Upload
brittney-miller
View
220
Download
0
Embed Size (px)
DESCRIPTION
Stratified Random Sample population is divided into groups of similar individuals (strata), then an SRS is chosen within each strata use when there’s reason to suggest different subsets would give different results capitalizes on pockets of homogeneity within population serves to decrease variability in results from different samples population is divided into groups of similar individuals (strata), then an SRS is chosen within each strata use when there’s reason to suggest different subsets would give different results capitalizes on pockets of homogeneity within population serves to decrease variability in results from different samples
Citation preview
•Chapter 4:Designing Studies...
Sampling
Convenience Sample
Voluntary Response Sample
Simple Random Sample
Stratified Random Sample
Cluster Sample
Types of Samples
Stratified Random Samplepopulation is divided into groups of similar individuals (strata), then an SRS is chosen within each stratause when there’s reason to suggest different subsets would give different resultscapitalizes on pockets of homogeneity within populationserves to decrease variability in results from different samples
Cluster Samplepopulation divided into groups (clusters) whose characteristics mirror those of the population, an SRS of the clusters is chosen, and all individuals within chosen clusters are includedcapitalizes on heterogeneity in populationgenerally does NOT serve to decrease variability, but often is more convenient to use
Why random samples?helps reduce bias in samplingallows us to infer information about the population from what we know about the samplevariability from sample to sample is NOT haphazard...follows laws of probability
How big a sample should we use?
depends....more specifics latergenerally larger samples give better info about population than smaller samples
Errors in Surveys
Sampling ErrorNonsampling Error
Sampling Erroruse of poor sampling methodsundercoverage - when some groups in a population are left out of the sampling process
Nonsampling Errornonresponse error - when individual chosen for sample can’t be contacted or refuses to participateresponse bias - when individual surveyed gives incorrect responsewording and order of questions can also influence the response given