Simple and Multiple Regression

PSYC234

What is regression?

We want to predict one continuous outcome (DV)

Does whole model work?

ANOVA (F and p-value)

How much variance is explained?

R squared

Is this predictor significant?

F and p-value

What direction is the relationship?

positive/negative B

Which predictor is strongest?

Standarised Beta

Is the estimate precise?

SE

Were assumptions, including multicollineraity met?

VIF = 1 (Excellenent)

VIF <5 = Acceptable

VIF >10 = Problem

Write up the regression model (4 marks)

1. Find DV

2. Variance (R squared)

3. Significant? (F and p value)

Write up individual predictors in continous prose? (14 marks)

1. Report positive/negative B, SE, p, CI.

2. Significant predictor = There was +/- association between X and Y. (B =, SE =, p value, CI)

3. Non-significant predictor = There was no significant association between X and Y.

Are there any variables causing problems?

VIF values

How to know if confidence intervals are significant?

No zero inbetween = Significant

Zero inbetween = Non-significant