This field study of nuclear power shiftworkers reports the use of hand
-held computers to record alertness levels, cognitive task performance
and workload ratings. Subjective ratings and performance tasks (Searc
h and Memory-5 (SAM-5); Choice Reaction Time (CRT)) were completed on
the computers at 2-h intervals during selected shifts in the work sche
dule by maintenance crews and reactor operators. Sleep duration and qu
ality data were also obtained during the 35-day shift cycle. Workload
remained relatively stable across shifts. The night-shift was associat
ed with shorter sleep durations and poorer sleep quality. Significantl
y lower levels of alertness and poorer performance for components of t
he performance tasks were also observed on the night-shift. Time-on-sh
ift effects on alertness were evident, whereby for the night-shift, es
pecially, alertness dropped dramatically during the early hours. Compo
nents of SAM-5 performance showed differential time-on-shift patternin
g relative to each shift whereas CRT was unresponsive to time-on-shift
effects. Significant variations were observed in alertness and perfor
mance components during the 12-h night-shift. Perhaps paradoxically, p
erformance on the 12-h night-shift appeared to improve rather than det
eriorate during the early-hour period that was associated with the gre
atest drop in alertness. This was suggestive of extra effort expenditu
re during this critical period.