STS 350 Study Guide - Final Guide: General Linear Model, Null Hypothesis, Explained Variation
23 Mar 2017
School
Department
Course
Professor
Document Summary
Format: 60 multiple choice & 5 short answer. Bivariate & multivariate regression: 16 questions, general theory, basic calculations. Logistical regression: 11 questions, no spss, know probability vs odds, odds ratio, walls stat, theory of logistic regression used to determine. Correlation: how one variable is related to another variable by looking at how it changes. Does not imply causality!!! (correlation does not = causation!) Linearity can be tested using the pearson product moment correlation coefficient. You are looking for a value close to +1 or -1. As the result becomes closer to 0, the less significant of a relationship the variables will have. This is ensuring that the x predictor variable is related to the y criterion variable. Normality is ensuring that the predictor variable and the criterion variable are each normally distributed. This can be tested using a histogram curve to see that there are normal kurtosis and skewness values as well as no outliers.
