Correlations

Question 1

What are correlations?

Answer 1

Correlations establish how two or more co-variables are related.
‘If something happens to X what happens to Y’

Question 2

How do we measure correlations?

Answer 2

1. Scatterplots; these visually show us the relationship between X and Y.
2. Correlation coefficient; (Pearson’s r) is a value between -1 to 1, which tells us the information of both the strength and the direction of the relationship.
3. p-value; tells us whether the relationship is significant (p < 05) or not (0 > .05).

Question 3

Describe a positive correlation.

Answer 3

When X increases, Y also increases.

r = 1

Question 4

Describe a negative correlation.

Answer 4

When X increases, Y decreases.

r = -1

Question 5

Describe a null correlation.

Answer 5

When X increases, Y shows no consistent change (there is no relationship).
r = 0

Question 6

Describe the strengths of correlations.

Answer 6

Cohen, 1988
Small: r > .10
Medium: r > .30
Large: r > .50
(ignore the +/- as they measure direction not strength).

Question 7

What is mediation?

Answer 7

This is where the relationship between an IV and a DV can be explained by their relationship to a third variable (the mediator).
It represents a ‘bridge’ and explains how/why the predictor influences the outcome.

Question 8

What is the difference between full and partial mediation?

Answer 8

Full mediation is where the direct is no longer significant (the mediator accounts for the full effect).
Partial mediation is where the direct effect remains significant (the mediator does not account for the full effect).

Question 9

What is moderation?

Answer 9

This is the combined effect of two variables on an outcome variable.
It represents a ‘switch’ and exxplains when a predictor influences the outcome and when not (interaction).
It affects the strength and direction of impact on Y.

Question 10

How do we report the correlation?

Answer 10

Report ‘r’ to 2 decimal palces, if p = .000 then report p < .001 and if p is not significant then report p > .05.

There was a significant, (if sig., then strength and direction) relationship between xx and yy.
OR
The relationship between xx and xx was not significant, r =
.xx, p = xx.
As xx increases, xx (increases, decreases, shows no consistent change).