COMM 101 Lecture Notes - Lecture 8: Pearson Product-Moment Correlation Coefficient, Observational Error, Weighted Arithmetic Mean

• School: U of M
• Department: Communication Studies
• Course: COMM 101
• Professor: Julia Lippman

Document Summary

Why use composite measures: every measure has error, but the errors differ for different measures, if we minimize both random and systematic error, we can increase both reliability and validity. Why would two things with error be better than one thing with error: football game: Which tells us more about how the wolverines compare to the buckeyes: how well they performed in the rst quarter of the game, how well they performed in all four quarters of the game. Any given quarter of the game is an indicator of overall performance, but putting them together makes a better measure. Take away #1: combining variables can lead to more accurate measurement. 3 basic types of composite measures: direct measures of a concept. Each variable gets at the whole idea: indexes, measures of parts of a complex concept. Each variable gets at some of the idea: indexes, measures of overlapping concepts. Each variable measures too much: typologies.