PSYCH 2220H Lecture Notes - Lecture 17: Multicollinearity, Standard Deviation, Coefficient Of Determination

• School: Ohio State University
• Department: Psychology
• Course: PSYCH 2220H
• Professor: Cravens- Brown Lisa Marie

Document Summary

Model summary box: r = multiple correlation. Relationship between combo the y and the combination of all 3 predictors (x: r square = proportionate reduction of error. How much more accurate regression line is than if just used mean. Amount of variability accounted for in y by a combination of all 3 predictors: adjusted r square. Compensates for model complexity to provide a more fair comparison of model performance better than r square. Can compare a predictor model to a 4 predictor model more easily: standard error of the estimate. Average distance of data points from the prediction line. B: df tells how many subjects in analysis, ignore mean squares, f and sig. Tells us whether the combo of predictors significantly predicts the outcome variable. Cannot tell which is best predictor b/c cannot compare raw coefficients (not same scale) Values change based on other predictors in model.