Sampling Participants Lecture PDF

SELECTING RESEARCH PARTICIPANTS Overview  Population and samples  Generalization  Sampling Methods  Probability  Nonprobability Simple random  Convenience Systematic  Snowball Stratified random Proportionate ran...

SELECTING RESEARCH PARTICIPANTS Overview  Population and samples  Generalization  Sampling Methods  Probability  Nonprobability Simple random  Convenience Systematic  Snowball Stratified random Proportionate random  Quota Cluster Multi-stage Choosing Participants  High school students’ attitudes towards unrestricted searches  Who should take the survey?  All high school students in the nation?  All 5,000 students in a local school district?  Only high schools that have reported problems with drugs & weapons?  Bottom line:  Not everyone can participate  Interpretation of data depends on who are participants are and how they were selected Populations & Samples  Population  Entire set of individuals of interest  Usually targeted, not everyone on the planet  Sample  Set of individuals selected from a population  Intended to represent the population Sampling & Generalization  Generalization  Apply results from sample to population  Depends on how representative the sample is  Selection bias  Sampling method used that leads to biased sample Sampling Methods  Probability sampling  Can determine probability that each member of population will be selected for sample  Population known – to establish a sampling frame  Even through random selection, sample may not be representative  Larger sample size increases the likelihood that sample looks like population Probability Sampling  Simple Random Sampling  Every member of population has an equal & independent chance of being selected  Reduces systematic bias Does not guarantee a representative sample  Can be difficult & time consuming, or even not possible Sampling  Systematic Sampling  Every kth element is sampled Probability Sampling  Stratified random  Population is divided into strata  Random sample drawn from each stratum  Increases chances of obtaining representative sample But may underrepresent or overrepresent some strata 150 Accounting Majors Sample (N = 50) 200 Biology Majors 10 Accounting 10 Biology 100 Art Majors 10 Art 10 Psychology 250 Psychology 10 Math Majors 100 Math Majors Probability Sampling  Proportionate random  Proportions of different groups in the population are reflected sample strata  Even greater chance that sample is representative Total Population = 800 Majors 150 Accounting Majors (19%) Sample (N = 50) 200 Biology Majors (25%) 10 Accounting (20%) 12 Biology (24%) 100 Art Majors (13%) 6 Art (12%) 16 Psychology (32%) 250 Psychology Majors 6 Math (12%) (31%) 100 Math Majors (13%) Probability Sampling  Cluster  Used w/ large populations  Sampling is at group level rather than individual level  Groups are randomly sampled School District 1 School District 2 Sample consists of School District 3 everyone from District 1 and 4 School District 4 School District 5 Probability Sampling  Multistage sampling  First, identify clusters Randomly sample  Next, sample individuals from clusters  Note: may be necessary to use more than 2 stages Nonprobability Sampling  Convenience Sampling  Sampling individuals who are readily available  Advantages  Time, money  Might be the only method possible  Disadvantages  Assumes characteristics that make your sample convenient are unrelated to variables of interest  Difficult to know if (or how) your sample is biased Nonprobability Sampling  Snowball sampling Identify some participant Have them refer others  Quota Mimics stratified, but not random  Mini-Review 1. Explain how stratified and proportionate sampling are similar and how they are different. 2. What sampling methods are likely to ensure the greatest chance that our sample will represent the population well? Why?