PSYC 202 Chapter Notes - Chapter 10: Cohort Study, Statistical Inference, Type I And Type Ii Errors
15 Jun 2018
School
Department
Course
Professor
WEEK 10 VIDEO NOTES
Video 1
Populations for observational designs
- Identify population because it is what the study is about
- Statistical pop. Is what you are sampling
- Can be more than one population
- The population is the group of all observation units that have a non-zero probability of
being included in your sample
- Be careful of bias and non-independence
- Simple random survey
o Take all possible units and randomly select
o Population is the group of all observation units that can be sampled
- Cluster survey
o Each cluster contains observation units
o Cluster is randomly sampled
o Population is all clusters
- Stratified
o Groups that are already grouped
o Each strata has a population
o Randomly select observation units from each strata
o Population is the group of all observation units that can be sampled within each
strata
- Case control study
o Control and case both have a population
o Population is the possible observation units in control and case groups
- Cohort study
o Randomly select observation units and look at relationship between dependent
and independent variables at some later time
o Population is the possible observations units that could be selected to be in the
cohort
Population for experimental design
- Experimental study design
o Control treatment design
▪ Each control and treatment has a population
▪ Populaiton is the group of all observation units that can be sampled within
the control and treatment
o Blocking deisgn
▪ Each block and treatment has a population
▪
o Multiple factor design
▪ Each combination of levels has a population
▪
▪ Population is the group of all observation units that can be sampled within
each cell
Video 2
Document Summary
Identify population because it is what the study is about. The population is the group of all observation units that have a non-zero probability of being included in your sample. Simple random survey: take all possible units and randomly select, population is the group of all observation units that can be sampled. Cluster survey: each cluster contains observation units, cluster is randomly sampled, population is all clusters. Stratified: groups that are already grouped, each strata has a population, randomly select observation units from each strata, population is the group of all observation units that can be sampled within each strata. Case control study: control and case both have a population, population is the possible observation units in control and case groups. Cohort study: randomly select observation units and look at relationship between dependent and independent variables at some later time, population is the possible observations units that could be selected to be in the cohort.
