a moderated regression

supported. Using a moderated regression analysis with mean transforma- tions, 1 found job complexity to moderate the relationship of cognitive com- plexity and promotability ratings, supporting Hypothesis 8. Table 3 shows the results of this analysis.

Statistical analyses that could be meaningfully conducted for the indi- vidual company data, testing Hypotheses 2, 3, 5, 6, 7. and 8, yielded results that are consistent with those for the combined data.

DISCUSSION

The results of the present researcb provide meaningful insights into the nature of job-related cognitive processes in two field settings. Given tbat the two companies studied were greatly different in terms of employee and organizational demographics, the consistency of tbe results obtained across company samples suggests that the results are robust.

Situational cbaracteristics and individual characteristics, witb the ex- ception of job tenure, are related to the use of cognitive processes in handling job problems. The number of years on a job does not account for the number of script tracks or the type of cognitive structure accessed. Future research should examine employee’s early training and experience on specific job problems ratber than job tenure.

The frequency of job problems’ occurrence was found to be related to tbe number of script tracks accessed (Hypothesis 1). but not to the type of cog- nitive structure (Hypothesis 2). The lack of support for Hypothesis 2 may be due to tbe lack of interrater reliability for the verbal protocol assessments. The verbal protocol measurement of cognitive processes may be more reli- able and valid for laboratory researcb than for field researcb. Additional field research using different procedures to measure cognitive structures is needed to examine Hypothesis 2.

Tbe data indicate that more cognitively complex employees access more script tracks. The results also confirm the relationship of the accessibility of script tracks witb their use. Consistent with the availability heuristic and with the satisficing principle, the results indicate that tbe script track tbat is most available in an employee’s memory is tbe one most likely to be selected for handling a job problem. Because tbis is not a causal analysis, I cannot

TABLE 3 Results of Moderated Regression Analysis