01:830:338 Lecture Notes - Lecture 2: Predictive Validity, Confounding, Normal Distribution
Document Summary
It is difficult to measure when there are so many different methods, approaches, etc: s-data: self-report (specifically asking the person about themselves) S-data and o-data agree with a lot of ideas. Allows respondents to structure their answers freely (ex: twenty statements. Test: i am ____ nice, tired, etc) woman who are married say i am a wife ; Students in college talk about their social group. The things that you say first are the most essential to you. (how many times do students talk about their social group to describe themselves?) Structured: close-ended (t/f, rate 1-5, interview, questionnaire, survey, etc) Requires respondents to answer within an imposed structure (ex: does this trait describe you?) Likert rating scale: typically 5-7 response alternatives with the end-points labeled to define the extremes. Coding doesn"t come an issue because there are numbers. Neo pi: the most common to measure the big five - i like most people i meet (1-5 likert scale)