Sociology 3306A/B Chapter Notes - Chapter 16: Scatter Plot, Random Assignment, Dependent And Independent Variables

• School: Western University
• Department: Sociology
• Course: Sociology 3306A/B
• Professor: Andrea Willson

Document Summary

Bivariate analysis provides basis as exemplified by regression slope coefficients for predicting values of one variable from values of another. If we want to go beyond prediction to testing causal hypotheses, then bivariate analysis is not enough. Need theoretical assumptions and empirical evidence about direction of evidence and nonspuriousness. Statisti(cid:272)al a(cid:374)al(cid:455)ses do(cid:374)"t pro(cid:448)ide (cid:271)asis for i(cid:374)ferring causal relationships on their own need theoretical model of causal process linking x and y: determine if data is consistent with theory. Direction of influence is esp based on theoretical or pratical knowledge. Experiments direction set by manipulation of iv which comes before measurement of. Natural time order of survey variables determines causal direction. When time order is indeterminate causal direction is determined on basis that one variable is more permanent than the other. Statistical model = simplified picture of reality. Arrow diagrams can indicate direction of relationship.