What is stratified sampling
- Supplementary info to help design sample
o Men generally earn more than women
- If variable we are interested in takes on different mean values in different sub populations
- Reasons
o Protection from possibility of obtaining a really bad sample (randomly chosen but not representative)
o Want data of known precision for subgroups
o Cost and convenience considerations
o Stratified sampling, if done correctly, will give more precise estimate for whole population
Allocating Observations to Strata
- Proportional allocation
- Optimal allocation
Post Stratification
- Avoid data snooping
Quota Sampling
- Many samples that masquerade as stratified random samples are actually quota samples
- Population divided into different sub populations
- Important difference à probability sampling is not used to chose individuals from sub populations for sample
- Don't know what is selection bias
No comments:
Post a Comment