Chapter 14.

Work Sampling 1

Work Sampling

Introduction

• The work sampling method uses statistical methods to determine the

proportions of total time for various activities that constitute a job.

• Applications
(1) Determining machine utilization
(2) Determining allowances
(3) Establishing standard times

• Assumption: Workers are using standard methods.

• Pros and Cons

Pros Cons
• Reduce tension caused by • Decrease the accuracy of
constant observation of performance rating (difficult to do
workers. a performance rating at each short
• Establish an average standard instant).
time over varying work • Require a relatively large number
conditions. of random observations for high
• Permit simultaneous accuracy.
development of standards for
various operations.

• Procedure

Step 1. Determine the sample size required.

Step 2. Develop a random observation schedule.
Step 3. Record and plot observations.
Step 4. Summarize the percentages of various activities observed,
performance ratings, and volumes of production.
Step 5. Determine machine utilizations, allowances, and standard times.
 Chapter 14. Work Sampling 2

Step 1. Determining a Sample Size

1. Estimation of Percentage

• A preliminary estimate of p is needed for an activity being studied to

determine the sample size of the study.

• The probability estimate can be obtained from historical data or a pilot

study for a short period of time (two or three days).

• The ratio of observations of a given activity to the total number of

observations is used to estimate the percentage of time for the activity.

ni
pˆ =
n
where: n = total number of observations
ni = number of observations for activity i

2. Estimation of Accuracy (Limit of Error)

• Determine the desired accuracy (tolerance, limit of error) of the results

with a confidence level designated.

• Calculation of Maximum Limit of Error (L)

 pq   p(1 − p) 
pˆ ~ N  p,  = N  p,  if np > 5 and nq = n(1 − p) > 5
 n   n 

 pˆ (1 − pˆ ) pˆ (1 − pˆ ) 
P  pˆ − z α ≤ p ≤ pˆ + z α  = 1− α
 2
n 2
n 

pˆ (1 − pˆ )
L = zα
2
n

• The maximum limit of error decreases as the size of a sample increases.

Exercise. Maximum limit of a 95% confidence error where pˆ = 0.8 and

n = 100 .
 Chapter 14. Work Sampling 3

3. Determination of Sample Size (N)

• Adequate sample size is needed to ensure statistical significance.

z α2 / 2 pˆ (1 − pˆ )
N=
L2

Exercise. pˆ = 0.8 , L = 2% , and confidence level ( 1 − α ) = 95%.

• There is a direct trade-off between the accuracy of the study and the
number of observations needed. Thus, we need to check if the extra
sample size is worth for the extra accuracy.
 Chapter 14. Work Sampling 4

Step 2. Developing Random Observation Schedule

• Random observations are needed when doing a work sampling study to

reduce any bias introduced by the worker’s anticipation of observation
times.

(Note) Bias Sources

(1) Worker’s Bias: The arrival of an observer at the work center

immediately influences the activity of the operator.

(2) Analyst’s Bias: There is a natural tendency for the observer to record
what has just happened or what will be happening, rather than what is
actually happening at the exact moment of observation.

• The longer period of time the time study, the better the chance of
observing average conditions.

• The frequency of daily observations depends on:

(1) The number of observations required,
(2) The time available to develop the data, and
(3) The number of analysts available.

Exercise. How many daily observations per observer should be conducted

for a total of 900 observations if 10 days and 3 observers are available for the
study?

• Methods available for random scheduling

(1) Random number table.
(2) Computer program.
(3) Random reminder: pocket-sized instrument which beeps at random
times.

Devilbiss Electronics, Ltd.

Exercise. Develop a random schedule of 100 observations for five workdays.

 Chapter 14. Work Sampling 5

Step 3. Recording and Plotting Observations

1. Guidelines for Observation and Recording

• Make a mark on the floor to show where the analyst should stand before
making an observation.

• Take a visual observation, and then make a written entry after leaving the
work area to minimize the workers’ feelings of being watched and to allow
them to perform in their accustomed manner.

(Note) A video camera can be used to save the travel time of the analyst
and reduce bias in the study.

• If the operator or machine under study is idle, the analyst must determine
the reason for the idleness, and confirm the reason with the line
supervisor before recording the data.

2. Recording Form

• Design a form to record observations. Keep unavoidable delays separate

from personal needs (trips to restroom or drinking fountain).

• A standard form does not exist because each work sampling study is
unique.

(e.g.)
Random Material Unavoidable Performance
No Machine Drilling Set-up
Time Handling Delay Rating
1 8:07 3 v 125
2 8:08 4 v 110
3 8:16 2 v 110
4 9:17 3 v 105
5 10:56 1 v 110
6 11:16 1 v 105
7 11:52 5 v 110
8 13:00 3 v 105
9 13:18 2 v 110
10 13:20 3 v 120
11 13:22 2 v 125
12 13:28 4 v 105
13 13:30 3 v 110
14 13:38 1 v
15 13:48 5 v 100
16 13:57 3 v 105
17 14:10 2 v 110
18 15:11 2 v 105
19 15:18 5 v 110
20 15:45 4 v 110
Total 12 3 4 1 2090
% 60.0% 15.0% 20.0% 5.0% 110
 Chapter 14. Work Sampling 6

3. Plotting Percentages

• Plot p values for each day on a control chart (used in statistical quality
control) using ±3σ limits (99.73%) as the control limits.

Exercise. Determine the control limit for pˆ = 0.6 and n = 500 .

Control Chart - Drilling

p
0.8

0.7

0.6

0.5

0.4

0.3

0.2

0.1

0
0 1 2 3 4 5 6 7 8 9 10
Day

• Points beyond the control limits of p are considered as being out of control.

(1) If the case occurred due to rare reasons (e.g., complete failure of
machine, strike), eliminate the outliers.

(2) If the case occurred due to change of the work methods, modify the
control chart and recalculate the sample size requirement.
 Chapter 14. Work Sampling 7

Step 4. Summarizing Work Sampling Observations

(e.g.) Length of the study period = 400 hrs

Total number of observations = 900
Average performance rating = 110
Production volume = 2,000 pieces.

Material Unavoidable
Machine Drilling Set-up Total
Handling Delay
1 101 26 44 3 174
2 102 34 35 6 177
3 119 34 25 5 183
4 109 24 33 8 174
5 127 30 29 6 192
Total 558 148 166 28 900
% 62.0% 16.4% 18.4% 3.1% 100.0%

Step 5. Analyzing Machine Utilization, Allowance, Standard Time

1. Machine Utilization

Exercise. % drilling task and 95% of confidence interval

z 0.05 = 1.96
2

pˆ (1 − pˆ ) pˆ (1 − pˆ )
pˆ − z α ≤ p ≤ pˆ + z α
2
n 2
n

2. Allowance

• Determine the percentage allowance for unavoidable delays.

Exercise. % unavoidable delay and 95% of confidence interval

 Chapter 14. Work Sampling 8

3. Standard Time

• Determine the observed time (OT) of each work element for a piece:

T × pˆ i
OTi =
P
where: T = total time of the study period,
p̂i = % element i , and
P = total production volume during the study period

Exercise: OT

T × pˆ drilling
OTdrilling =
P

T × pˆ set −up
OTset −up =
P

T × pˆ material _ handling
OTmaterial _ handling =
P

• Determine the normal time (NT) of the work process for a piece:

NT = ∑ OT × R , where
i
i R = average rating of performance

Exercise: NT

• Determine the standard time (NT) of the work process for a piece:

STi = NTi × (1 + Allowance)

Exercise: ST