Track 6 · Sampling and the Central Limit Theorem · Lesson 8 of 8

Sample size, precision and cost

Plan the sample size needed for a target margin of error, explain the diminishing returns of larger samples, and allow for nonresponse.

30 minIntermediate7 graded exercisesMembers

A survey team quotes 400 interviews for a ±5-point margin of error. The client asks for ±2.5 points “for a bit more money” — and is told it will cost four times as much. Precision grows only with the square root of the sample size, and every planning decision follows from that fact.

Because SE⁡=σ/n\SE = \sigma/\sqrt{n}, precision improves with the square root of the sample size. Doubling nn cuts the SE by only 1−1/2≈29%1 - 1/\sqrt{2} \approx 29\%; halving the SE needs four times the sample; cutting it to a tenth needs a hundred times. Costs, meanwhile, usually grow in proportion to nn. The first few hundred observations buy a lot of precision; each extra thousand buys less.

Included with a Formula Quest membership

It has 7 graded exercises with feedback, hints and worked solutions. After it you will be able to:

  • Plan the sample size for a target margin of error for a mean or a proportion, rounding up
  • Explain why halving the margin of error needs four times the sample
  • Inflate a planned sample for the expected response rate
  • Weigh extra precision against cost, and separate precision from bias