Sampling 03: Stratified Random Sampling

Definition: Stratified random sampling is a method used to increase the representativeness of a sample by dividing the population into distinct groups (strata) and then sampling from each group.

Stratification: Dividing the population into groups based on specific characteristics. These groups are called strata.
- Example: Population consists of 40 women and 60 men.
- Aim: To ensure representation across different strata in the sample.
Random Sampling: Selecting individuals randomly from the strata to create a sample.
- Simple random sampling ensures that each individual has an equal chance of being selected.

Scenario: Measuring cholesterol levels in a population of 100 individuals (40 women and 60 men).
Simple Random Sampling Example: Selecting 10 individuals randomly.
- Possible outcomes: 6 women & 4 men, 3 women & 7 men, etc.
- Issue: Although randomly chosen, samples may not be representative of the population.

Increased Representativeness: Helps ensure that every subgroup is accurately represented in the sample.
- Example of stratified selection:
  - Pick 4 women from the group of 40.
  - Pick 6 men from the group of 60.
Reduced Variability: Within strata, individuals are homogeneous, leading to lower variability within each group.

Extended Example: Population is further divided into four strata based on smoking habits:
- Women who smoke
- Women who don’t smoke
- Men who smoke
- Men who don’t smoke
Sample Selection from Each Stratum:
- Example: 3 from non-smoking women, 1 from smoking women, different combinations are chosen randomly each time.
Final Sample Construction: After sampling from each stratum, combine all sampled individuals to form a stratified random sample.

Stratified random sampling is an effective method for ensuring that important subgroups are represented in research studies, enhancing the overall validity of the conclusions drawn from the sample.