Sampling Techniques

by
Variables
 Topic Outcome:
 At the end of this topic, students will be able to:
 Discuss the advantages and disadvantages of
variable sampling.
 Apply Process Fraction Nonconforming Type of
variable sampling.
 Use ANSI/ASQ Z1.9 – 1993.
 Describe the Shainin Lot Plot Plan.
 Outline:
 An Introduction.
 Types of Sampling Plan.
- Shainin Lot Plot Plan.
- ANSI/ASQ Z1.9
 Variability Unknown
Standard Deviation Method
Range Method
 Variability known
 An Introduction
 When a quality characteristic is measurable on a
continuous scale, and is know to have a distribution of a
specific type (e.g. normal distribution); it is possible to use
a sampling based on sample measurements such as:
Mean and
Standard Deviation.
 Variables Sampling Plans.
 Population

Sample
LSL USL
Compare

LSL USL

ACCEPT Compare to required Quality Level REJECT

 pu
pL
- +

LSL USL

Percent nonconforming below and above specifications.

 Advantages of Variable Sampling Plan:
 Same protection with smaller sample size.
 Feedback of data on Process.
 Extent of conformity given weightage.
 Errors in measurement more likely detected.
 Disadvantages of Variable Sampling Plan:
 Applicable to only one quality characteristic at a time.
 Higher inspection cost – Measurement.
 Higher clerical cost – Calculation.
 Possibility of no nonconformity in a rejected lot.
 Dependence on assumption of distribution shape.
Types of Variable Sampling Plan
Variable Sampling Plan

Percent nonconforming Process Parameter

To determine the To control the average

proportion of product that and standard deviation of
is outside specification the distribution of the
product to specified level
E.g.:
E.g.: E.g.:
Shainin
Shaininlot
lot plot
plot Acceptance control chart
ANSI/ASQ
ANSI/ASQZ1.9-1993
Z1.9-1993 Sequential sampling for variables
Hypothesis testing
 Shainin Lot Plot Plan
 Developed by Dorian Shainin (Hamilton Standard Division
of United Aircraft Corporation, 1950).
 The plan uses a plotted frequency distribution
(histogram) to evaluate a sample for decisions
(accept/reject a lot).
 It is a practical plan for in-house inspection & receiving
inspection.
 Advantages:
 Applicable to both normal and non-normal frequency
distributions.
 Simple to use.
Lot Plotting Method

 The method for obtaining the lot plots is as follows:

 A random sample of 10 subgroups (g) of 5 each for a
total of 50 items is obtained from the lot.
 The average, X-bar and range, R, are calculated for
each subgroup.
 A histogram is constructed (number of cell: 7-16).
 Calculate the average of averages, X-double bar, and
average of ranges, R-bar.
 Calculate the upper and lower lot limits (ULL & LLL)
Q&A
  X 976.8
X   97.7
g 10
3
 R 13.7
control R   1.37
limits g 10

3R (3)(1.37)
ULL  X   97.7   99.5
d2 2.326
3R (3)(1.37)
LLL  X   97.7   95.9
d2 2.326

Chart for Ranges:

Factor for Central Line  d2
 Fig 10-14: Lot plot histogram
Lot Plot Evaluation
 Decision is based on a comparison of the lot plot with 11
different types of lot plots.
Type Distribution Comments
-within spec.limits
1
- accepted without calculate lot limits.
-lot limits within spec. limits.
2
normal -accept.
3 -lot limits outside spec. limits.
-percentage of product beyond spec is obtained.
4 -Review board determines the final disposition of the stock.
5 skrewed
nonnormal
6 Lot was screened & sorted
7 Bimodal condition
8
nonnormal Lot was screened & sorted
9
10 Bimodal condition
11 Stray values
 Summary

 Once learned, the lot plot procedure is relatively simple and

has resulted in improved quality and lower inspection
cost.
 Unacceptable lots are returned to the producer, and this
action will cause a subsequent improvement in quality.
 Inspectors can accept lots; however, disposition of
unsatisfactory lots is left to a material-review board.
 Many users of the lot plot method have modified the Shainin
method for their own situation.
 The major criticism of the plan is that the shape of the plot
does not always give an accurate indication of the true
distribution. Shainin states that the plot is close enough to
have no practical effect on the final decision, or if there are
any errors, they are in a safe direction.
 ANSI/ASQ Z1.9
 MIL-STD-414 ASQ  closely match ANSI/ASQ Z1.4
and ISO/DIS 3951.
 Indexed by AQL (0.10-10.0%).
 Assumption: normally distributed random variable.
 Composition of the standard: 9 different procedures to
evaluate a lot.
Variability Unknown Variability Unknown Variability known
(Standard Deviation Method) (Range Method)

LSL or •LSL and

Single Specification Double Specification
USL USL
•1AQL or
Form 1 Form 2 Form 2
2AQL
(k method) (m method) (m method)
Quality Index  k Quality Index  p0  M

k = Acceptability Constant; M = Maximum Allowable Percent Nonconformity

 It is divided into 4 sections:
 Section A: General description of sampling plan, sample
size, code determination, OC curves.
 Section B: Unknown Variability – Standard Deviation
Method.
 Section C: Unknown Variability – Range Method.
 Section D: Know Variability.
 It could apply either with a single specification limit or two
specification limits.
 Specification limits:
 It is the requirement that a quality characteristic should meet.
 This requirement may be expressed as an upper spec. limit
(USL); or a lower spec. limit (LSL), called herein a single
specification limit; or both upper and lower specification limits,
called herein a double specification limit.
 For single specification Form 1 and Form 2 give identical
end results.
 Form 1: Decision on Acceptability is to compare calculated
Quality Index to Acceptability Constant, k.
 Form 2: Decision on Acceptability is to convert calculated
Quality index to Percent Nonconformity in lot p, and then
compared to Maximum Allowable Percent Nonconformity,
M.
 IMPORTANT (Normality Assumption):
 This standard assumes the underlying distribution of
individual measurements to be normal in shape. Failure of
this assumption  affect OC curves and probabilities
based on these curves.  affect the estimate of percent
nonconforming calculated from mean and standard
deviation. The assumption should be verified prior to use of
the standard. A variety of statistical tests and graphical
techniques are available.
Determination of Sample Size

 Relative samples are designated by code letters.

 The sample sizes code letter depends on the inspection
level and lot size.
 There are 5 inspection levels:
 General Levels
 I, II, and III
 Special Levels:
 S3 and S4
 The sample size code letter applicable to specified
inspection level and for lots of given size shall be obtained
from Table A-2
 Note:
 Unless otherwise specified, inspection Level II shall
be used.
 Inspection Level I may be specified when less
discrimination is needed.
 Inspection Level III may be specified for greater
discrimination.
 Level S3 and S4 may be used when relatively small
sample sizes are necessary and large sampling
risks can be tolerated.
 Switching Rules for ANSI/ASQ Z1.9

• Preceding 10 lots START • 2 out of 5

Accepted with total
nonconforming less than consecutive
limit number, lots not
• Steady production, and accepted
• Approval from
responsible authority.

NORMAL TIGHTENED
REDUCED

• Lot not accepted, or • 5 consecutive

• Lot accepted but lots accepted
nonconformities found
lie between Ac and Re • 10 consecutive lots
of the plan, or
Remain on Tightened
• Irregular production, or
• Other conditions
warrant.
• Discontinue
inspection
Variability Unknown – Standard Deviation Method

 There are two parts in this method

 Single Specification Method  Form 1 and 2.
 Double Specification Method  Form 2.
 Single Spec. Method – Form 1
 A) Determine sample size code (Table A-2) by using
lot size and inspection level.
 B) Select plan from Master Table B-1 and B-2. Obtain
sample size, n, and the acceptability constant, k.
 C) Select at random the sample of n units from the lot;
inspect and record the measurement of each unit.
 D) Compute x-bar,sample standard deviation, s,
upper/lower specification limit.
 E) If the upper/lower spec limit  k  lot accepted.
Variability Unknown – Standard Deviation Method
 Single Spec. Method – Form 2
 It converts the computed results into percent
nonconforming (p0) into lot through Table B-5 and
compares it with Maximum Allowable Percent
Nonconforming, M.
 A) Determine sample size code (Table A-2) by using lot
size and inspection level.
 B) Select plan from Master Table B-3 and B-4. Obtain
sample size, n, and M.
 C) Select at random the sample of n units from the lot;
inspect and record the measurement of each unit.
 D) Compute x-bar, sample standard deviation, s, Quality
Index, QU or QL.
 E) If lot percent nonconforming, pu or pL  M  lot
accepted.
Table B-5
Q&A

 The minimum temperature of operation for a certain device is specified as

180oC. A lot of 40 items is submitted for inspection where inspection level
II, normal inspection,and AQL = 1.0% are the criteria. (Form 1 & 2)

From Table A-2, the code letter D, which gives a sample size n=5
(Table B-1). The temperatures for 5 samples are 197, 188, 184, 205, and
201oC.
 X 197  188  184  205  201
X   195o C
n 5
2

 X
2
X 190.435  190.125
s n   8.80
n 1 5 1
X  L 195  180
low quality index, QL    1.70
s 8.80
Acceptability Cons tan t , k  1.53
 QL  k  Accepted
 Estimated a lot percent nonconforming below L: pL
 From Table B-5, pL = 0.66%
 Maximum allowable percent nonconforming, M
 From Table B-3, M=3.33%
 pLM  accepted
Q&A

 The maximum temperature of operation for a certain device is specified

as 208oC. A lot of 40 items is submitted for inspection where inspection
level II, normal inspection,and AQL = 1.5% are the criteria. (Form 1 & 2)

From Table A-2, the code letter D, which gives a sample size n=5
(Table B-1).

X
X  195o C
n
2

 X
2
X 190.435  190.125
s n   8.80
n 1 5 1
U  X 208  195
low quality index, QU    1.48
s 8.80
Acceptability Cons tan t , k  1.40
 QU  k  Accepted
 Estimated a lot percent nonconforming above U: pU
 From Table B-5, pU = 4.22%
 Maximum allowable percent nonconforming, M
 From Table B-3, M=5.83%
 pUM  accepted
Variability Unknown – Standard Deviation Method
 Double Spec. Method – Form 2
 It can be either:
 One AQL value for both Upper and Lower
Specification Limit Combined, or
 Different AQL values for Upper and Lower
Specification Limit.
Variability Unknown – Standard Deviation Method
 Double Spec. Method – Form 2
 One AQL value for both Upper and Lower Specification
Limit Combined
 A) Determine sample size code (Table A-2) by using lot
size and inspection level.
 B) Select plan from Master Table B-3 and B-4. Obtain
sample size, n, and M
 C) Select at random the sample of n units from the lot;
inspect and record the measurement of each unit.
 D) Compute x-bar, sample standard deviation, s, Quality
Index, QU or QL.
 Determine the estimated lot percent nonconforming,
p=pL+pU (Table B-5).
 E) p  M  lot accepted.
Q&A

 The minimum and maximum temperatures of operation for

a certain device is specified as 180oC and 209oC. A lot of
40 items is submitted for inspection where inspection level
II, normal inspection,and AQL = 1.0% are the criteria.

From Table A-2, the code letter D, which gives a

sample size n=5 (Table B-1). The temperatures for 5
samples are 197, 188, 184, 205, and 201oC.
  X 197  188  184  205  201
X   195o C
n 5
2

 X
2
X 190.435  190.125
s n   8.80
n 1 5 1
X  L 195  180
low quality index, QL    1.70
s 8.80

Estimated a lot percent nonconforming below L: pL

From Table B-5, pL = 0.66%

 U  X 209  195
QU    1.59 ( say 1.60)
s 8.80

Estimated a lot percent nonconforming above U= pU

From Table B-5, pU = 2.03%

The lot meets acceptance criteria if pL+pU  M

From Table B-3, M=3.33%

 (0.66+2.03)%  3.32  accepted

Variability Unknown – Standard Deviation Method

 Double Specification Method  Form 2.

 Different AQL values for Upper and Lower Specification
Limit.
 A) Determine sample size code (Table A-2) by using
lot size and inspection level.
 B) Select plan from Master Table B-3 and B-4. Obtain
sample size, n, Mu (AQL for upper spec limit) and
ML(AQL for lower spec limit).
 C) Select at random the sample of n units from the lot;
inspect and record the measurement of each unit.
 D) Compute x-bar, sample standard deviation, s,
Quality Index, QU or QL.
 E) Determine the estimated lot percent
nonconforming, pL and pU (Table B-5), and p (= pL +
pU )
 F) Accept the lot if the following 3 conditions are met:
pU  MU
pL  ML
p  (MU or ML, which ever is larger)
Q&A

 The minimum and maximum temperatures of operation for

a certain device is specified as 180 and 209oC. A lot of 40
items is submitted for inspection where inspection level II,
normal inspection,and AQL = 1.0% for upper and
AQL=2.5% for lower specification limits are the criteria.
From Table A-2, the code letter D, which gives a sample
size n=5 (Table B-1). The temperatures for 5 samples are
197, 188, 184, 205, and 201oC.
Line Information needed Value Obtained Explanation
1 Sample size, n 5
2 Estimated Lot Standard Deviation, s 8.81
3 Sample mean, X-bar 195
4 Upper Specification Limit, U 209
5 Lower Specification Limit, L 180
6 Quality Index: QU 1.59
7 Quality Index: QL 1.70
8 Est. Lot Percent Ncf above U, pU 2.19% Table B-5
9 Est. Lot Percent Ncf below L, pL 0.66% Table B-5
10 Total Est PercentNcf, p 2.85%
11 Max. Allowable Percent Ncf above U, M U 3.32% Table B-3
12 Max. Allowable Percent Ncf below L, ML 9.80% Table B-3
13 Acceptability Criteria:
a) Compare pu with MU 2.19%<3.32%
b) Compare pL with ML 0.66%<9.80%
c) Compare p with ML 2.85%<9.80%
 ANSI/ASQ Z1.9
Variability Unknown Variability Unknown Variability known
(Standard Deviation Method) (Range Method)

LSL or •LSL and

Single Specification Double Specification
USL USL
•1AQL or
2AQL
Form 1 Form 2 Form 2

Quality Index  k Quality Index  p0  M

k = Acceptability Constant; M = Maximum Allowable Percent Nonconformity

Variability Unknown – Range Method

 Basic approach is similar to those of Standard Deviation

Method.
 Master Tables used are Tables C-1 to C-5, instead of B-1
to B-5.
 Quality indices for Form 1 are calculated as follows:

QU 
U  X  ; QL 
 X  L
R R

 As calculation of Quality indices for Form 2, there is an

additional factor c included:

QU 
U  X c
; QL 
 X  L c
R R
Q&A

 A lower specification limit for electrical resistance of a certain electrical

component is 620 ohm. A lot of 100 items is submitted for inspection.
Inspection Level II, normal inspection, with AQL=0.4% is to be used.
From Tables A-2 and C-1 it is seen that a sample of size 10 is
required.Suppose the values of the sample resistance in the order
reading from left to right are as follows:

643, 651, 619, 627, 658 (R1=658-619=39)

670, 673, 641, 638, 650 (R2=673-638=35)
and compliance with the acceptability criterion is to be determined. (using
Form 2)
Line Information needed Value Obtained Explanation
1 Sample size, n 10
2 Sum of measurement 6470
3 Sample mean, X-bar 647
4 Average Range, R-bar 37 (39+35)/2
5 Factor c 2.405 Table C-3
6 Spec limit (lower), L 620
7 Quality Index: QL 1.76
8 Est. Lot Percent Ncf below L, pL 2.54% Table C-5
9 Max. Allowable Percent Ncf, M 1.14% Table C-3
10 Acceptability Criteria 2.45%>1.14%
a) Compare pL with M

The lot does not meet the acceptability criterion.

 ANSI/ASQ Z1.9
Variability Unknown Variability Unknown Variability known
(Standard Deviation Method) (Range Method)

LSL or •LSL and

Single Specification Double Specification
USL USL
•1AQL or
2AQL
Form 1 Form 2 Form 2

Quality Index  k Quality Index  p0  M

k = Acceptability Constant; M = Maximum Allowable Percent Nonconformity

Variability known

 Basic approach is similar to those of Standard Deviation

Method.
 Master Tables used are Tables D-1 to D-5, instead of B-1
to B-5.
 Quality indices for Form 1 are calculated as follows:

QU 
U  X  ; QL 
 X  L
 

 As calculation of Quality indices for Form 2, there is an

additional factor v included (Table D-3 and D-4):

QU 
U  X v
; QL 
 X  L v
 
Q&A

 The specified minimum yield point for certain steel castings is 58,000 psi.
A lot of 500 items is submitted for inspection. Inspection Level II, normal
inspection, with AQL=1.5% is to be used. The variability is known to be
300 psi. From Tables A-2 and D-1 it is seen that a sample size of 10 is
required. Suppose the yield points of the sample specimens are:

62,500; 60,500; 68,000; 59,000; 65,500

62,000; 61,000; 69,000; 58,000; 64,500
and compliance with the acceptability criterion is to be determined.
Line Information needed Value Obtained Explanation
1 Sample size, n 10
2 Known Varibility:  3,000
2 Sum of measurement 630,000
3 Sample mean, X-bar 63,000
4 Spec limit (lower), L 58,000
5 Quality Index: QL 1.67
6 k 1.70 Table D-1
7 Acceptability Criteria 1.67%<1.70%
a) Compare QL with k

The lot does not meet the acceptability criterion.

 Summary - ANSI/ASQ Z1.9
Variability Unknown Variability Unknown Variability known
(Standard Deviation Method) (Range Method)

LSL or •LSL and

Single Specification Double Specification
USL USL
•1AQL or
2AQL
Form 1 Form 2 Form 2

Quality Index  k Quality Index  p0  M

k = Acceptability Constant; M = Maximum Allowable Percent Nonconformity

QL or QU  k
Accept
p, pU, or pL  M or k