Failure time analysis of office system use

Given a choice, individuals may experiment with certain types of office automation and accept them wholeheartedly, while others may reject them. This article develops models to characterize the probability of continued use of an office information system. It then tests these models with usage patterns of 210 study participants of the IBM PROFS Professional Office System during a two-year period at the University of California at Berkeley. Two types of longevity analyses were performed, one using survival functions and the other using proportional hazard models. The survival function analysis showed that users did not try PROFS out and then immediately quit. Nevertheless, over a two-year period, 50% stopped using the system. Subjective analysis of the data suggests a considerable number of these users moved to other systems rather than completely stopping use of any office automation software product. Proportional hazard models were employed to determine how the probability of terminating use related to user characteristics. It was found that the longer individuals had access to the system, the lower was the dropout rate. No relation was found between dropout rate and gender, age, or years working for the university. There was a direct relation between the probability of discontinuance and the individual's departmental affiliation, and that a participant was a faculty member. The data did not support conclusions about the relation between other job classifications and discontinuance of use.