Observational Study Designs

Observational Study Designs

1. Simple Random Survey

   • randomly selected from sampling pop

2. Stratified Survey

   • statistical population has inherent grouping (strata) that can influence results
     • know ahead of time

3. Cluster survey

   • natural clustering, not independent within cluster
     • ex. all individuals within household
     • clustering group is chosen randomly then ask all within that group

4. Case-Control Survey

   • Trying to explain known outcome
     • strong risk of spurious relationships---intentional bias
     • looking at cases with outcome want but no control
   • restrospective

5. cohort survey

   • follow sampling units through time to see dev of response variable
     • less chance of spurious relationship
   • prospective