1 Work Sampling Can provide information about men and machines in less time and lower cost. It has three main uses: 1.Activity and delay sampling To measure the activities and delays of workers and machines. 2.Performance sampling To measure working time and performance index of a person on a manual task. 3.Work measurement To establish a time standard for a manual task.
2 How many observations? The percentage number of observations recording idle state (worker or machine) is a reliable measure of the percentage time that the operation is in the delay state. The accuracy increases as the number of observations is increased. Work sampling in its simplest form: 1.Make observations of one or more operators or machines at random intervals. 2.Note if they are working or idle. 3.Give a tally mark under "working" if the operator or machine is working and give a tally mark under "idle" if the operator or machine is idle.
3 4.Obtain the ratio of number of idle tally marks to the total number of idle and working tally marks to calculate the percentage of the day that the worker or machine is idle. Three methods could be used to obtain the sample size: 1.The formula for determining the sample size for a confidence level of 95% (2 sigma) is: Where S = desired relative accuracy p = percentage expressed as a decimal (percentage occurrence of an activity or delay being measured using a trial study) N = number of random observations (sample size) 2.Using the alignment chart. 3.Using tables.
4 After the start of the work sampling, it is advisable to recalculate the sample size based on the current results (calculate p based on the current results for idle and working). This will better evaluate the progress of the study and the resulting sample size might be lower which will save time and cost. Recalculation is advised at regular intervals (perhaps at the end of each day). How can we determine whether the results are within the desired accuracy after the study is completed? This can be done in two ways: 1.First obtain percentage occurrence of an activity or delay (p) using the results for the number of idle and working observations and then use the above formula to calculate S instead of N (which is already known).
5 2.Using tables (see table 57). The resulting accuracy should be compared to the desired accuracy. The result will be satisfactory if the resulting accuracy value is lower than the desired accuracy. If the resulting accuracy value is higher than the desired accuracy value, then N should be recalculated and work sampling should be continued. What is the true accuracy value? The resulting accuracy value will help us to obtain the upper and lower value for the delay or working percentage with a 95% confidence level.
6 Example: After a work sampling study of a machine, the results showed that the percentage time the machine was idle was 35% (p=.35, idle observations=1400, and total number of observations=4000). The accuracy was determined using the above formula and was ±4.3% (S = ± 0.043). Multiply the value for S into the p value to get the upper and lower value for the accuracy (±4.3% X 35% = ±1.5%) and then add and subtract the resulting value to the percentage idle time to get the upper and lower value for the delay ( and 35─1.5). In this case, the true value was between 36.5% and 33.5%. This means that we are 95% confident (we are using 2 sigma) that the percentage idle time will be between 33.5% and 36.5%.
7 Control Charts Control charts in work sampling studies enable the analyst to plot the daily results. If a plotted point falls outside the control limits, this is likely to indicate that some unusual or abnormal condition may have been present during that part of the study. In control charts, 3 sigma is usually used in determining the upper and lower control limits. This means that there are only three chances in 1000 that a point will fall outside the limits due to a chance cause. It can be safely assumed that when a point falls outside the limits, there is a reason for it. The formula for determining the control limits for p is: Where n= number of daily observations= total number of observations/number of days studied.
8 See figure 246 as an example. Alignment charts could also be used to determine the upper and lower control limits for p. See figure 247 as an example. Use of Random Number Tables Work sampling requires the observations to be random, unbiased, and independent. To ensure that the sample is random, random numbers tables are used. The table will help to determine the time of day that an observation should be made. It may also show the order in which the operators should be observed, or the specific location in the department or plant where a reading should be taken. See tables 61 and 62.
9 Performance Sampling Performance sampling is simply assigning a rating of the operator's performance or speed for each working observation. The rating (in percent) is recorded on the observation sheet the same as in making a time study. Each working observation will have its own performance rating. The performance ratings are then added up and divided by the total number of working observations to represent the average performance of the observed operator for the period covered by the study.
10 Continuous Performance Sampling To have a control on labor cost, time standards for specific operations could be established (by the use of motion and time study techniques such as MTM systems) and then a count of the number of units finished each day could be obtained. The actual results could be compared to time standards to obtain a performance index for a worker or department. This method is widely used and is very effective in many situations. But sometimes much work could not be measured directly, the cycles may be long and varied, methods may not be standardized, and it is often difficult to obtain a count of the units of completed work. In these situations, some control of labor cost could be achieved by the use of work sampling. Continuous performance sampling can be performed by making observations of all workers in a department at random during the entire week or month and computing the results for this period.
11 The process could be repeated week after week to provide management with information related to the work force such as: 1.Percentage of time working. 2.Percentage of time out of department. 3.Percentage of time idle. 4.Average performance index while working. 5.Labor effectiveness (item 1 x item 4).
12 Productivity Sampling Performance sampling could be used to improve productivity. Boeing has used performance sampling to improve productivity for many years and at the present time their program called "Productivity Sampling" covers more than 3000 people in 120 shops. The company uses productivity sampling to increase productivity and reduce costs in the shops. Areas in need of improvements are pinpointed and specific problems are revealed by the use of productivity sampling. See figure 250 and pages 434 and 435 (Motion and time study design and measurement of work by Barnes) for more details of the Boeing program.
13 Work Sampling in Non-manufacturing Activities Work sampling is used in non-manufacturing environments such as banks, restaurants, hospitals and department stores to determine daily and hourly personnel requirements and for cost control. Large companies usually make work sampling studies of the activities of personnel in many departments. Determining Time Standards by Work Sampling In short-cycle repetitive work situations; time study, standard data, or predetermined time data such as MTM systems are usually preferred for establishing time standards. But long- cycle operations and group work could benefit more from work sampling studies in determining time standards.
14 The steps for determining standard time using work sampling are as follows: 1.Obtain the percentage of the day that the operator is working. 2.Determine the average performance index for the operator during the working portion. 3. Determine the total working time during the day of the study. 4.Determine the number of finished pieces during the day of the study. 5.Multiply items 1, 2 and 3 together. 6.Divide item 5 by item 4. 7.Add the allowances. See figure 251 as an example. Work sampling could also be used to obtain the time standard of a group of operators performing a manual assembly task. See figures 252 and 253 as an example.