Audience: closed class
Presentation for Prof. Peter Bentler’s class. Subjects covered will include Data Archive services and demonstration on finding data in public archives. The emphasis will be on helping students find material for their final project.
Students need to find, and download, a data set of individual respondent responses on questionnaires, interviews, tests, etc. to analyze. There are several requirements:
Students have to choose some variable or construct to focus on. This could be a personality trait (extroversion, impulsivity, orderliness etc.), a psychopathology trait (anxiety, depression, neuroticism etc.), a social or political attitude (favorability toward abortion, opinion about Obamacare, endorsement of a presidential candidate, attitude about nuclear power plants, etc.) a health-related behavior (amount of exercise, extent of cigarette smoking, blood pressure, etc.). Anything where one could think people vary along a continuum (high to low, for vs against, agree vs disagree).
This is the key part. Students have to find a data set where there are at least 15-20 items (questionnaire items, interview questions, etc.) that deal with their chosen construct. Each of those items is presumably measuring a somewhat different aspect of the construct, or asking about it in a different way. If only 12-13 items are available, that might do; 10 or less, definitely no.
Does not matter, except that the data should be at the individual level (not household, or census, or other grouping). It could be a random sample of US population, a sample of seniors over 70, college students, mothers of small children – really, anything will do.
Number of respondents should be anywhere from approx. 200 to 10,000. Less than 200 makes the results unstable, more than 10,000 is just too huge to work with easily.
Download could be in various formats, probably being relevant to a student’s favorite statistical program.
Once the data is downloaded, students can use statistical programs to do recoding (if necessary) and analyses as described in class.