ERSH 8320_DOS

What is the main purpose of hierarchical regression?

• To find the best-fitting model without theory

• To screen variables based on VIF

• To test competing models based on theory

• To center variables automatically

What is the first step in building a multiple regression model with “few” independent variables?

• Run stepwise regression

• Hypothesize the model using theory

• Remove multicollinearity

• Check for quadratic terms

What does R² represent in regression?

• The predicted values

• The variance in Y explained by X

• The intercept

• The residual

What’s the key difference between partial and semi-partial correlation?

• Partial removes shared variance from both X and Y

• Semi-partial removes from both X and Y

• Partial is always smaller

• They measure the same thing

What does a squared semi-partial correlation tell us?

• Error variance in Y

• Redundancy among predictors

• Strength of multicollinearity

• Unique variance in Y explained by a predictor

What makes a model “second-order”?

• It only includes main effects

• It includes an interaction term

• It includes squared terms and interactions

• It includes control variables

What is a suppressor variable?

• A variable that increases Type I error

• A variable that doesn’t correlate with Y but controls irrelevant variance

• A quadratic term

• A mediator

Why must we be cautious when interpreting t-tests after multiple comparisons?

• They may exaggerate significance

• The alpha level may inflate Type I error

• They always contradict F-tests

• They use semi-partial correlation

What does a VIF greater than 10 typically indicate?

• Multicollinearity

• Good model fit

• High prediction accuracy

• Low residual variance

What is the rule of thumb for Tolerance to detect multicollinearity?

• Tolerance < .25

• Tolerance > .90

• Tolerance < .10

• Tolerance = 1