Saker, Jackie, M.S., Northern Kentucky University; Noblick, Jacob, M.S., Northern Kentucky University; Jenkins, Sylvia, M.S., Northern Kentucky University; Johnson, Te’A, M.S., Northern Kentucky University; Moberg, Philip J., Ph.D., Northern Kentucky University
The present study pursued three objectives, (1) to examine evidence for the dimensionality of workaholism proposed in prior research, (2) to aggregate existing conceptual interpretations to develop a new workaholism scale, and (3) to generate preliminary evidence of dimensionality and construct validity.
In 1971, Oates defined workaholism as employee tendencies to work incessantly. Beyond its core element of substantial investment in work, academic consensus on a construct definition has been elusive (Harpaz & Snir, 2003) due to proposed discrepancies in dimensionality. For example, Spence and Robbin (1992) conceptualized a “workaholic triad” of work drive, work involvement, and work enjoyment. Supplementing this construct definition, Burke (2002) added a dimension of preference for a results-focused orientation. Contending that work enjoyment was not part of the workaholism construct, Schaufeli, Shimazu, and Taris (2009) instead proposed a construct of work compulsivity. One goal of the present study was to examine evidence for the presence of these dimensions by developing a more comprehensive workaholism scale. Thus, by integrating conceptual propositions and empirical findings from prior research, we defined workaholism as a multidimensional construct characterized by elevated levels of work (a) addiction/compulsivity, (b) drive, (c) enjoyment, (d) involvement, and (e) results orientation and then hypothesized that these five dimensions would emerge from factor analysis of newly-constructed items representing these five dimensions.
We constructed an initial item pool to represent the five proposed dimensions, then distilled to 51 for the final scale, and presented the resulting workaholism scale using digital survey administration software. Undergraduate students (N = 203) attending an urban, midwestern university participated in exchange for course credit.
To generate evidence to evaluate inferences of construct validity, we included previously validated measures of three workaholism or workaholism-related constructs. Two measures assessed workaholism directly, the Dutch Workaholism Scale (Schaufeli, Shimazu, & Taris, 2009) and the Workaholism Battery (Spence & Robbins, 1992). The third measure assessed work involvement, the High Involvement Work Practices Scale (Searle, Den Hartog, Weibel, Gillespie, Six, Hatzakis, & Skinner, 2011). Data have been collected and analyses completed.
To examine the dimensionality embedded in responses to the proposed workaholism scale, we conducted an exploratory factor analysis using principal axis factoring with oblique rotation to allow for correlated factors. Three interrelated factors emerged. Factor one was comprised of 11 items (α= .92), explained 25.18 percent of scale variance, and was labeled, “Affective Pleasure.” Factor two was comprised of 6 items (α= .70), explained 8.65 percent of scale variance, and was labeled, “Work Intensity” Factor three was comprised of 5 items (α= .76), explained 8.03% of scale variance, and was labeled, “Compulsive Effort.” Factor intercorrelations were moderate ranging from r = .34 to .48.
To generate evidence supporting validity inferences, we examined factor correlations with external scales. Factor one, Addictive Pleasure, correlated marginally with the Dutch Workaholism Scale (r = .21), but moderately with both the Workaholism Battery (r = .59) and High Involvement Work Practices Scale (r = 56). Factor two, Work Intensity, correlated moderately with all three external scales, the Dutch Workaholism Scale (r = .50), the Workaholism Battery (r = .46), and the High Work Involvement Scale (r = .46). Factor three, Compulsive Effort, correlated moderately with two external scales, the Dutch Workaholism Scale (r = .54) and the Workaholism Battery (r = .47), but marginally with High Work Involvement (r = .25).
We constructed the new workaholism scale to represent five dimensions of workaholism identified in prior research. Examination of the emerging factor structure provided partial support. Although we hypothesized and constructed items to represent five dimensions, addiction/compulsivity, drive, enjoyment, involvement, and results orientation, only three emerged in the present study, Affective Pleasure (enjoyment), Work Intensity (results orientation), and Compulsive Effort (addiction/compulsivity). Two hypothesized dimensions derived from prior research, work drive and work involvement, did not emerge as independent factors.
Based on these analyses, we believe that three dimensions, rather than the five identified in prior research, may characterize workaholism, a sense of pleasure/enjoyment, an orientation toward results and achievement, and a tendency to invest effort compulsively.
A primary limitation of the present study, however, is a threat to generalizability represented by the nature of the focal sample, part- and fulltime employed undergraduate students whose perspectives likely reflect limited work experiences in terms of breadth, complexity, and duration. Continued scale development efforts will require multiple samples of employed incumbents to examine factor invariance and enhance generalizability.
We believe that the present study provides preliminary evidence supporting the development of a new scale to assess workaholism. Additional efforts are needed to (a) clarify the construct definition, (b) revise and generate items, (c) gather data from incumbent samples, (d) incorporate additional measures of potentially workaholism-related constructs (e.g., work engagement, job satisfaction facets, organizational commitment) to facilitate interpretation of the emergent factors.