationship, Johnson, Hall and Theorell further modified the model to show that jobs that pose the greatest risk to health and mortality are those Job strain, excessive work and work-family conflict 6 with high demands, low decision latitude and low social support (Johnson, Hall, & Theorell, 1989). The job strain model has been used extensively to predict all-cause mortality (Amick et al., 2002; Falk, Hanson, Isacsson, & Ostergren, 1992) cardiovascular disease (Alfredsson, Hammar, & Hogstedt, 1993; Schnall, Landisbergis, & Baker, 1994) and mortality (Alterman, Shekelle, Vernon, & Burau, 1994; Johnson et al., 1989; Kivimaki et al., 2002), precursors to cardiovascular disease (Schnall, Schwartz, Landsbergis, Warren, & Pickering, 1992), and other outcomes associated with psychological and physical strain, such as musculoskeletal disorders (Nahit, Pritchard, Cherry, Silman, & Macfarlane, 2001; Skov, Borg, & Orhede, 1996; Waldenstrom et al., 2002), psychological disorders (Cropley, Steptoe, & Joekes, 1999; Mausner-Dorsch & Eaton, 2000; Stansfeld, Fuhrer, Shipley, & Marmot, 1999) and workplace injuries (Murata, Kawakami, & Amari, 2000; Myers et al., 1999). It has been used only infrequently in studies of work-family conflict (Duxbury, Higgins, & Lee, 1994), even though identified antecedents to WFC are often characterized in similar terms to components measured within the model: high psychological demands (including time demands), lack of control or autonomy, and lack of social support (Voydanoff, 1988a, 1988b). Workaholism Separate from, but possibly related to issues of job strain (Greenberger, Porter, Miceli, & Strasser, 1991) is the concept of workaholism. This term, although commonly used and frequently seen and heard in the popular press, has not been as thoroughly studied as job strain and its definition, meaning, and potential sequelae remain understudied. In this section we describe and summarize how various disciplines have conceptualized workaholism. Job strain, excessive work and work-family conflict 7 Within the psychological literature, three inter-related explanations have been proposed for why people engage in excessive work behavior. One line of thinking about workaholism has taken to heart the term’s implied reference to chemical addictions such as alcoholism (Killinger, 1991; Robinson, 1998; Schaef & Fassel, 1990), and has focused on the potentially addictive nature of the rewards (both intrinsic and extrinsic) frequently associated with work (Oates, 1971; Porter, 1996). A second line of thinking is that excessive work effort may reflect an underlying obsessivecompulsive personality disorder (Naughton, 1987). A third, less dramatic psychological explanation of excessive work is based on the notion that some people simply derive such fulfillment, pleasure, and enjoyment from their work that they have a tendency to lose themselves in it (Machlowitz, 1980; Scott, Moore, & Miceli, 1997). What these three perspectives share in common is a view of the phenomenon of excessive work focused inwardly on the foibles, frailties, or pathologies of individual workers. While such aspects of individual workers almost certainly come into play for some individuals, these perspectives are limited by their inattention to potential systemic or environmental factors that may explain the tendency toward excessive work for many. Over the past several years, a variety of writers in the popular press have argued that work in the Western world during the late 20 and early 21 centuries has become, for many, a hyperindividualistic, and often all-consuming quest for meaning, self-actualization, and social status(Ciulla, 2000; Hochschild, 1998; Nash, 1994; Schor, 1993). While the explanations described above focus on largely non-rational responses to work, others have sought to understand workaholism in terms of a generally rational response on the part of individuals to both positive and negative environmental conditions and incentives. Clearly, the possible explanations for why people work excessively are many and varied. We find the concept of individuals’ responses to incentives a potentially important unifying principle. Individuals may work Job strain, excessive work and work-family conflict 8 excessively in response to positive incentives such as the intrinsic rewards they derive from their work, or from the more utilitarian aspects of monetary gain and material consumption that such work may provide. By contrast, individuals may also work excessively in response to negative incentives such as attempting to avoid or escape consumer debt. More darkly, in a competitive work culture, some may work excessively out of fear that if they are not perceived by their bosses, supervisors, and peers as striving to excel, they may be passed over for promotions or raises, and may be at greater risk of layoff than their more workaholic peers. Excessive work may also derive from the normative standard of judging work productivity in terms of total per-worker output, rather than worker output per hour worked. Finally, for some, working excessively may provide a positively sanctioned refuge from difficult family relationships, responsibilities, or obligations. The U.S. has been characterized as having a “culture of over-work.” (Andresky Fraser, 2001)(p. 200). Most recently, Joe Robinson has argued that excessive work can be described as a sort of bravado, arising naturally from the fact that we live our lives embedded in a culture that values productive endeavor above all else (Robinson, 2003). With such cultural imperatives operating, the potential linkages between excessive work and WFC become obvious, with or without suppositions of illness or pathology of individual workers. And yet, while the term “workaholism” is commonly used in the popular press and everyday conversation, the concept is currently not well defined. Over the past two decades, research in the business management, organizational behavior, and human resources literatures has begun to explore the role of person-level factors and coping strategies, (excessive work-effort and other control-enhancing strategies, in particular), as determinants of outcomes directly affecting organizations, including job satisfaction, productivity, and turnover (Porter, 1996; Rabinowitz & Hall, 1977; Scott et al., 1997; Spence & Robbins, 1992). Unfortunately, little of this Job strain, excessive work and work-family conflict 9 work has formally examined the impact of workaholism on families or on outcomes of work to family conflict. In our review of the academic literatures, we have found only one study attempting to develop empirical measures of workaholism (Spence & Robbins, 1992), one study that has applied these measures to examine the relationship between workaholism and work-life conflict (Bonebright, Clay, & Ankenman, 2000), and one study that has endeavored to offer a theoretically based definition of the term (Scott et al., 1997). Our paper extends this line of research by integrating the empirical work of Spence and Robbins with the theoretical expectations of Scott et al. to develop a new empirical measure of workaholism. As with work itself, the idea of excessive work effort is complex and multi-dimensional. We have chosen here to measure the concept in terms of experiential and behavioral components, since motivational and attitudinal aspects are not as readily captured with a cross-sectional survey study design. This study combines theory from the job strain literature with more recent organizational literature on person-level factors. Common themes in these literatures are the importance of “control” in the work setting, and excessive or compulsive work effort. Examined together, these literatures suggest the possibility that excessive work behavior may be an expected response to perceptions of a lack of control, or a reduction in control among workers who formerly felt a high level of control. For this reason we hypothesized that a combination of person level, namely excessive or compulsive work behaviors, and psychosocial work characteristics would predict work to family conflict. Methods Sample Data from this study come from a postal survey of working men and women who were insured and received health care from a large Midwestern managed care organization. Participants Job strain, excessive work and work-family conflict 10 were identified using a two-stage sampling procedure. First a subset of both large and small employer groups that purchase health insurance from the organization and represent a range of companies and institutions with predominately salaried, professional employees were selected. The employer groups chosen included colleges and universities, law firms, publishers, insurance, financial, and investment firms, defense contractors, medical groups, hospitals, real estate firms, marketing companies, and consultant firms. For the second stage of the procedure, a random sample of 3,000 male and female employees, between the ages of 30 and 65, from all of the selected employer groups were chosen and asked to complete the study’s mailed survey. We targeted salaried workers for two reasons. First, the nature of salaried work may be more conducive to excessive work patterns than hourly work. In general the expectations for salaried workers are that they complete their work duties and meet their deadlines whatever the time requirements may be. Second, salaried work may be differentially selective of individuals with personality traits that impel them towards excessive work. Measures Dependent variable Work-family conflict. To assess WFC, participants were asked to report how frequently they missed family activities or obligations because of work or work-related responsibilities and the level of distress caused by the conflict (Fox & Dwyer, 1999). Frequency and distress scores ranging from 1 to 4 were then multiplied to create a WFC score from 1 to 16. Independent variables Work environment scales. For this study we assessed job demands, decision latitude and social support using a series of statements based on items in Karasek’s Job Content Questionnaire (JCQ) (Karasek, Gordon, Pietrokovsky, Frese, & Pieper, 1985). Fifteen statements were used to Job strain, excessive work and work-family conflict 11 assess decision latitude (9 for decision authority and 6 for skill discretion), 4 for job demands, and 6 for social support. Participants responded to statements using one of 5 response categories (rarely/never, sometimes, often, usually, always/most of the time). Items for each scale were summed and used in analyses as a continuous measure. Work behavior. Work behavior was assessed using a set of seven summative scales. Six of these scales assessed work attitudes and habits and were reproduced or adapted from work by Spence and Robbins (Spence & Robbins, 1992). Reproduced scales included those that measured ambition (driven), job involvement, perfectionism, and non-delegation. Adapted scales included items from other Spence and Robbins scales but were reconfigured to address two theoretical constructs proposed in the literature on excessive work patterns: thinking about work outside of work and discretionary time spent on work (Scott et al., 1997). The seventh scale, a validated subscale from the Zimbardo Time Perspective Inventory (Zimbardo & Boyd, 1999) assessed the extent to which an individual adopts a future-oriented view of the world. For all items in each of these scales, participants were asked to respond to each item using a 5-point scale ranging from very untrue (1) to very true (5). The Driven scale contained seven items assessing intrinsic work drive or ambition. Example items from this scale include: “I seem to have an inner compulsion to work hard, a feeling that it's something I have to do whether I want to or not,” and “I feel obligated to work hard, even when it's not enjoyable.” Job Involvement was measured with a seven item scale assessing identification with and involvement in one’s job. Example items from this scale include: “My job is a very important part of who I am,” and “I am deeply committed to my job.” The Perfectionism scale contained eight items assessing perfectionistic behavior. Example items from this scale include: “I can't let go of projects until I'm sure they’re exactly right,” and “My Job strain, excessive work and work-family conflict 12 standards sometimes make it difficult to get everything done.” Non-delegation was assessed using an seven item scale, with example items including: “I'd rather work overtime to finish a project myself than delegate portions of it to someone else,” and “I double check other’s work so thoroughly that delegating tasks saves me no time at all.” Thinking about work outside of work was assessed using an four item scale, with items such as: “I leave my worries at the office,” (reverse coded) and “I often find myself thinking about work even when I want to get away from it for a while.” A seven item scale was used to assess spending discretionary time on work activities. Sample items from this scale include: “I usually take some job-related work along on vacations,” and “I work most nights and weekends.” Future time orientation was assessed using a thirteen item scale, sample items from which include: “Meeting tomorrow’s deadlines and doing other necessary work comes before tonight’s play,” and “I take each day as it is rather than plan it out,” (reverse coded). Scores for each scale were then summed and, in order to standardize the variance for the scales, transformed into z-scores. Using a method proposed by Spence and Robbins, we conducted an exploratory cluster analysis of z-scores to classify workers into a set of work pattern profiles using Ward’s minimum variance method. In preliminary analyses we examined solutions that produced between three and eight clusters. The three cluster solution was conceptually distinct and interpretable and we labeled the clusters “average,” “workaholic,” and “unengaged” work behavior. All cluster solutions identified the same “pure” workaholic group, as well as more variable definitions of groups that could be classified as “average” and “unengaged.” Solutions with more than three clusters tended to pull people from the average group into less conceptually distinct clusters, some of which appeared to represent other types of workaholism as proposed by Scott et al. (e.g., achievement-oriented workaholics, compulsive workaholics). These results Job strain, excessive work and work-family conflict 13 support the proposition set forth by Scott, Moore & Miceli (Scott et al., 1997) that “workaholism” may not be a single behavioral pattern, but may represent various sub-types of excessive patterns, such as compulsive-dependent, perfectionistic, and achievement-oriented habits. Results of an eight cluster solution have been presented elsewhere (Martinson, 2002). However, since the threecluster solution subjectively produced clusters that were conceptually more distinct and interpretable we use it in all subsequent analyses presented here. Covariates Because work family conflict has been shown to differ by age, gender, and marital status, we controlled for their effect in multivariate models. Educational attainment and occupation class are also included as control variables in multivariate models. Job title and usual activities were used to classify individuals into 8 occupational classes based on the first two digits of the Standard Occupational Code (SOC). Analysis strategy The first part of these analyses was to describe both the study sample and scale characteristics. For the second part we created a series of multiple linear regression models in order to address whether work characteristics and work behavior was associated with WFC, while adjusting for age, gender, marital status, occupational code, and educational attainment. In the first model of this series we fit the control variables. We then added the work characteristics, including an interaction term between demands and decision latitude to the model. In the final model we added work behavior. Beta and standard deviations are presented in respective tables.