WORK SAMPLING

Work Sampling
• A technique that discovers the proportions of
total time that various activities contribute to
the job by taking a relatively large # of
observations at random intervals.
• Used to determine: production standards,
machine and personnel utilization, and job
allowances. Typically faster and
cheaper than other techniques.
 Advantages
• Well suited for high cycle time and low
repetition rate jobs.
• Does not require the analyst to continually
observe the job.
• Reduced clerical time.
• A smaller # of hours are required to collect the
data.
• Operator not subjected to long
periods of observations.
 Advantages (con’t)
• Single analyst can study a small crew (or # of
subjects)
• Cost of the study may be cheaper (up to a certain #
of samples)
• Conducted over a long period of time (more of a
normal performance)
• Study can be postponed if something more
urgent arises (no change in study criteria)
 Disadvantages
• Will not produce as good a standard as direct time
study
• Cost maybe higher than suspected if sampling rates
are frequent
• Operators need to be located close together to
eliminate travel
• Doesn’t provide the detailed data (C/I ideas) that
direct study does
• Statistics of why it works is confusing for
most personnel
 Number of Sample
• Probability based on the binomial distribution
(p + q)n = 1
• Mean = np, Variance = npq
• As n becomes large, the binomial distribution
approaches the normal distribution
• This proportion has distribution of
Mean = p, Variance = √(pq/n)
• There are many sources of tabulated data for
determining the # of samples required at
various confidence levels and accuracy’s.
• Formula for calculation purposes simplifies to:
n = (Zα/2)2 x P x (1-P)/(l2)

• n = total number of random observations upon which P is based

• l = error, 1 – confidence level
• P = true percentage occurrence of element being sought, expressed as a
decimal
 Example
An analyst wants to determine the amount of
downtime of 10 CNC machines. A pilot study
indicates that out of 25 observations, only 1
machine was down (p= 0.04). Confidence level
95%, how many observations needed?

n = (1.96)2 x 0.04 x (1-0.04) / (0.05)2 = 59

 Standard Time
OT = (T/P) x (ni/n)
• T = total time
• P = total production
• ni = number of occurrence for element i
• n = number of observation

Ingat kembali!
OT = Observed Time
NT = Normal Time
 Study Plans (Protocol)
• Start with preliminary estimate of the variable. This
can be historical data, conducting pilot study, or an
educated guess (least desirable)
• Determine the desired accuracy of the results
(confidence interval)
• Estimate the # of observations
• Develop a sampling schedule
• Design the data collection form, and
control charts if applicable
Work Sampling Form (Instrument)
• Forms should be custom designed to
accommodate the specific data of interest to
your study. Make the spaces large enough to
easily record the data. Provide summary and
calculations spaces right on the form.
• Commercial software available for this
application
 Technique
• Locate yourself at the same place each time
prior to observing the operation
• Try to intentionally distract yourself from the
variable of observation as you approach the
site (think safely)
• Limit your time at the site to that
actually needed for the observation
 Technique (con’t)
• Try to record only the minimum data that you
will need to successfully reconstruct the
observation
• Verify any discrepancy with the supervisor or
foreman
• Make notes on the form after the operator
can no longer see you
• Keep a pleasant attitude
 Computerized Work Sampling
• IE time increased by a reduction in clerical
time, results realized faster
• More professional appearing report
• $ of performing studies is cheaper
• Improved accuracy
• Reduced errors by analysts
• Greater use made of work sampling,
because it is less painful
 Summary
• “Sell” the work sampling method before using
it
• Use as large of a sample size as is practical
• Take individual observations at random times,
balance the study if necessary
• Take the observations over two weeks or
more
EXERCISE
 Example 1
Information Source Data
Total working day Time card 480 min
(working + idle)
Number of units drilled Inspection department 420 unit
Working fraction Work sampling 85 percent
Average rating Work sampling 110 percent
Allowances Work sampling 15 percent

Then OT, NT, and ST ?

OT = Observed Time
NT = Normal Time
ST = Standard Time
 Observation Element 1 Element 2 Element 3
Number
1 90

Example 2 2
3
4 95
110

5 100
6 100
An analyst made 30 7 105
8 90
observation over 15 min 9 110
10 85
on a work assignment 11 95
involving three elements, 12
13
90
100
during which 12 units 14
15 80
95

were produced. 16 110

17 105
Performance rating is in 18 90
19 100
Table 14.2. Allowances is 20 85
21 90
10 percent. Then OT 1-3, 22 90
23 110
NT 1-3, and Total Standard 24 100
Time ? 25
26
95

27 105
28 100
29 110
30 110 ___ ___
ΣRating 860 705 1180