Describe the effect of sample size on X̄, R and s charts for the sample sizes n ≤ 5, 5 < n ≤ 10 and n > 10.
10 marks · directive: Describe
The core of a good answer
Larger subgroups make the X̄ chart limits (±3σ/√n) tighter and more sensitive to small shifts; for n ≤ 5 the range is almost as efficient as s (95·5% at n = 5) and the R chart has no lower limit, for 5 < n ≤ 10 the range loses efficiency (85% at n = 10) though either chart works, and for n > 10 the s chart should replace the R chart because the range wastes the information in the middle observations.
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
A descriptive question, but it needs numbers: the control chart constants (A₂, d₂, D₃, B₃) and the relative efficiency of R versus s for estimating σ. Organise by the three size bands asked, and cover all three charts in each band, plus practical considerations (cost, rational subgroups).
Answer outline
Introduction
Variables control charts monitor the mean with X̄ and the spread with R or s. The subgroup size n enters every limit through the constants, so it changes sensitivity, false alarm behaviour and which spread chart is efficient.
Step 1: X̄ chart and n
- Limits X̿ ± 3σ/√n = X̿ ± A₂R̄ (or ± A₃s̄); A₂ = 0·577 at n = 5, 0·308 at n = 10, 0·180 at n = 20.
- Larger n narrows the limits, so a shift of δσ in the mean is detected faster: the probability of detection depends on δ√n.
- Large n raises sampling cost and risks mixing assignable causes within a subgroup (rational subgrouping).
Step 2: n ≤ 5 (small subgroups, the usual practice)
- R chart preferred: easy to compute on the shop floor, and its efficiency relative to s is high (100% at n = 2, 99·2% at n = 3, 97·5% at n = 4, 95·5% at n = 5).
- D₃ = 0 (no lower control limit on R, valid for n ≤ 6) and B₃ = 0 for s (n ≤ 5), so a decrease in variability cannot be signalled.
- X̄ chart is sensitive only to fairly large shifts (about 1·5σ or more).
Step 3: 5 < n ≤ 10
- Efficiency of R falls: about 93% at n = 6 and 85% at n = 10, because the range uses only two observations.
- R chart gets a positive lower limit from n = 7 (D₃ = 0·076); s chart from n = 6 (B₃ = 0·030).
- Either chart is acceptable; s is preferred when computation is automated. X̄ chart detects moderate shifts (about 1σ) quickly.
Step 4: n > 10
- Use X̄ and s charts: s uses all n deviations and is clearly more efficient (R efficiency about 81% at n = 12, 70% at n = 20).
- σ is estimated by s̄/c₄ (c₄ = 0·9727 at n = 10, 0·9869 at n = 20) rather than R̄/d₂.
- Larger n also makes X̄ nearly normal by the central limit theorem even for skewed data, and the X̄ chart can detect small shifts (under 1σ).
Conclusion
Small subgroups (n ≤ 5) favour X̄ and R charts, the middle band allows either spread chart with s gaining ground, and n > 10 calls for X̄ and s charts with tighter, more sensitive limits. (Constants recomputed in Python from the normal range distribution.)
Examples and data to cite
- The X̄ and R chart is described for relatively small subgroups (n ≤ 10) and the X̄ and s chart for relatively large ones (n > 10).
Facts to remember
- σ̂ = R̄/d₂ (d₂ = 2·326 for n = 5, 3·078 for n = 10) or s̄/c₄.
- X̄ limits: X̿ ± A₂R̄ with A₂ = 3/(d₂√n).
- R limits: D₃R̄, D₄R̄; s limits: B₃s̄, B₄s̄.
- Relative efficiency of the range to s falls as n grows: 0·955 at n = 5, 0·850 at n = 10.
- Larger n gives a smaller β risk (higher power) for a given mean shift.