What is a correlation?
The relationship between variables
What is a positive relationship?
When if one variable scores high the other will too
What is a negative relationship?
If one variable is scored low the other will too
What do correlations aim to do?
Test the strength of a relationship
What statistical test is generally used for correlation?
Pearson’s Product Momentum Correlation Coefficient
What is Pearson’s Product Momentum Correlation Coefficient?
A statistic with a range of 1 to -1
What two things does Pearson’s r tell us about the relationship between variables?
The strength of the relationship
The direction of the relationship
It can also determine whether it is significant or not
What does a correlation coefficient give you?
A measure of the strength of the relationship
How much the two variables vary together
What does a correlation coefficient score mean the further it gets from zero?
The stronger the relationship
What does +_ .1 mean compared to +_.5?
This is how big the relationship is, generally .1 is a small effect and .5 is large effect
What is the null hypothesis for correlation?
That r=0
Meaning there is no relationship between variables
What is an important thing to remember with correlational data?
Strength, direction and significance of relationship are all importand
What is the most important thing to remember about correlation?
Correlation does not equal causation
Correlation coefficients give no indication of direction of causality
For what reasons can we not assess causality from correlation?
Direction
Third Variable
Why is direction an issue in determining causation in correlations?
We don’t know which variable comes first
Why is third variables an issue in determining causation in correlations?
Correlation could be not because the two variables are related but because they are both related to a third variable
What are the assumptions of correlations?
Linearity
Normality
Why is the third variable problem an issue in determining causality with correlations?
The relationship between two variables is determined by each variables relationship to a third variable
The relationship is only present due to this and without it the relationship is gone
What are non-parametric methods to correlations?
Spearman’s correlation co efficient (large data set)
Kendall’s T (small data set)
Why is spearman’s cc a good alternative to correlations?
No normality assumptions
