Doctoral Dissertations

Date of Award

3-1988

Degree Type

Dissertation

Degree Name

Doctor of Philosophy

Major

Industrial and Organizational Psychology

Major Professor

John M. Larsen

Committee Members

Russell,

Abstract

This investigation focused on the decision-making process in the area of absence inception--the decision to be absent from work. Absenteeism research over the past several decades has been largely concerned with correlates of job satisfaction and summary measures of absenteeism and has produced very disappointing results (Hackett & Guion, 1985). As an attempt to rectify some of these previous shortcomings, in the present study, absence potential was examined in relation to; a) work and nonwork issues, b) absence control policies, c) job involvement, and d) events as stimuli for deciding to be absent as opposed to state-based attitudes.

The research proceeded in three phases. Subjects in each phase were students from a local undergraduate institution evening school program for working adults, representing a broad cross-section of different occupational types and organizational backgrounds.

In the first phase of the investigation, a brainstorming technique was used to develop relevant categories of events that might be related to absence. From this base, a questionnaire was developed which asked subjects about the potential for absence associated with specific events as well as whether certain events had ever contributed to an absence for them. Four work-related categories were identified (perceived job stress, boredom on the job, sense of inequity, and avoidance of problems at work); and five nonwork related categories were identified (feeling badly/sick, family responsibilities, personal business, transportation problems, and leisure preferences). The first phase of the study also determined that typically only one or two issues are considered important to individuals as they make decisions about absence.

The second phase of the investigation was a scaling study where the intent was to develop continuous scales for work events and nonwork events. Several statements were generated for each of the four work events categories and each of the five nonwork events categories. Subjects rated the degree to which each work event would cause them to want to be somewhere other than work and how much the nonwork events would create a pull away from work for the individual. The mean response to each item represented the scale value.

On the basis of the items means and standard deviations, items were then chosen to comprise a work events scale and a nonwork events scale. However, because of generally large standard deviations for most items, the original scaling study sample was reduced by one-third in order to eliminate those subjects with very small standard deviations across items. Item means were recomputed and revised scale values were used in the third phase of the study. Even so. the scales were not conclusively accepted as Thurstone type scales.

The main study again utilized a survey questionnaire where subjects were asked to give the probability that 14 work events, 14 nonwork events and 9 combination work and nonwork events would lead them to be absent on that day. They also responded to six absence control policy items, six job involvement items and several demographic questions.

Nonwork events were shown to be a much stronger predictor of subjective probability of absence than work events in a regression analysis, accounting for an R2 increment of 23% compared to 4% for work events. Similarly, when contrasted against work events in a repeated measures analysis of variance, the average subjective probability of absence for nonwork events was significantly greater than work events. Neither the salience of the work events scale nor of the nonwork events scale was moderated by absence control policies or job involvement. The concept of absence proneness was supported by a positive correlation ( r=.58, p≤ .001) between the average subjective probability of absence for work events and that of nonwork events. Lenient absence control policies and iow job inv'olvement were positively related to average subjective probability of absence, each accounting for about 7% of variance. No interaction effects were found for these two variabies. The demographic variabies studied were insignificant predictors of subjective probability of absence either as single variables or as interaction effects. Total variance accounted for by job involvement, absence control policies and the demographic variables as predictors of average subjective probability of absence was 14%.

Download
Files over 3MB may be slow to open. For best results, right-click and select "save as..."

Share

COinS