Three-process model of supervisory activity over 24 hours.

OBJECTIVES This study used endogenous and exogenous factors to develop a model for sampling supervisory activity that implied consulting numerous data on computer systems. METHODS The study was carried out in an automated workshop of a chemical plant. Five crews worked on a 3x8 hour shift system, with changes at 0400, 1200 and 2000. Each team included 2 experienced controllers who supervised the process from computer systems in a control room. Starting and ending duty at the same time as the operators, the researchers used real time to code of all the screen pages selected on the computers by 8 controllers (from 4 teams) over 18 shifts. RESULTS The time-of-day fluctuations of the call frequency confirmed the existence of endogenous factors (biological rhythms or fatigue), but they cannot explain all the variations observed. The modeling of the data yielded the following 3 explanatory factors: (i) the cognitive demands of the tasks, as an external factor, mainly concern the impact of shift changeover characterized by a strong peak of information gathering at the beginning of the shift, even at 0400; (ii) the shift duration appears as a "fatigue" factor and reflects a rapid reduction of information gathering over the shift; (iii) circadian rhythms are characterized by a minimum of activity at night and a maximum in the afternoon. CONCLUSION In other similar work situations, in addition to an essential work analysis, our model could help the design of shift duration or schedules for shift changeover.