Work Sampling

Various IE (Industrial Engineering) techniques are utilized in

productivity improvement activities. “Work Sampling” is one of the
IE techniques and it is the economical method based on the concept
of probability.
In this lecture, including the theoretical probability, there may be
some parts that are a little hard to understand. However, it is not
necessary to think of them deeply.
The important thing is how to utilize Work Sampling in the
productivity improvement. Learn the basics of Work Sampling, and
think about how to utilize it in productivity improvement activities.

Written by Nobuo Kanai and Naoya Nishigaki

1
 Table of Contents
1. Introduction: Work Sampling method is one of the IE techniques
2. The system of IE (in a narrow sense)
3. Ratio-delay Study
3-1 The main purpose of Ratio-delay Study
3-2 Kinds of Ratio-delay Study
4. Step of IE for work improvements
5. The outline of Work Sampling
5-1 What is work sampling
5-2 How is work sampling used?
5-3 How is work sampling performed?
6. Feature of Work Sampling
6-1 Features of Work Sampling
6-2 The important notices when applying Work Sampling
6-3 The comparison of advantages or disadvantages with other methods
7. Recording sheet to be used in Work Sampling
8. Determination of observation hours (use of “ Random Time Table”)
8-1 Random Time Table
9. Decision of observation numbers
9-1 The grounds for determining the number of observations
9-2 Meaning of 「 Accuracy: 8% 」 , 「 Confidence level: 90% 」
9-3 Two kinds of accuracy
9-4 Calculation formula of the necessary number of observations “N”
10. Confirmation of the results of Work Sampling
10-1 Confirmation items before analyzing the results of Work Sampling
10-2 Examination of the accuracy of the operating rate by the work sampling
11. The viewpoint of the results of Work Sampling
12. The usage of Work Sampling
13. Case of implementing Work Sampling
13-1 The procedure to perform Work Sampling in this case
13-2 The cases of Work Sampling 2
1. Introduction:
Work Sampling method is one of the IE techniques.

(1)What is IE (Industrial Engineering)

The branch of engineering that is concerned with the efficient production of industrial
goods as affected by elements such as plant and procedural design,
management of materials and energy, and the integration of workers in the overall
system.

(2)The purpose of IE

1) The improvement of work methods

(Discovery of the most efficient method. Elimination of “Muda”, “Mura” and “Muri”)
2) Standardization(Work methods, Materials, Facilities, Tools, Environment)
3) Setting of the standard time
4) Instruction and maintenance
(Instruction and the entrenchment of new work methods)

Note: “Muda” means Waste “Mura” means Fluctuation “Muri” means Over burden

3
2. The system of IE (in a narrow sense)

(Process chart analysis

Process
・ Flow analysis)
analysis
(Therblig analysis)
Methods
study
Motion
study
IE (Work study)

Time (Standard
study time)
Work
measurement
Ratio-delay 1) Continuous reading method
study 2) Snap reading method (Work
Sampling)

4
3. Ratio-delay Study
3-1 The main purpose of Ratio-delay Study
- Investigation of causes of reducing operating rate
- Improvement to enhance the rate of utilization.

In addition, Ratio-delay Study is used

for setting the standard time or grasping Operation or Motion in a Factory
work loads. Miscellaneous Education
operation Meeting

An example of the element Repair work

Tool
in Ratio-delay Study Exchange
Waiting for
Parts
1) Machine stop due to the trouble
2) Analysis of various “Waiting” MUDA Work
(Waiting for materials, Waiting for parts, Repair of
Waiting for jigs and tools etc.) Machine Physiological
Trouble Reasons
3) Work stop by the physiological reason ( toilet etc.)
4) Machine stop or work stop for set-up Waiting for
material Transportation
5) Machine stop for tool exchange of Jigs for Set-up

5
3-2 Kinds of Ratio-delay Study

Ratio-delay study is the technique to investigate the “ratio of time composition“ of a

working rate of a worker or the machinery and operation contents.

There are two types of Ratio-delay Study method. Continuous reading method

(1)Continuous reading method:

The method to observe and analyze the operation status of a worker
or machinery in succession all day long

(2)Snap reading method ( Work Sampling )

The method to observe and analyze the operation status of a worker Work Sampling method
or machinery by momentarily observing the frequency for a long
time according to the classification of an action.
For example, When observing the operation situation and checking it
30 times a day, if a worker works 27 of 30 times, calculated as
27÷30=0.9. The rate of utilization is 90%.

6
4. Step of IE for work improvements
An example of the approach of the work
improvement using the IE technique Discovery of problems
9) Discovery of bottle neck processes and analyzing
1) Determination of objects
10) Finding and analyzing “Muda / Muri” works
Understanding of current state or processes
2) PQ analysis
Improvement
3) Process chart analysis
11) Improvement of work methods (Elimination
3) Work Sampling of Muda) ・ Process change etc.

4) Work flow analysis Standardization

6) Line balance analysis 12) ・ Standardization of improved methods

・ Setting of standard works or processing
7) Material handling analysis time

It may not always be necessary to follow this order. Its

8) Therblig analysis
principal objective is to grasp current states or
problems, and link to the elimination of “Muda, Muri and
Mura.” 7
 The source of the item No.5 :
5. The outline of Work Sampling World academy online

5-1 What is work sampling?

Work sampling is a work measurement methodology that estimates the proportion of time
an employee utilizes in performing assigned jobs/tasks.
The methodology uses random observations of actual worker’s activity and is dependent on
the laws of probability.

Since it does not require a formalized time study procedure conducted by qualified
stopwatch analysts, it is less costly.
The methodology requires that the observer simply determine whether an employee is
actually working or is idle during any particular observation.

After all the observations are completed, the percentage of working time is computed from
the total number of observations.
The greater the number of observations, the more accurate the technique is.

The calculation of operating rate in Work Sampling

Operating rate=(Sample number in actually working ÷Total sample number)X100
8
5-2 How is work sampling used?

Work sampling is used for the following:

1. Ratio-delay studies.
Worker’s allowances are determined by calculating the percentage of time
an employee spends in unavoidable delays.

2. Percentage of utilization of equipment.

The technique is used to determine the actual utilization of machinery and
other equipment.

3. Determining labor standards.

The technique is useful in determining work standards for various tasks by
rating the employee’s performance.

4. Evaluating an employee’s performance.

A performance standard can be calculated utilizing the work sampling
procedure and resulting standards.

9
5-3 How is work sampling performed?

(1) Sample observations.

Several sample observations are performed as the basis for developing a correct sample size
based on problem parameters.

(2) Compute the actual sample size.

The sample size is dependent on the desired level of statistical confidence and accuracy.
Normally, the acceptable level of confidence is 95% with the accuracy level of ± 5%.
The following formula determines the actual sample size necessary for a work
sampling procedure:

The necessary sample size (in the case of 95% confidence)

N=4 p (1-p) / e2

p=estimation of the number of times that are spent in an activity

1-p=estimation of the number of times that are not spent in an activity
e=absolute accuracy

Higher confidence level or higher accuracy level will increase the required sample size.
The 95% confidence level and ± 5% accuracy level mean that the results of 95 times of work
sampling executed 100 times will be within ±5% of accuracy error margin.

10
Cont… 5-3 How is work sampling performed?
(3) Prepare a random schedule for work observations.
(4) Observe and rate the employee’s work performance.
(5) Totalize the number of produced units and calculate the average time per unit.
(6) Compute the standard time per unit.

Reference: meaning of “absolute accuracy”

2σ area of the normal distribution curve is 95.45%. Therefore,
the size of a sample satisfying 95% of confidence is 2 σ area.
If the confidence for the whole sample is decided, it is
necessary to decide the tolerance of the sample.
Formula as 95% of confidence,
tolerance of sample: “e”,
e=2 P(1-P) incidence of phenomenon: ”P”
√ N sample size: “N”
( just memorize this formula)

The formula shows that a mean value of P obtained from the

sample size N is in the range of ± e from a true value as a
center with 95% of confidence.

It means that the mean value of P is somewhere in P ± e. An area (probability) in the

The tolerance “e” is called “ absolute accuracy” normal distribution curve
11
6. Features of Work Sampling
6-1 Features of Work Sampling
1) It is possible to observe a large number of working conditions alone.
2) Observation is easy.
3) It is possible to make an investigation within a short period.
4) The observation of almost actual situations is possible, because the observed
operator/worker is rarely aware of being observed.

6-2 The important notices when applying Work Sampling

1) Grasp the purpose of the analysis enough.

2) How fine has to grasp the operating rate?
3) How precise has to grasp the rate of operation?
4) Decide time with a random schedule so that the observer dose not enter
the selfish judgment to observation hours.
5) Investigate “the work state , the management state and the tendency“ of
the target operation beforehand.

12
 6-3 The comparison of advantages or disadvantages with other methods.

Type Advantage Disadvantage

1) Continuous Correct time values can be ・ An observer can observe only
reading method obtained. one object.
* Fatigue of an observer or a worker occurs.
2) Snap reading * Many objects can be observed * The qualitative analysis of occurred
method (Work by one observer. contents is not possible.
Sampling) * The fatigue of an observer or a * Grasping of qualitative occurrence
worker is little. frequency or the time required for such
occurrences is not possible.
3) VTR method * The work situation can be * The area of observations is limited.
(One of the replicated * No information about various work
means of * The analysis of working hours environment can be provided.
continuous and work contents is possible (noise, illumination, temperature, humidity
reading)
at the same time. etc.)
* Analysis of minute movements
is possible.

13
 Work Sampling Observation Sheet Name Approved Checked Made by

7. Recording Operation Name Assenbly

Taro YAMADA Y.KOYAMA G.MURATA Y.INUI

sheet to be
Operator 8 persons(aaa,bbb, ・・・・・・・・・・,hhh) ○○○Division ××Dept. △△Factory
Equipment □□□Section ＊＊＊Group

used in Work
Classification Main Work Preparation Reserve time Non-operating

Preparation of materials

Transportaion of parts
Tightening of screws

Arrangement of parts

Preparation of Jigs
Assembling of part

Handling of parts

Repair of defects
Insertion of part
Items

Transportaion
Sampling

Soldering

Meeting

Waiting
Total

Others

Others

Others

Others
Toilet

Chat

Rest
Time

1 8: 23 / / // / / / / 8
2 30 // / / / / / / 8
3 34 // // / / / / 8

Sample 4
5
37
58
/
///
/
/
/ ///
// /
/ /
/
8
8

format of 6
7
9: 27
36
//
/
/
/
/
/
/
//
/
// /
// 8
8

recording 8
9
10: 35
11: 16
// /
/ / //
///
//
/ /
/ /
8
8

sheet-1 10
11
20
35
/
//
/
/ //
///
//
/ /
/
/ 8
8
12 13: 15 // / / / / / / 8
13 29 // / / / // / 8
14 41 / // // / / / 8
15 55 /// / / /// 8
16 14: 32 / / //// / 8
17 15: 05 // // // / / 8
18 45 // / / /// / 8
19 16: 02 / / // /// / 8
20 34 // / // // / 8

Total
33 17 9 17 45 2 3 4 4 2 4 2 3 2 4 1 2 2 3 1 160
of each item
% of each item 20.6 10.6 5.6 10.6 28.0 1.3 1.9 2.5 2.5 1.3 2.5 1.3 1.9 1.3 2.5 0.6 1.3 1.3 1.9 0.6 100
Source: Total of each
123 13 18 6 160
IE text in work place classification
By Katsuyoshi ISHIHARA % of each
76.7 8.1 11.4 3.8 100
classification
(Issued by JUSE) 14
 Operation analysis totaling sheet
Cont…7. Recording sheet Name of object Taro
Yamada
Approved Checked Made by

to be used in Work Sampling Observation period

Remarks
26 Feb.~2 Msar.(5days)
○○Division ×××Dept.
△△section ＊＊＊Group

A llo w a n ce o f m a n p o w e r
A llo w a n ce in w ork g ro u p
P re p a ra tio n o f m ate ria ls
Classification Main operation Preparation Allowance

A llo w a n ce o f o p e ra tio n
Sample format of

A rra n g e m e n t o f p a rt
T ig h te n in g o f scre w

A sse m b lin g o f p a rt

H a n dlin g o f p a rts
In se rtio n o f p a rt

N o n -o p e ra tin g
Items
recording sheet-2

S o ld e rin g

O th e rs
Date
or
Time
26 Feb.1974(Mon) 33 17 9 17 47 7 4 2 9 6 3 6

27 Feb.1974(Tue) 35 16 11 14 53 6 5 1 7 8 2 2

28 Feb.1974(Wed) 33 15 13 19 48 9 2 2 11 5 0 3

1 Mar. 1974(Thu) 31 19 8 15 54 5 5 0 6 9 4 4

2 Mar.1974(Fri) 38 17 11 21 51 3 1 0 10 5 2 1

Total of each item 170 84 52 86 253 30 17 5 43 33 11 16

% of each item 21.2 10.5 65 10.8 31.6 3.8 2.1 0.6 5.4 4.1 1.4 2
Total of each
645 52 8.7 16
classification
% of each
80.6 6.5 10.9 2.0
classification
Source:
Problem,Conclusion, Proposal etc.
IE text in work place
By Katsuyoshi ISHIHARA
(Issued by JUSE) 15
 Work Sampling Totaling Sheet
Approved Checked Made by

Cont…7. Recording sheet Process name

Operation Name
AV assembly process
Terminal caulking work
Name:
Taro YAMADA

to be used in Work Sampling Part, Product

Facility (Worker)
Terminal box
N.1○○No.2××No.3□□ Production Section
△△Division ▽▽Department
Sub-ass'y Group
Clarification Operating Set-up Non-operating

P reperation of m aterial
Sample format of

M achine adjustm ent

E xchange of punch

S orting of Products

A bsence of w orker
M achine trouble
part assem bling
Items
recording sheet-3

C aulking

O thers

O thers
Total

Worker
Facility

1. Caulking ///// ///// //// // / // / ///

machine / /////
30
No.1 /

2. Caulking //// ///// /// / / /// // // /

machine /////
30
No.2 ///

3. Caulking //// ///// //// / / // / ////

Machine / /////
30
No.3 /

Total of each item 16 35 11 4 2 1 7 4 9 1 90

% of each item 17.7 38.9 12.2 4.5 2.2 1.1 7.8 4.5 10.0 1.1 100%
Total of each
classification
% of each
classifiction
No. 1 2 3 4 5 6 7 8 9 10
time 8:41 9:00 9:15 9:57 10:11 10:16 10:28 10:30 10:47 10:52
Source: No. 11 12 13 14 15 16 17 18 19 20
IE text in work place Observation time 10:54 11:05 11:08 11:22 11:35 11:54 12:55 1:06 1:27 1:58
time No. 21 22 23 24 25 26 27 28 29 30
By Katsuyoshi ISHIHARA time 2:19 2:24 2:57 3:01 3:06 3:20 3:48 4:02 4:11 4:31
(Issued by JUSE) No.
time
31 32 33 34 35 36 37 38 39 40 16
8. Determination of observation hours (use of “ Random Time Table”)

It is necessary in the “Work Sampling method” to determine observation hours at random.

Random hours are usually determined based on a “Random Time Table”.
There are various kinds of “ Random Time Table”. One of the tables is shown on the next slide.
As for this “Random Time Table,” 40 times of “hour” are written by unit of one minute for about
ten hours from 0:00 to 9:59.

As for two sets of numerical values in two columns, the sequence No.(1-40) is written on the
left column and the hour is written in the right column.

8-1 Random Time Table

(A part of the Random Time Table)

No. Hour No. Hour No. Hour No. Hour No. Hour

17
 Random Time table(1) 40 times for 10 hours Random Time table(2) 40 times for 10 hours

No. Hour No. Hour No. Hour No. Hour No. Hour No. Hour No. Hour No. Hour No. Hour No. Hour

18
Source:“Basics of IE”by Akihisa FUJITA(published by Kogakusha)
Cont….. 8-1 Random Time Table
An example:
The number of days for observation: 5 days
Period of one observation: 8:00~17:45(rest time 12:00~12:45)

Procedure 1: Choose the column number for a particular day of observation.

In this example, the column No. 3 is used on the 1st day, the column No.2 is used on
the 2nd day and the column No.8 is used on the third day.

No. hour No. hour No. hour

Each random observation hour is shown in a sub-column depending on

the number of observations.
Choose only the necessary number of times of hour.
In the case of 20 times of observation, choose the numbers from 1 to 20. 19
 Cont…8-1 Random Time Table The hours for 20
Procedure 2: No times of observation
Hour
.
Add the observation start hour to each hour of the time table. 8.25
8.35
8.52
9.13
Column No. 9.23
8.25
9.30
No Hour 9.34 8.35
. Add all 8 at the hour of
9.37
9.58 8.52
9.23
the time table, because 10.27
10.36 9.30
the observation start 10.48
9.34
11.02
hour is 8:00, 11.15 10.27
11.35
12.03 10.36
Rest 12.16 11.02
12.20
11.35
Procedure 3: time 12.35
12.42 13.02
Exclude the time of non-observation. 13.02
13.07 13.29
13.21
14.15
The rest time, from12:00 to 12:45, cannot be observed. 13.29
14.29
13.41
Delete such hours from a time table. 14.15 14.41
14.29
For example, delete hours such as 12.03,12.16,12.20,12.35 14.35
14.55
15.32
and 12.42 with a side line. 14.41
14.55 16.05
15.32
Procedure 4: 15.49 17.02
16.05 17.34
Choose only the No. that fit the number of observation times from the 16.25
16.45
No. line. 16.57
17.02
In this example, the hours for 20 times of observation were chosen. 17.21
11 8 2 16 4 15 9 19 22 1 7 21 13 12 5 6 14 18 10 3 17.30
17.34 20
 35 Additional 5
Cont…. 8-1 Random Time Table observations
No. Hour observations No. Hour
8.05
Notes on determining the random hours 8.25
8.35
8.17
8.28
8.52
8.42
1) The column using the random time table 9.13
9.23
8.43
9.01
should be changed at random every day. 9.30
9.34
9.20
9.35
9.37
9.49
9.58
9.59
10.27
2) When the observation time of the day is 10.36
10.08
10.24
10.48
over 10 hours , for example 16 hours, 11.02
10.23
10.36
choose two rows at random and use it in 11.15
11.35
10.45
11.20
combination. 12.03
12.16 Rest 11.29
12.20 11.32
12.35 time 11.45
12.42 12.39 Rest
3) When observing 40 times or more in a day, 13.02
13.07
12.57
13.12
time
choose two rows at random. 13.21
13.29
13.52
14.02
At first, determine the time with one column. 13.41 14.34
14.15 14.39
And make up for shorten times from the 14.29 14.41
14.35 14.49
second column. 14.41 15.13
When the hour is overlapped, add another No. 14.55
15.32
15.50
16.06
15.49 16.25
16.05 04.32
16.25 16.49
16.45 16.52
16.57 17.06
17.02 17.09
17.21 17.33
17.30 17.45
17.34 17.53 21
9. Determination of observation numbers
9-1 The grounds for determining the number of observations
1) Acceptable levels of difference between obtained results and true values(accuracy)
2) How can the results be trusted?
9-2 Meaning of “Accuracy: 8%”, “Confidence level: 90%”:
When the real operating rate is 80%, the operating rate is inferred between 72%(=80-8) and
88%(=80+8) against 80%.However, it must not be deviated more than ten times when analysis is
carried out 100 times
Necessary accuracy or confident level are determined depending on purposes and situations.
9-3 Two kinds of accuracy: 9-4 Calculation formula of the necessary
1) Absolute accuracy ”e” number of observations “N”
For example, the accuracy of 2% means “the real operating rate
is within 70% ±2%”. This is called the absolute accuracy.
e=2 P(1-P)
√ N
2) Relative accuracy “s”
The difference between analysis result and true value is shown by ratio.
４ P(1-P)
For example, in the case of operating rate is 80% , N=
e2
when absolute accuracy is 8% → Relative accuracy=8/80=0.1 (10%)
e: absolute accuracy
Generally, in the case of work sampling, the relative accuracy is set to N: number of observation
5~10%. When the purpose of analysis is to grasp problems or simply to P: operating rate by
make investigations, around 10-30% of relative accuracy is reasonable preliminary investigation
22
10. Confirmation of the results of Work Sampling
10-1 Confirmation items before analyzing the results of Work Sampling

1) Confidence level of the results of Work Sampling

2) Was Work Sampling implemented appropriately?

When it is felt through daily observations or experiences that there are some doubts about the
results of operating rates, restudy whether or not the way of Work Sampling had problems.
Check the obtained rate of utilization from the following standpoint:

1) There is a large gap between the obtained operating rate and estimated operating
rate from the results of the production volume or the observation on a daily basis.
2) There is a large difference in the operating rate from day to day, and the cause of the
difference cannot be grasped clearly.
3) There is a difference in the operating rate among each of the facilities or workers that are
observed, and the cause of the difference cannot be grasped clearly.

When it was confirmed that Work Sampling was carried out appropriately, confirm the
accuracy and the confidence level of the operating rate by the Work Sampling .
23
10-2 Examination of the accuracy of the operating rate by the work sampling
For example, at first, the observation number “N” is calculated with the accuracy of 3%,
the confidence level of 95% concerning the operating rate of approximately 80%,
4X0.8X(1-0.8)
N= = 712
0.032

Observation Number N=712 was applied to Work Sampling,

→observed operating rate became 70%.
Calculation of the absolute accuracy as the confidence level of 95%.
The accuracy fell
e=2 0.7X( １ -0.7) ＝ 0.034 to 3.4%.
712

When the sustenance of the accuracy: 3% is expected strictly:

Recalculation of necessary number of observations
furthermore, it is necessary to take
N==9394X0.7X(1-0.7)
2
0.03 222 data (= 939-712).

Confirm the validity of the way of the Work Sampling in this way.
And watch the results from various viewpoints about operating rate. 24
11. The viewpoint of the results of Work Sampling
1) Check the gap to the standard operating rate.
2) Check the content of non-operation by Pareto analysis and improve non-operation items.
3) Check the change on a day or time.
4) Check the difference of operating rate between workers or facilities

12. The usage of Work Sampling

1) Grasping the cause of low operating rate and improving it.
2) Planning the adequacy of work loads
3) Studying the handling number of facilities
4) Planning the standardization of indirect works
5) Planning the improvement of facilities
6) Finding out the allowable rate for setting of standard time
7) Obtaining basic data for setting of standard time or pricing of products
Working hour per one piece:
＝ operation time÷production volume ＝ observation period(hour)÷production volumeXoperating rate
As a result of observation for 5 days with 8 hours a day, when the operating rate is 80%,
When the production volume is 2000 pieces for a observation period (hours)
Working hour per one piece=[(8X5X60)/2000]X0.8 ＝ 0.96 (minutes)
Calculation of true standard time:
Standard time
＝ (observed total operation time÷production volume)Xrate of main workXratingX(1 ＋ allowed rate)

25
13. Case: Implementation of Work Sampling
13-1 The procedure to perform Work Sampling in this case

Procedure 1 : 1) make the purpose of the study clear

Definition of problem(s) 2) Clarify the purpose of the item which
is going to be measured.

Procedure 2:
Get the approval by the manager or
the foreman of the target section

Procedure 3:
Determine reliability, accuracy to be
expected for results.

It is desirable to conduct the preliminary

Procedure 4: sampling for about one day.
Estimate the incidence of phenomenon The samples for the preliminary observation
by the preliminary sampling may be included in the number of samples of
the main observation.

26
 Cont… 13-1 The procedure to perform Work Sampling in this case
Procedure 5: 1) Determine the number of the necessary samples.
Make an implementation plan. 2) Determine the sampling period and the sample number per day.
3) Determine a daily interval of rounds depending on the number
of workers or the number of machines for observation.
4) Choose the observation hour at random.
5) Prepare a recording sheet for observation.

Procedure 6: 1) Record observed data.

Conduct sampling according 2) Aggregate data at the end of each observation day.
to the plan 3) Determine a management limit, and write a control chart to
exclude an abnormal value. ( ３ σ control limit)
4) When it is believed that an stable result was provided
approximately, check whether the result enter a allowable
level, and stop the sampling if the result is good.

Procedure 7: 1) Summarize the result.

Get a result 2) Suggest the improvement.

27
 Source: “IE techniques in a workplace”
by Mr. Kinichi IKENAGA (published by JUSE)
13-2 The cases of Work Sampling
(1) Example 1: The operating rate of trucks for transportation of products
A request for increasing the number of the truck for transportation was delivered to the general
manager by the manager of the shipment section.
Procedure 1: The definition of problem(s)

When the operating rate of the trucks is high, the proposal is appropriate.
When the operating rate is low, an action is necessary to find the causes of low rate of utilization.
It should be decided whether trucks should be increased or not, after verifying the rate of non-
utilization that cannot be avoided.
Therefore the general manager requested the IE section to conduct Work Sampling in order
to grasp the actual situation of the non-operating rate of the trucks.

Procedure 2:Get the approval by the manager or the foreman of the target section.

The staff of the IE section explained the purpose of the investigation to the manager and the
foreman of the shipment section. The shipping section has had no reliable information about
the operating rate of the trucks till now.
Therefore the manager and the foreman cooperated with this investigation proactively.

28
 Cont…13-2 The cases of Work Sampling

Procedure 3: Determine the confidence and accuracy to be expected for the result.

Confidence level was set to 95% (95 of 100 times of probability) that is the most common
practice to work sampling. And the absolute accuracy was set to ±3%.

Procedure 4: Estimate the incidence of phenomenon by the preliminary sampling.

The preliminary sampling was carried out for estimating the non-operating rate P and to know
the approximate causes of non-operating conditions.
→ the observation was made by going around outskirts of a garage and a shipment yard ten times.
As for 14 trucks, 140 samples were provided by 10 observations.
Judgment standard:
・ The absent truck which went out for shipment→ “operating”
・ The truck which is in the garage→ “non-operating”
・ The absent truck for which the shipment instruction is not made →“non-operating”

Observation result→54 “non-operating“ trucks in total

Non-operating rate→54/140 ＝ 38.6 ％

29
 Cont…13-2 The cases of Work Sampling

Procedure 5: Make an implementation plan.

The number of necessary samples was calculated as follows:

P=0.386, e=0.03, N=?

e=2 P(1-P)
N=4P(1-P) / e2 =4X0.386(1-0.386)/0.032=1,050 √ N
The necessary sampling size is 1,050 from this calculation.
140 samplings have been already taken in the preliminary sampling. ４ P(1-P)
Therefore, 910 samples should be newly taken. N=
e2

Sampling methods for 910 samples:

1) 14 trucks→65 rounds are necessary to take 910 samples.(910/14=65)

2) the completion of an investigation is expected in approximately ten days.
→ the average frequency of daily visits is 6 times/truck (65/10≒6).
3) Therefore, 84 samples a day should be taken.(14X6 ＝ 84)

The implementation plan was made as above.

30
 Cont…13-2 The cases of Work Sampling
Procedure 6: Make sampling according to the plan
Sampling were performed 1064 times for two weeks.
The result of the sampling is shown in the following table.
15.5% of data on October 5 was excluded from the ample data as an abnormal value
because it was too different from other data. Therefore a new sample of 15 October was added .

The non-operating rate of the truck

Total Sample The sample of

Observation date Rate % Remarks
Number non-operation
1 Oct.（Mon) 140 54 38.6 The data on 5 Oct.
2 Oct.(Tue) 84 35 41.7 was excluded from
3 Oct(Wed) 84 24 28.6 the sample data
4 Oct(Thu) 84 36 42.9
The reason that is judged
5 Oct(Fri) 84 13 15.5 abnormal
to be abnormal is shown
6 Oct(Sat) 84 36 42.9
on the following slide
8 Oct(Mon) 84 37 44
9 Oct(Tue) 84 26 31
10 Oct(Wed) 84 26 31
11 Oct(Thu) 84 24 28.6 The data on 15 Oct.
12 Oct(Fri) 84 33 39.3 was added instead
13 Oct(Sat) 84 27 32.1 of the data on 5 Oct.
15 Oct(Mon) 84 33 39.3 added

31
Average of the non-operating rate=36.7%
Cont…13-2 The cases of Work Sampling
The inspection of the abnormal value and check of validity of the non-operating rate
The 3 σ limit was applied to inspect the accuracy of the data on 5 October.
The range of 3 σ in the normal distribution is 99.73% of the entire samples. Therefore, there is
extremely the small probability of deviation from the 3σ limit. Therefore, it should be considered
that some abnormality occurred when the sampled data go out of the 3 σ limit.
The calculation formula for the 3 σ limit(UCL & LCL) for mean P of the non-operating rate

UCL= P+3 P(1-P) LCL= P-3 P(1-P) P=mean % of the non-operating rate
√ n √ n n=mean number of samples in a day
The calculation about the truck
P=(404÷1.148)X100 ＝ 0.35X100 ＝ 35.2 ％
n ＝ 1.148 samples÷13 days ＝ 88 samples per day
0.35(1-0.35)
Control limit of P ＝0.35±3 √ ＝ 0.35±0.15
88
Upper control limit(UCL)=0.35+0.15=0.50=50% The datum (15.5%) of 5 Oct. was excluded,
Lower control limit(LCL)=0.35-0.15=0.20=20% because it was out of the UCL.

Mean non-operating rate that is calculated excluding the data on 5 October

P=(391÷1.064)X100=0.367X100=36.7 ％
Check whether the absolute accuracy enters the tolerance range or not
e=2 P(1-P) = 2 0.367(1-0.367) =0.0296 Absolute accuracy e=0.0296
√ N √ 1064 Less than ±3% that is judged first 32
Cont…13-2 The cases of Work Sampling
(2) Example 2: The average time of the assembly work
The case of the assembly work in a certain factory:
As for their products, a component and the structure are almost the same. However, their
appearances and dimensions are slightly different. The time standard was not clearly fixed
so far, therefore some problems always occurred in the manpower planning till now.
Therefore work sampling was performed to know the average assembling time per one unit in a
short term since the rough result was accepted.

Number of workers in this work: 15 women workers The relative accuracy“S” is the
ratio of the absolute accuracy
Mean working hour: 8 hours per day
”e” for the mean P (mean non-
Approximate operating rate : 75% from the past data operating rate)
Expected tolerance by the work sampling:
±4% with the relative accuracy.
Calculating formula of
1) Calculation of the required number of samples relative accuracy “S”

P=0.75, S=0.04, N=? S=2 1-P

√ NP
N=4(1-P)/ S2P ＝ 4X(1-0.75)/0.042X0.75 ＝ 833 」
Calculating formula of N in
・ Sample number 900 was chosen, because the completion case of relative accuracy
of study was expected in 3 days.
・ The daily sample number=900÷3=300 samples 4(1-P)
N=
・ The daily frequency of observations is 300÷15=20 times S2 P
33
Cont…13-2 The cases of Work Sampling
2) The result of Work Sampling
Operating rate P=718÷900 ＝ 79.8 ％
Operating Idling Total(N)
Rate of idling =182÷900 ＝ 20.2 ％
1st day 240 60 300
The relative accuracy of the rate
2nd day 242 58 300 of utilization with 900 samples
3rd day 236 64 300
Total 718 (79.8%) 182 (20.2%) 900 S=2 1-P = 2 1-0.798 =0.034
√ NP √ 900X0.798
Relative accuracy is ±3.4 ％ → within ±4 ％ of tolerance

Calculation of mean assembling time per one unit

The number of assembled unit in 3 days was 9,325 units by a production volume investigation.
Mean assembling time per one unit Idle rate 15 of 20.2% is
Operation time 1 permitted as the allowed
＝ X
Number of assembled units 1-Ratio of allowance level, but 5.2%(=20.2-15)
cannot be accepted as a
Operating time:8hoursX60minutesX3daysX15personsX0.798 ＝ 17,237minutes factor, therefore 5.2% was
Number of assembled units:9,325 units excluded
Rate of allowance:15%(for the working hours included allowable time)

Mean assembling time per one unit ＝ (17,237÷9,325)X(1 / (1-0.15))=2.17 minutes

Various standards are set in this company based on the result. 34

 End

The following books were referred to for preparing this manual on Work Sampling.

1) “IE method in work place” by Mr. Kinichi IKENAGA (published by JUSE)

2) “IE text in work place” by Mr. Katsuyoshi ISHIHARA (published by JUSE)
3) “Production control in parts factory” by Mr. Kunihiko HUTABA (published by Kogyo Tosho)
4) Material by World academy online

35