CRIM 320 Lecture Notes - Lecture 6: Situation Two, Level Of Measurement, Negative Relationship
Document Summary
Expected: values we would expect if variable 1 and 2 are independent. Tell us if variables are independent of each other. Difference of zero means independent limitations of chi-square. Expected cell counts must be greater than 5. Relationship between political party and whether gangs are a. Smaller the sample is, more categories there are, much more difficult to find significant results, whether variables are independent of each other lambda. Like chi-square, is measure of association for nominal data. Instead of association, lambda treats one variable as dependent and the other as independent. Unlike chi-square (and its extensions), it is a pre (proportional reduction error) measure proportional reduction in error (pre) The proportion of prediction error that can be reduced by knowing the independent variable. If (cid:449)e (cid:272)a(cid:374) (cid:862)predi(cid:272)t(cid:863) depe(cid:374)de(cid:374)t (cid:448)aria(cid:271)le, (cid:449)e are, i(cid:374) a statisti(cid:272)al se(cid:374)se, (cid:862)e(cid:454)plai(cid:374)i(cid:374)g(cid:863) it. Being able to account for why people have different responses on outcome lambda ( ) as a pre measure.