Sampling © 2009 University of Wisconsin-Extension, Cooperative Extension, Program Development and Evaluation 1 A sample is a portion or a subgroup of an entire group (called the population) from which an estimate can be made for the entire group. See the booklet, Sampling http://learningstore.uwex.edu/pdf/G3658-03.pdf Also, see Bill Trochim’s Social Science Research Methods at http://www.socialresearchmethods.net/kb/sampling.php © 2009 University of Wisconsin-Extension, Cooperative Extension, Program Development and Evaluation 2 Check your ‘sampling’ IQ by answering the following questions. 1. 2. 3. 4. 5. 6. Why do people use a sample? Does a sample always refer to people? Is there one best way to sample? Should we always sample? What makes a “good” sample? What size does the sample need to be? Answers follow © 2009 University of Wisconsin-Extension, Cooperative Extension, Program Development and Evaluation 3 1. Why do people use a sample? • It is cheaper • It can result in more accurate data. – Managing a representative sample to obtain a high response rate may yield more accurate information than surveying everyone and having response bias. © 2009 University of Wisconsin-Extension, Cooperative Extension, Program Development and Evaluation 4 2. Does a sample always refer to people? NO You might sample • 4-H records • activity logs • gardens • demonstrations • clubs… © 2009 University of Wisconsin-Extension, Cooperative Extension, Program Development and Evaluation 5 3. Is there one best way to sample? NO Depending upon your purpose, there are various sampling strategies. If you want to generalize from the sample to the population, probability sampling is necessary. Sometimes probability sampling is impossible or you don’t want or need to generalize to the larger group. Then, nonprobability sampling is appropriate. © 2009 University of Wisconsin-Extension, Cooperative Extension, Program Development and Evaluation 6 Types of sampling strategies: Probability: • Generalize to population. Some examples: – Simple random sample – Stratified sample – Cluster sample – Systematic sample Nonprobability: • Generalizability not as important. Want to focus on “right cases.” Some examples: – Quota sample – “Purposeful” sample – “Convenience” or “opportunity” sample © 2009 University of Wisconsin-Extension, Cooperative Extension, Program Development and Evaluation 7 4. Should we always use a sample? NO • If the population is small, you may choose to include everyone. • If you want to be inclusive giving everyone the chance to be involved or to know about the issue under study, you might include everyone. © 2009 University of Wisconsin-Extension, Cooperative Extension, Program Development and Evaluation 8 5. What makes a “good” sample? When the sample accurately represents the population. . © 2009 University of Wisconsin-Extension, Cooperative Extension, Program Development and Evaluation 9 6. What size does the sample need to be? “It all depends”… upon the size of the population (the entire group), how varied the population is, how much sampling error can be tolerated, your resources. If your purpose is to generalize or show representativeness, then sample size is a concern. Use a sample size calculator to determine sample size at different levels of precision © 2009 University of Wisconsin-Extension, Cooperative Extension, Program Development and Evaluation 10 Anticipate nonresponse A certain number of people will not respond for one reason or another. Increase the sample size to account for nonresponse. © 2009 University of Wisconsin-Extension, Cooperative Extension, Program Development and Evaluation 11 What do you think? Scenario: The local library wants to know what impact its services are having on the community. It creates a survey, announces the purpose and process of the survey in the local newspaper, places the questionnaires and locked boxes in strategic places around the community for people to complete and securely submit the survey, and posts visible notices at each location to bring attention to the survey. © 2009 University of Wisconsin-Extension, Cooperative Extension, Program Development and Evaluation 12 The previous scenario represents the potential for sampling bias. Bias is a systematic error that can prejudice your evaluation findings in some way. So, sampling bias is consistent error that arises due to the sample selection. For example, a survey of high school students to measure teenage use of illegal drugs will be a biased sample because it does not include home schooled students or dropouts. A sample is also biased if certain members are underrepresented or overrepresented relative to others in the population. For example, distributing a questionnaire at the end of a 3-day conference is likely to include more people who are committed to the conference so their views would be overrepresented. Interviews with people who walk by a certain location is going to over-represent healthy individuals or those who live near the location. Selecting a sample using a telephone book will under-represent people who cannot afford a telephone, do not have a telephone, or do not list their telephone numbers. © 2009 University of Wisconsin-Extension, Cooperative Extension, Program Development and Evaluation 13 What can you do to address sampling bias? Consider potential sampling bias during your evaluation planning and correct for potential biases. If you do identify differences between respondents in your sample and in the target population, make it clear in your reporting who the results represent. © 2009 University of Wisconsin-Extension, Cooperative Extension, Program Development and Evaluation 14 Re-check your ‘sampling’ IQ. What did you learn? What did you already know? What do you need more help with? 1. 2. 3. 4. 5. 6. Why do people use a sample? Does a sample always refer to people? Is there one best way to sample? Should we always sample? What makes a “good” sample? What size does the sample need to be? © 2009 University of Wisconsin-Extension, Cooperative Extension, Program Development and Evaluation 15 Congratulations! We hope this presentation helped you better understand the ins and outs of sampling! © 2009 University of Wisconsin-Extension, Cooperative Extension, Program Development and Evaluation 16