Sampling INFO 271 B LECTURE 9 COYE CHESHIRE

Agenda 2 Non-probability Sampling Probability Distributions Info 271 B

Why Sample? 3 Drawing populations versus ‘samples’ E. g. , survey of i. School masters students Reducing Error Info 271 B

Non-Probability Sampling 4 Quotas Pick key groups of interest and find individuals to fill specific goals (i. e. , 100 people in each key group). Quotas are fulfilled without using random sampling Purposive Sampling Find key groups and only study them Info 271 B

More Non-Probability Sampling 5 Convenience Sampling Taking anyone you can get to participate Snowball Sampling Info 271 B Find starting point and use these individuals to get next participant…and so on

Probability and Probability Sampling 6 Are things that we observe different from what would be expected by chance?

Real-World Example: Attitudes About the President (1 -5 Likert Scale) 7

Do men and women differ in their assessment of the US president? 8

Probability Sampling 9 Sampling Frames A list of units of analysis from which you take a sample and to which you generalize Directories, local census, etc. But, often cannot get an adequate sampling frame Info 271 B Field research

How Large of a Sample? 10 Sample Accuracy versus Sample Precision Key Issues for Determining Sample Size: 1) 2) 3) 4) Info 271 B Heterogeneity of Population Number of Subgroups Size of Subgroup Precision of sample statistics

Randomized Samples 11 Simple Random Sample Requires numbering all potential participants in a given sampling frame (N) Pull random numbers from any source, use these as the sample (n) Select n units out of N such that each NCn has and equal chance of being selected. Issues… Info 271 B True randomization? Replacement “in the field”

Systematic Random Samples 12 Random start and sampling interval (Sample x Interval = Population) Issues Periodicity Poor sampling frames (www. socialresearchmethods. net) Info 271 B

Info 271 B 13

Stratified Random Sampling 14 Key Issue: Representation of salient sub-populations Maximize between-group variance while minimizing within-group variance Proportionate Samples Do you know the key independent variables? If not, may be better off avoiding stratification Disproportionate Samples Weighting Info 271 B

Complex Sampling Designs 15 Cluster Sampling No available frames Based on areas, institutions, or ‘clusters’ Info 271 B