Work Sampling

Sections:
1. How Work Sampling (WS) Works
Chapter 16 2. Statistical Basis of Work Sampling
3. Use of Work Sampling to Measure
Average Task and Standard Times
4. Defining the Activity Categories in WS
5. Scheduling Observations

1
 Work Sampling (WS) Defined
 Work Sampling is statistical technique for determining the
proportions of time spent by subjects (e.g., workers, machines) in
various defined categories of activity (setting up a machine,
producing parts, idle, etc).

 Large number of observations are made over an extended

period of time.

 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

2
When is Work Sampling Appropriate?
 Sufficient time should be available to perform the
study
 Several weeks 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

3
Example: How Work Sampling Works
 A total of 500 observations taken at random times
during a one-week period (40 hours) on subjects of 10
machines with results shown below. In each observation
the machine subject was identified into one of three
categories:

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

spend in each category?

4
 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

5
 Work Sampling Applications
 Machine utilization - how much time is spent by
machines in various categories of activity
 Previous example

 Worker utilization - how workers spend their time

 Average unit time - determining the average time on each

work unit

 Time standards - limited statistical accuracy when

standards set by work sampling

6
 Statistical Basis of Work Sampling
 Binomial distribution, in which parameter p = true
proportion of time spent in a given category of activity

 There are usually multiple activity categories, so we

have p1, p2, . . , pk, . ., pK proportions for K different
activity categories

 The binomial distribution can be approximated by the

normal distribution, where mean and standard deviation

 = np
= np1  p 

7
 Alternative Parameters
 The parameters  and  can be converted
back to proportions by dividing by the number
of observations n

μ np p1  p 
p= = p 
n n n

 = mean of the proportion p

 p = standard deviation of the proportion p
n = total number of observations
p = the proportion of the total number of observations (time spent) devoted to an
activity category of interest

8
 Confidence Intervals
 Our aim is to estimate p within a defined error range at a
confidence level

 The statement of a confidence interval p̂ relative to p can

be expressed as follows

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

 The probability that the actual p lies within p-z*sigma and p+z*sigma is (1-
alpha) 9
 Example: Confidence Intervals in
Work Sampling
 Determine the 95% confidence interval for the proportion
of time spent setting up the machines, category (1) in
previous example (see: slide 4)

Solution:

10
 Example: Determining the Number of
Observations Required
 Determine how many observations required to estimate
the proportion of time used to set up 10 machines. The
confidence interval must be within ±0.03 of the true
proportion, which the foreman initially estimates to be
0.20. A 95% confidence level is to be used.

Solution:

Zα/2 = 1.96, c = 0.03, = 0.20

n = (1.96)2 × 0.20 × (1 – 0.20) / (0.03)2
= 683

11
 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

where
Tci = average task time,
pi = proportion of observations associated with category i,
TT = total time for the work sampling study,
Qi = total quantity associated with category i

12
 Example: Determining average task times
 Consider the example in slide 4. A total of 1572 units were
completed by the 10 machines and that a total of 23 setups
were accomplished during the 5-day period.
Determine (a) the average task time per work unit during
production (b) the average setup time.

 Remember that proportion of observations associated with

category “running production” was found as 0.6. For “being set
up” it was 0.15.

 Solution:
TT = 40 hr (10 machines) = 400 hr
(a) Tproduction = 0.60(400)/1572 = 0.1527 hr/unit = 9.16 min/unit
(b) Tsetup= 0.15(400)/23 = 2.609 hr/setup = 156.5 min/setup

13
 Determining Standard Times
 When the purpose of the work sampling study is to set time
standards, the analyst must rate the performance of the worker
during each observation

 First determine normal time for activity i

pi (TT ) pi (TT )(PR i )
Tci = Tni =
Qi Qi
where
Tni = normal time for work unit associated with activity i
PRi = average value of the performance ratings for all
observations in category i,

Then determine standard time

Tstdi = Tni(1 + Apfd)

14
 Example

 In the welding department a work sampling study was conducted. Only

two activity categories were considered: (1) welding and other productive
work, and (2) personal time, rest breaks, and delays for the workers. Over
a four-week period (40 hours/week), 125 observations were made at
random times. Each observation captured the category of activity of each
of eight welders in the shop section of interest and the performance rating
of the welders was 98%, 105%, 95%, 110%, 90%, 95%, 85%, 89%,
respectively. Results indicated that category 2 constituted 33% of the total
observations and the allowance factor Apfd is 15%.
(a) Define the limits of a 96% confidence interval for activity 2.
(b) If a total of 725 work units were produced during the 4 weeks, and all
category 1 activity was devoted to producing these units, what was the
average time spent on each unit?
(c) Determine the standard time for the category 1 activity.

15
 Example
Solution:
With 8 welders and 125 observations of each welder, the total number
of observations n = 125(8) = 1000 observations.
(a) For the 96% confidence level, Zα/2 = 2.05.
p1  p  0.331  0.33
p   = 0.01487
n 1000
= 0.33 – 2.05(0.01487) = 0.33 – 0.0305 = 0.2995
= 0.33 + 2.05(0.01487) = 0.33 + 0.0305 = 0.3605
(b) Given Q = 725 work units
P1 = 1 – 0.33 = 0.67
Total time TT = (4 weeks) × (40 hr/week) × (8 welders) = 1280 hr

= 0.67 × (1280)/725 = 1.183 hr/work unit

(c) PRi = (98+105+95+….+ 89)/8 = 96% Tn = 1.183×0.96 = 1.136 hr/unit

Tstd = Tn (1 + Apfd) = 1.136(1+ 0.15) = 1.306 hr/unit
16
 Defining the Activity Categories
Some guidelines:
 Must be defined to be consistent with objectives of study

 Must be immediately recognizable by observer (each

category must be distinguishable from the others)

 If output measures are included, then activity categories

must correlate with those measures

 Helpful to limit the number of categories to ten or fewer

(fewer categories will result in higher average which will
reduce variance and improve confidence interval).

17
 Defining the Activity Categories (Cont.)

 Example in machine utilization: is it sufficient to use single category

“ machine idle”

OR
Will it be more informative to know the reason of machine being idle by
including categories such as:
“downtime due to breakdown”
“waiting for work” and
“no schedule work”

18
 Defining the Activity Categories (Cont.)

 Two examples of possible categories for

possible types of work (for worker): First
example (see table 16.2, page 433)

19
 Defining the Activity Categories (Cont.)

 Two examples of possible categories for

possible types of work (for worker): second
example (see table 16.3, page 434)

20
Work Sampling Observation Form

21
 Scheduling Observations
 Preparing a schedule of randomized observations
 Improve the statistical accuracy
 Reduce bias

 Sampling stratification: Total number of observations is

divided into a specified number of time periods (e.g., days,
half-days and hours) so that there are an equal number of
samples taken each period
 Observation times in period are randomized
 Reduces the variance

22
 Scheduling Observations (Cont.)

 Example: Total observations 100: the duration of the work sampling

study is two weeks

so, 100/2= 50 round per week vs 50/5 =10 rounds per day

 Total number of observations is divided into a specified number of time

periods (e.g., days, half-days and hours)

 Due to the randomization, there would not be exactly 10 rounds/day.

Some days would have more, other days have fewer.

23
 Example: Generation of random
observation times
 For the machine utilization example, generate the schedule of 10
observation times for the first day. The shift hours are 8:00 a.m.
to noon, then 1:00 p.m. to 5:00 p.m.

 Solution: Generate a set of three digit random numbers between 1

and 999 (using a pseudo random number generators in Excel).
Conversion of numbers to clock times

 Numbers with first digits= 8,9,1,2,3 and 4 are read directly as

the clock hour

 Numbers with first digits= 0 and 6 are read as clock hours 10

and 11, respectively

 Numbers with first digits= 5 and 7 are discarded

 Numbers with second digits 6 through 9 are discarded

24
Example: Generation of random
observation times (Cont.)

25