One operator on each shift was responsible for collecting and tabulating the sampled data from all the operators on that shift. This was not much of a burden since the sampling rate chosen (0.4 samples/hour) generated only about 20 to 25 samples per shift. By combining the sampled data from all the operators, we accomplished two important objectives. One, we got an appropriate total sample size while keeping the recording requirement extremely low for individual operators. Two, the process guaranteed anonymity. Anonymity might not be possible in every situation, but where possible, it is desirable.