The following work sampling data was collected from observing a retail assistant: 

Activity                                                Number of observations

Attending to customers                                  60

Telephone enquiries                                        150

Walking around                                              40

Reading data                                                 30

Talking to manager                                         40 

Talking to other workers                                 20 

Lunch/coffee/comfort breaks                          50 

Dealing with customer complaints                  110 

Total                                                          500

The figures mean that, say, over the course of a day, it is estimated that 22 per cent of the time was devoted to dealing with customer complaints. 

(a) Determine the limits of accuracy of the above calculation 

(b) If the accuracy needed was to be ±1 per cent, determine the appropriate sample size. 

(c) How would an operations manager use the data obtained from the exercise described?

Answer: (a) N = 4P (100 - P) /L2 

 N = 500 

 P = 22 per cent 

 L = SQRT (4 × 22 (100 - 22))/500) = 3.7 per cent 

Therefore, P = 22 per cent ±3.7 per cent (ie we are 95 per cent confident that the true value lies between 18.3 per cent and 25.7 per cent) 

(b) 88(78)/1 = 6864. Unlikely to be economically worthwhile - a sample size of 1716 would give ±2 per cent accuracy 

(c) For setting targets of improvement, determining areas of improvement, cost savings, staff training.