standard multiple linear regressions

Question 1

what is tabachnick and Fidell’s 2007 criteria?

Answer 1

for sample size
sample size = <50+M8 (M being the number of predictor variables in the study)

Question 2

Multicollinearity assumptions

Answer 2

Tolerance values need to be greater than 0.5 and VIF values need to be lower than 10

There shouldnt be correlations between predictors (correlations need to be smaller than +/- 0.9)

Question 3

homoscedacity assumption

Answer 3

must look at the scatterplot-should be evenly spread (homoscediacity is there if there is for example large numbers on the left hand side)

Question 4

assumption of normality

Answer 4

check histogram with for a normal distribution

check p plot of standardised residuals- the residuals should fall close to the line on the standards residuals p-p plot

Question 5

REPORTING RESULTS -

Answer 5

look at model summary R2 tells us % if variants are explained by predictor variables. So if R2 = 0.026, we can say 2.6.% of the DV can be explained by the IV. Report the adjusted R2 as well!

USE ANOVA to report the results= f (df (regression, df(residual)= F score, p< sig score.
EXAMPLE= F(2,421)= 4.78, p<0.013
Report if this is significant or not

Use coefficient table to report which IV is significant, and which is more significant than the other IV (using sig scores). To formally write each one you write IV= standardised coefficients beta score, p< sig score.
EXAMPLE= Extraversion= 0.156, p<0.0017

Question 6

Regression calculations!

Answer 6

use coefficient table- use the B column

Y= constant + Iv B score (Iv name), Iv B score (Iv name) + error