The authors conducted a comparative study of continuous timing and work sampling, as applied by different observers. Twelve researchers were split into two groups and carried out the same study, switching method every half an hour. The study has lasted 3 hours, corresponding to 6 iterations (3 iterations per method). Statistical analysis of data confirmed that both methods may determine elemental breakdown with the same accuracy. Observer-induced variability had a minor effect, and only on those time elements that were short and occurred less frequently. Error increased as the study progressed, as the likely result of observer fatigue. Again, this trend was significant for short elements and not significant for longer ones. This study may offer a warning against over-detailing elemental breakdown. Work cycles should be split in as many elements as strictly necessary for achieving the specific purpose of the study. Splitting them in too many elements just for the sake of description is counterproductive, because it places an excessive strain on the observer and it may increase the risk of errors. A simple study design will facilitate replication by a number of different researchers, regardless of method.

Key words: time studies, methodology, accuracy, work sampling, operations