Population and Sampling Methods

Population

  • Defined as the group to which study results apply (target group).

Non-Probability Sampling (no random selection)

  • Key feature: individuals do not have equal selection chance.
  • Types:
    • Convenience (reliance on available subjects)
    • Purposive (select pre-defined group)
    • Snowball (participants recruit others; useful for hard-to-reach groups)
    • Quota (match sample to population on control categories, e.g., age, gender)
    • Consecutive (include every eligible subject until required size reached)

Probability Sampling (uses random selection)

  • Ensures every population unit has equal chance.
  • Limited by incomplete frames, time, or resources.
  • Types:
    • Simple Random (assign numbers, draw via random method)
    • Systematic (select every kthk^{th} unit; k=Nnk = \frac{N}{n})
    • Stratified (divide into strata; sample within each — proportional or disproportional)

Sample Size Guidelines

  • Determined by method & study design.
  • Focus group: 6−126{-}12 participants (not generalizable).
  • Per variable cell: 50, 75, 10050,\ 75,\ 100.
  • Multivariate studies (total n):
    • 100100 poor | 200200 fair | 300300 good | 500500 very good | 10001000 excellent.
  • Allow extra 10−25%10{-}25\% for expected drop-outs (e.g., longitudinal studies).
  • Fewer than 3030 per cell ⇒ unstable results.

Sampling Error (Standard Error)

  • Arises because sample data substitute for full population.
  • Only calculable for probability samples.
  • Must be estimated to understand result variability.