Training in self-management was given to 20 unionized state government employees to increase their attendance at the work site. Analyses of variance revealed that compared to a control condition (n = 20), training in self-regulatory skills taught employees how to manage personal and social obstacles to job attendance, and it raised their perceived self-efficacy that they could exercise influence over their behavior. Consequently, employee attendance was significantly higher in the training than in the control group. The higher the perceived self-efficacy, the better the subsequent job attendance. These data were significant at the .05 level.

Determined the long-term effects of self-management training given to 20 unionized state government employees to increase their job attendance in a 6-month and a 9-month follow-up study. A repeated measures analysis of variance revealed that enhanced self-efficacy and increased job attendance were effectively maintained over time. Perceived self-efficacy at the end of training predicted subsequent job attendance. The control group (n = 20) was then given the same training in selfmanagement by a different trainer. Three months later, this group showed the same positive improvement as the original training group with regard to increased self-efficacy and job attendance. These findings lend support to a self-efficacy based theory of job attendance.In 1962, As argued that employee absenteeism was a social fact in need of a theory. Twenty years later, Johns and Nicholson (1982) found that despite a heavy investment of research effort, no major breakthrough in the prediction, understanding, and control of absenteeism had occurred.There are at least two explanations for this impasse. The first is the measurement problem. Latham and his colleagues Latham & Napier, 1984;Latham & Pursell, 1975 have maintained that the primary interest of researchers should be to find a reliable measure on which to build theory and on which to evaluate the validity of predictors and the effectiveness of interventions. For this reason, they argued for the necessity of measuring job attendance in addition to, if not in place of, absenteeism. This argument was based on findings from four samples of employees that test-retest reliability coefficients of job attendance are significantly higher than typical measures of absenteeism Latham & Pursell, 1975). This occurs because absenteeism measures are almost always contaminated heavily by human judgment. The measures reflect the categorization behavior (e.g., jury duty, illness, injury, bereavement, vacation) of the recorder rather than the measurements of the researcher. Conversely, the recording of job attendance is relatively straightforward. The measurement is based on observations of employee presence at the work site.For reasons that are not altogether clear, Nicholson and his colleagues (Chadwick-Jones, Nicholson, & Brown, 1982;Johns, 1984; have argued against relying on traditional test-theory concepts such as test-retest reli-Preparation of this article was supported in part by the Ford Motor Company Affiliate Fund to Gary P. Latham.We wish to thank Albert Bandura and Fred Kanfer for their helpful comments on a preliminary draft of this article, and Dawn Winters for her editorial assistance.

In a control-group field experiment using a reversal design, 30 insurance salespeople were randomly assigned to an experimental group that received self-management training. A multivariate analysis of variance and subsequent repeated-measures analyses of variance revealed that, compared with a control condition (n = 30), training in self-management skills significantly improved job performance as assessed through both objective and subjective measures. Performance improvement continued with time, and increases were sustained across a 12-month period posttraining. Subsequent training of the control group produced similar increases in self-efficacy, outcome expectancies, and job performance. Potential mediating effects of self-efficacy and outcome expectancies on the self-management-performance relationship were explored and partially supported.

No abstract

No abstract