Chapter 16

Work Sampling
Outline:

1. How Work Sampling Works

2. Statistical Basis of Work Sampling

3. Application Issues in Work Sampling

Work Sampling Defined

Statistical technique for determining the

proportions of time spent by subjects in various
defined categories of activity

 Subjects = workers, machines

 Categories of activity = setting up a machine,

producing parts, idle,…… etc.
Work Sampling Defined

 For statistical accuracy

◦ Observations must be taken at random times

◦ Period of the study must be representative of

the types of activities performed by the
subjects
When is Work Sampling Appropriate?

 Sufficient time is available to perform the study

◦ Several weeks are usually required for a work sampling
study

 Multiple subjects
◦ Work sampling suited to studies involving more than one
subject

 Long cycle times for the jobs covered by the study

 Nonrepetitive work cycles

◦ Jobs consist of various tasks rather than a single
repetitive task
Example 16.1: How Work Sampling Works
 A total of 500 observations taken at random times during
a one-week period (40 hours) on 10 machines with
results shown below.

Category No. of observations

(1) Being set up 75
(2) Running production 300
(3) Machine idle 125
500
 How many hours per week did an average machine spent
in each category?
Example: Solution

 Proportions of time determined as number of observations in each

category divided by 500

 Time in each category determined by multiplying proportion by total

hours (40 hr)

Category Proportion Hrs per category

(1) Being set up 75/500 = 0.15 0.15 x 40 = 6
(2) Running production 300/500 = 0.60 0.60 x 40 = 24
(3) Machine idle 125/500 = 0.25 0.25 x 40 = 10
1.00 40
Work Sampling Applications
 Machine utilization - how much time is spent by
machines in the production category.

 Worker utilization - how workers spend their time

 Allowances for time standards - assessment of delay

components in PFD allowance factor

 Average unit time - determining the average time on

each work unit

 Time standards - limited statistical accuracy when

standards set by work sampling
Statistical Basis of Work Sampling
 Binomial distribution, in which parameter p = true proportion of time spent in a given
category of activity (probability of each category/activity)

 There are usually multiple activity categories, so we have p1, p2, . . , pk, . ., pK
proportions for K different activity categories. In the previous example, k=3.

 The binomial distribution can be approximated by the normal distribution, where:

 n p Mean is the average of data

p1  p 
Standard deviation shows you how much your data is spread
out around the mean or average. For example, are all your

p  scores close to the average? Or are lots of scores way above (or
way below) the average score?
n
Estimating the Proportion p

 In a sampling study, we let

p̂ = the proportion of the total number of

observations devoted to an activity category of
interest

 The proportion p̂ is our estimate of the true

value of the population proportion p
Confidence Intervals

 The general statement of a confidence interval for p̂

relative to p can be expressed as follows

 ˆ
p  p 
Pr  z / 2    z / 2  1  
 ˆ p 
 
 Work Sampling Method
 Select a sample size so that the estimate of the proportion
of time spent on a particular activity that does not differ
from the true proportion by more than a specified error, so

pˆ  c  p  pˆ  c
where
p̂ = sample proportion (number of occurrences
divided by the sample size)
c = maximum error in the estimate
 Work Sampling Method
 As the binomial distribution applies, the maximum error of
the estimate is

pˆ 1  pˆ 
c  z / 2
n
where
n = sample size
z = number of standard deviations needed to
achieve the desired confidence

 Solving for n
2
 z / 2 
n  pˆ 1  pˆ 
 c 
Confidence Intervals
 This can be rearranged to the following:

Pr pˆ  z / 2ˆ p  p  pˆ  z / 2ˆ p 1  

pˆ  z / 2ˆ p  p  pˆ  z / 2ˆ p
 Example
 Based on the previous example, determine the 95% C.I. for
the proportion of the time spent setting up the machine
(Category 1)

 Solution:
From the results of the previous example, n= 500:
p̂ = 0.15

0.1187  p  0.1813
Number of Observations Required

 How many observations are required to achieve a given

confidence interval about the estimate of p?

 We need to decide two parameters:

1. Confidence level 1 - 

 This allows us to find the corresponding value of z/2

2. The half-width c of the confidence interval, defined as

the desired acceptable deviation from p

 Thus, we have p  c
Number of Observations Required

 Given z/2 and c, the number of observations

required to achieve the specified confidence
level is given by the following.

z / 2  p̂1  p̂ 
2
n 2
c
Example:

Example:

n
 z / 2  pˆ 1  pˆ  1.96  0.21  0.2
2

2
 683
2 2
c (0.03)
 The n is estimated before starting the work sampling
study.

 After certain number of observations (30% to 40% of n),

you have to re-estimate n again. If it is greater than the
original n, add the remiaing number of observations.

 It is a good practice to re-estimate n again after

performing about 75% of n.
Problem 16.1
 For the data in Example 16.1, determine a
90% confidence interval for the
proportion of time spent running
production, category (2).
 Solution
 Solution: = 0.0219
 For 90% confidence interval, = 1.65
pˆ  z / 2ˆ p= 0.60 – 1.65(0.0219) = 0.60 –
0.0361 = 0.5639

pˆ  z / 2ˆ p
= 0.60 + 1.65(0.0219) = 0.60 +
0.0361 = 0.6361
Problem 16.2
 The allowance factor for personal time, fatigue,
and delay (PF&D) is to be determined in the
machine shop area. If it is estimated that the
proportion of time per day is spent in these
three categories (personal time, fatigue, and
delay are grouped together to obtain one
proportion) is 0.12, determine how many
observations would be required to be 95%
confident that the estimated proportion is
within ±0.02 of the true proportion?
Solution:
 For 95% confidence interval, z / 2 = 1.96;
and interval value c = 0.02

 n = (1.96)2(0.12)(0.88)/(0.02)2 = 1014.2
rounded to 1015 observations
Determining Average Task Times

 Average task time for a given work category is

determined by computing the total time
associated with the category and then dividing
by the total count of work units produced by
that category.
Determining Average Task Times

pi (TT )
Tci =
Qi

Where;

Tci = average task time,

pi = proportion of observations associated with category i,

TT = total time,

Qi = total quantity associated with category i.

 Determining Standard Times
 Similar to determining average task time, except
performance rating must be factored in

 First determine normal time for activity i

Tni 
pi TT  PRi 
Qi
The average value of the performance ratings for
PR 
i
all observations in category i
 Then determine standard time

Tstdi = Tni(1 + Apfd)

Example:
 In the previous example, assume that 1572 units were
completed by the 10 machines and that a total of 23
setups were performed during the 5 days period.
Determine

(a) The average task time per unit during the production

(b) The average setup time.

Solution:
The duration of the study is 40 hours, during which 10
machines were observed.
Example:
(a) Hence, the total time is 10x40 = 400 hrs.

TT = 400 hrs

p̂ of production is 60% = 0.60

pi TT  0.6400 
Tc    0.1527 hr  9.16 min
Qi 1572

Note: all produced parts are the same?

Example:

(b) The 23 setups are associated with “being in setup”

category with:
p̂ = 0.15

pi TT  0.15400 
Tc    2.609hr 156.5 min
Qi 23
 Problem 16.6
 The Chief IE in the production department wanted to know what
value of PFD allowance to use for a particular section of the shop.
A work sampling study was authorized. Only three activity
categories were considered: (1) production work, (2) personal time,
rest breaks, and delays, and (3) other activities. Over a four-week
period (40 hours/week), 100 observations were made at random
times. Each observation captured the category of activity of each of
22 production workers in the department. Results indicated that
category 2 constituted 19% of the total observations. (a) Define the
limits of a 95% confidence interval for activity 2. (b) If a total of 522
work units were produced during the 4 weeks, and the 1540
observations in category 1 activity were all devoted to producing
these units, what was the average time spent on each unit?
Solution:
 With 22 workers and 100 observations of each worker, the total
number of observations n = 100(22) = 2,200 observations.
 For the 95% confidence level, = 1.96
 σ̂ p = 0.00836
 pˆ  z / 2ˆ p = 0.19 – 1.96(0.00836) = 0.19 – 0.0164 = 0.1736

 pˆ  z / 2ˆ p = 0.19 + 1.96(0.00836) = 0.19 + 0.0164 = 0.2064

(b) Given Q = 522 work units

 p̂1 = 1540/2200 = 0.70
 Total time TT = (4 weeks)(40 hr/week)(22 workers) = 3520 hr
 Tc = 0.70(3520)/522 = 4.72 hr/work unit
Problem 16.8
A work sampling study is to be performed on an office pool consisting
of 10 persons to see how much time they spend on the telephone.
The duration of the study is to be 22 days, seven hours per day. All
calls are local. Using the phone is only one of the activities that
members of the pool accomplish. The supervisor estimates that 25%
of the time of the workers is spent on the phone. (a) At the 95%
confidence level, how many observations are required if the lower and
upper limits on the confidence interval are 0.20 and 0.30. (b)
Regardless of your answer to (a), suppose that 200 observations were
taken on each of the ten workers (2000 observations total), and
members of the office pool were using the telephone in 590 of these
observations. Construct a 95% confidence interval for the true
proportion of time on the telephone. (c) Phone records indicate that
3894 phone calls (incoming and outgoing) were made during the
observation period. Estimate the average time per phone call.
Solution:
(a) For 95% confidence interval,
zα / 2 = 1.96; and interval value c = 0.05

n = (1.96)2(0.25)(0.75)/(0.05)2 = 288.1 rounded to

289 observations

(b) Given n = 2,000 observations, p̂ = 590/2000 = 0.295

σ̂ p 0.2950.705
2000

= 0.295 – 1.96(0.0102) = 0.295 – 0.020 = 0.275

= 0.295 + 1.96(0.0102) = 0.295 + 0.020 = 0.315
(c) Given Q = 3894 calls
Total time TT = (22 days)(7 hr/day)(10 workers) = 1540
hr
Tc = 0.295(1540)/3894 = 0.1167 hr/call = 7.0 min/call
Problem 16.11
A work sampling study is to be performed on the art department in a
publishing company. The department consists of 22 artists who work at
computer graphics workstations developing line drawings based on
authors’ rough sketches. The duration of the study is 15 days, seven hours
per day. Line drawings are the main activity performed by the artists, but
not the only activity. The supervisor of the department estimates that the
proportion of time spent making line drawings is 75% of each artist’s day.
(a) At the 95% confidence level, how many observations are required if the
lower and upper confidence limits are 0.72 and 0.78, respectively. (b)
Regardless of your answer in preceding part (a), a total of 1000
observations were actually taken, and artists were making line drawings in
680 of those observations. Construct a 97.5% confidence interval for the
true proportion of time making line drawings. (c) Records for the period of
the study indicate that 5,240 line drawings were completed. Estimate the
average time per line drawing? (d) Determine the standard time for one
line drawing, given that the average performance rating for the artists was
observed to be 90%, and the allowance for personal time, fatigue, and
delays is 15%.
Solution
(a) For 95% confidence interval, = 1.96; and interval value c = 0.03
 n = (1.96)2(0.75)(0.25)/(0.03)2 = 800.3 rounded to 801 observations
(b) Actual = 680/1000 = 0.68
 For 97.5% confidence interval, = 2.24
 σ̂ = 0.0148
p

 pˆ  z / 2ˆ p = 0.68– 2.24(0.0148) = 0.68 – 0.0331 = 0.647

 pˆ  z / 2ˆ p = 0.68+ 2.24(0.0148) = 0.68 + 0.0331 = 0.713

(c) Given Q = 5,240 drawings

 Total time TT = (15 days)(7 hr/day)(22 artists) = 2,310 hr
 Average time per drawing = 0.68(2,310)/5,240 = 0.2998 hr = 17.99 min

(d) Standard time per drawing Tstd = 17.99(0.90)(1 + 0.15) = 18.62 min
 Schedule of Randomized
Observation Times

 Observation times in Work sampling study

must be randomized. Why?

- To improve the statistical accuracy of the

data.
- To reduce the bias that might be
introduced by the subjects.
The procedure for randomizing the
times
 Use any source of Random Numbers.

 Convert the random numbers to clock times

during the time period of interest.
- Certain numbers would have to be discarded for
example 2:75 “is not a time”.
- Other adjustment must be made to accommodate
the given shift

 Resequenced into chronological order

Example
 For a machine utilization example,
generate the schedule of 10 observation
times for the first day. The shift hours are
from 8:00 a.m. to noon, then from 1:00
p.m. to 5:00 p.m.
Scheduling Observation Times

Solution:
 Obtain or generate a sequence of enough random
numbers with values between 0 and 999 [021, 542, 865,
804, 023, 488, 587, 743, 570, 722, 308, 118, 431, 465, 337,
605, 229, 325].

 Let us use the following convention:

 The first digit represents the hour, and the second and third digits
represent the minutes.

 Numbers with first digits 0 and 6 are read as clock hour 10 and 11
  Numbers with first digits 5 and 7 are discarded

 Numbers with second digits 6 through 9 are discarded.

 See the following table

RN Clock RN Clock RN Clock
Time Time Time
021 10:21 587 Discarded 431 4:31
542 Discarded 743 Discarded 465 Discarded
865 Discarded 570 Discarded 337 3:37
804 8:04 722 Discarded 605 11:05
023 10:23 308 3:08 229 2:29
488 Discarded 118 1:18 325 3:25

 Morning Schedule: 8:04, 10:21, 10:23, 11:05,

 Afternoon Schedule: 1:18, 2:29, 3:08, 3:25, 3:37, 4:31

 Work Sampling Study Procedure
It is recommended that a uniform procedure should be
followed to perform a work sampling study is to

1. Establish the Purpose

First, the objective of the study should be established. Work
sampling can be used to determine an overall perspective on
the work done.

2. Identify the Subjects

Second, the people performing the task must be identified, i.e.
general office work is being studied with the objective of
determining overall productivity.
3. Identify the Measure of Output
The third step in making the study is the identification of the
measure of the output produced or the types of activities
performed on the jobs being studied. This step is especially
important if the objective of the study is to measure productivity
with the intent of setting a standard.

4. Establish a Time Period

Fourth, the time period during which the study will be conducted
must be established. Starting and stopping points for the study
must be defined as well.

5. Define the Activities

This step involves defining the activities that are performed by
the people under study. For example, the definition used in a
machine utilization study, including only the categories of
working, idle, and idle-mechanical breakdown.
6. Determine the Number of Observations Needed
After the work elements are defined, the number of observations
for the desired accuracy at the desired confidence level must be
determined. If a reasonable guess cannot be made, then a trial
study of perhaps 20 to 40 observations should be made to get an
estimate.

7. Schedule the Observations

Once the number of required observations has been determined,
either from appropriate statistical calculations or from tables, and
the actual observations must be scheduled. Typically, the analyst
will assign an equal number of observations each day during the
course of the study. For example, if 800 observations are required
and 20 work days are established as an appropriate observation
time, 40 observations should be recorded each day.

A random number table can be used to establish the random

times for each observation.
8. Inform the Personnel Involved
Before the study is actually performed, the personnel
involved should be informed about the objective of the
study and the methodology that will be employed.

9. Record the Raw Data

The next is the actual recording of the raw data. Although
this recording can be performed by anyone, it is desirable
that a trained analyst be employed.

It is also very important that the observations be made at

exactly the same location every time.

10. Summarize the Data

After the data have been collected, they must be
summarized.
Work Sampling Observation Form
Advantages of Work Sampling

 Can be used to measure activities that are impractical to

measure by direct observation

 Multiple subjects can be included

 Requires less time and lower cost than continuous direct

observation

 Training requirements less than DTS or PMTS

 Less tiresome and tedious on observer than continuous

observation

 Being a subject in work sampling is less demanding than

being watched continuously for a long time
Disadvantages and Limitations
 Not as accurate for setting time standards as other work
measurement techniques

 Usually not practical to study a single subject

 Work sampling provides less detailed information about work

elements than DTS or PMTS

 Workers may be suspicious because they do not understand the

statistical basis of work sampling

 Behavior of subjects may be influenced by the act of observing

them