# Regression How to write up

What does multiple linear regression predict?
DV based on IVs
1 of 27
What are the advantages of stepwise?
It is useful in exploratory phase of research
2 of 27
What are the advantages of hierarchical?
Allows the researcher to validate the hypotheses used to determine the order of predictors
3 of 27
What is a disadvantage of stepwise?
It is greatly influenced by random variation and rarely gives replicable results
4 of 27
What table is given after?
A table of Mean and SDs
5 of 27
Why are Pre Analyssi checks done?
Pre analysis checks were conducted before the study, which showed that none of the variables deviated from normal
6 of 27
Sample size for medium size effect?
N>50 + 8 * M
7 of 27
For a small effect size
N> (8/f2) + (m - 1)
8 of 27
What is the sample size for step wise regression?
N>40*m
9 of 27
How do you work out cohens F?
R2/1-R2-R2
10 of 27
What is cohens F?
This checks the relation between number of cases and number of predictors
11 of 27
What are the values for Cohens F?
F2 = 0.02, F2= 0.15, F2 = 0.35
12 of 27
What would you say about multi-collinearity and singularity?
‘There was no evidence of singularity (all tolerances greater than .1) or multicollinearity (all r’s less than .9 and VIF<10).
13 of 27
Why do we not want singularities?
They present mathematical problems because we dont want to measure the same thing twice
14 of 27
What happens if we have multi-collinearity?
We should compute correlations amongst IVs and remove appropriate IV
15 of 27
What is singularity?
Variable is redundant (Combination of two or more variables)
16 of 27
What do we says about outliers?
Cooks distance was always less than 1
17 of 27
Heteroscedasity, normality?
The histogram, normal pp plot of regression standardised residuals and scatter plot shows normality of residuals. This shows there was no heteroscedasity
18 of 27
How do you start a model summary?
A significant regression equation was found with an R2 = ..., explaining .... % of variance
19 of 27
What else is reported?
Adjusted R 2 and R 2, Beta values
20 of 27
What does Adjusted R2 show?
Loss of predctive power: Adjusted R2 = How much variance in the outcome would be accounted for if the model had been derived from the population rather than from the ample
21 of 27
In stepwise regression, what does SPSS search for?
the best predictor and enters it into the model first. It will continue doing this until the predictor does no longer significantly improve the fit of the model, at which point it will reject the predictor and stop the process.
22 of 27
What do we report for ANOVA?
• ‘A within-subjects ANOVA was conducted. A significant regression equation was found (F(__)=__, p<__), therefore explaining the variance at a level that was higher than predicted by chance.’
23 of 27
What does the F Value show?
• ‘TheF value shows by how much the model has improved the prediction of the outcome – if p<.05 this improvement was significant.’
24 of 27
What does Beta show?
• One unit increase in (one IV) corresponds to (Beta) units increase in (DV). -Say for all significant ones. ‘This indicates that (specify IV’s) all account for a significant amount of variance in the model.
25 of 27
Why is Beta better than B values?
’The beta value is a standardized value which represents the strength of the relationship between the DV and predictor variables
26 of 27
What does the B value represent?
The b value represents the slope of the line between the DV and predictor variable, however is not standardized and as such can’t be compared between predictor variables
27 of 27

## Other cards in this set

### Card 2

#### Front

What are the advantages of stepwise?

#### Back

It is useful in exploratory phase of research

### Card 3

#### Front

What are the advantages of hierarchical?

#### Back ### Card 4

#### Front

What is a disadvantage of stepwise?

#### Back ### Card 5

#### Front

What table is given after?

#### Back 