Why does experimental research have to have random samples? What’s the advantage or reason? What mistake did RA Fisher make in his logic many years ago that has misled researchers ever since? What type of sampling was used in the description of Geraldine Ferraro’s observation that she would probably win? What type of sampling is used in comparing states by SAT scores and what is the basic problem with Terrel Bell’s chart? Why did “fixing” the problem statistically so that SAT’s could be compared not work? What was wrong with the method?
Thinking Target: Analysis, Comprehensive
Remember that sampling can be described from best to worst as:
- Random – every subject has an equal chance of being chosen.
- Stratified- sample is proportional to the populating in characteristics.
- Systematic-every nth subject is taken from the population.
- Cluster-naturally occurring groups (convenience)
- Volunteer-the subject chooses to participate or not.
Pay close attention to the parts highlighted in the article and answer the following questions.
- Why does experimental research have to have random samples? What’s the advantage or reason?
- What mistake did RA Fisher make in his logic many years ago that has misled researchers ever since?
- What type of sampling was used in the description of Geraldine Ferraro’s observation that she would probably win?
- What type of sampling is used in comparing states by SAT scores and what is the basic problem with Terrel Bell’s chart?
- Why did “fixing” the problem statistically so that SAT’s could be compared not work? What was wrong with the method?
- What is the point of the example of bullet holes and Abraham Wald as advice for better sampling?
- What is “nonignorable nonresponse” and why is that one of the most important concepts you can remember to tell if research is making good claims?
- What is the only real solution to solving research problems that would like to test hypotheses about causality?