SOCIOL 5 Lecture Notes - Lecture 14: Internal Validity, Statistic, Peer Pressure
Document Summary
Think how to percentage the table for bivariate analysis. Table should always be percentaged within each category of independent variable. Bivariate inferential statistic: testing hypothses about the population. Statistically significant relationship b/n variables or is apparent difference due to random sampling error. Something statistically significant if we can show with confidence that chance of no relationship b/n variables is small. Goal is to evaluate whether iv affects the dv, while holding constant other potential explanatory variables or controlling for their effects. Extends the logic of bivariate analysis with more variables to better identify possible causal relationships - rule out spurious causation. Peers have powerful effect on individual behaviour. Ask them about behaviour and infer peer pressure. Observe group behaviour (wave at a game, chants at a protest) Examine diffusion f behaviour in a network (facebook data on pictures of cats) Informational conformity (conforming to group due to the informer) Focus attention on few variables and holding others constant.