Each day, a sample of 50 items from a production process was examined. The number of defectives found in each sample was as follows :
6, 2, 5, 1, 2, 2, 3, 5, 3, 4, 12, 4, 4, 1, 3, 5, 4, 1, 4, 3, 5, 4, 2, 3.
Draw a suitable control chart and check for control. What control limits would you suggest for subsequent use ?
15 marks · directive: Draw
The core of a good answer
An np chart suits a constant sample size of 50: with 88 defectives in 24 samples, p̄ = 88/1,200 = 0·0733, centre line np̄ = 3·67, UCL = 3·67 + 3√(50 × 0·0733 × 0·9267) = 9·20 and LCL = 0 (negative); sample 11 (12 defectives) lies above the UCL so the process is out of control, and after removing it the revised limits for future use are p̄ = 76/1,150 = 0·0661, CL = 3·30, UCL = 8·57, LCL = 0, within which all remaining samples fall.
An outline written by upsc.help, not an official UPSC answer; every fact checked by a separate review. How we work
What the question asks
Choose the chart (number of defectives with constant n means an np chart; a p chart is equivalent), compute trial limits, identify out-of-control points, remove points with assignable causes and recompute standard limits for future use. Describe the plot.
Answer outline
Introduction
k = 24 samples, n = 50 each; the statistic is d = number of defectives, d ~ Binomial(50, p) under control.
Step 1: Centre line
- Σd = 88, total inspected = 24 × 50 = 1,200
- p̄ = 88/1,200 = 0·07333
- CL = np̄ = 50 × 0·07333 = 3·667
Step 2: Trial control limits
- σ_d = √(np̄(1 − p̄)) = √(3·667 × 0·92667) = 1·843
- UCL = 3·667 + 3(1·843) = 9·197
- LCL = 3·667 − 5·530 = −1·863, taken as 0
Step 3: Plot and check
- Plot sample number 1 to 24 on the x-axis and d on the y-axis with lines at 0, 3·67 and 9·20.
- Sample 11 has d = 12 > 9·20: out of control. All others lie between 1 and 6, within limits, with no long runs on one side of the centre line.
- Conclusion: the process was not in statistical control during the period.
Step 4: Revised limits for future use
- Assume an assignable cause is found for sample 11 and removed; drop it.
- Σd = 76 from 23 samples, p̄ = 76/1,150 = 0·06609
- CL = 50 × 0·06609 = 3·304
- UCL = 3·304 + 3√(3·304 × 0·93391) = 3·304 + 5·270 = 8·574
- LCL = 3·304 − 5·270 < 0, taken as 0
- All 23 remaining points (maximum 6) are inside, so these are suggested as standard limits.
Step 5: Equivalent p chart
- Trial: CL 0·0733, UCL 0·184, LCL 0; revised: CL 0·0661, UCL 0·171, LCL 0.
Conclusion
Suggested limits for subsequent use: np chart with CL = 3·30, UCL = 8·57 (so 9 or more defectives in a sample of 50 signals trouble) and LCL = 0. Numbers computed in Python.
Examples and data to cite
- np charts are used for count of defective units with constant sample size, for example daily inspection of 50 castings or bulbs.
- Because the LCL is zero, unusually good samples cannot be detected; larger n would be needed to obtain a positive LCL (n > 9(1 − p)/p, about 127 here).
Facts to remember
- np chart: CL = np̄, limits np̄ ± 3√(np̄(1 − p̄)).
- p̄ = total defectives / total inspected.
- Negative LCL is set to 0.
- Revise limits after removing points with identified assignable causes.