SOCIOL 5 Lecture Notes - Lecture 12: Quota Sampling, Level Of Measurement, Sampling Distribution
Document Summary
Describes variation in dv explained by iv. Lambda = 0: knowing distribution of iv is no help at all in estimating value of dv, compared to estimating dv solely on basis of its own frequency distribution. Lambda = 1: knowing distribution of iv makes it possible to perfectly predict dv. Lambda increasing, association between iv and dv, and your ability to predict dv based on iv, also increasing. Lambda = . 16: knowing iv reduces errors in guessing dv by 16% Describe percentage in variation in dv explained by iv. Number of concordant pairs - pairs of observations ranked in same order of both variables. Number of discordant pairs - ranked in different order of 2 different variables. Ex. american opinions on aca = pull first ten people you see (not a good approach but a nonprobability sample with some bias)