Correlation Flashcards
What is a correlation?
It assesses the degree to which scores on two variables co-relate
What is the correlation coefficient?
r provides an index of the extent to which two variables are linearly correlated
What is a scatterplot?
A graph of respondent’s scores on one variable plotted against their scores on a second variable
What is one problem that might be encountered in a Pearson’s correlation?
Things are measured on different scales so it is difficult to look at how they co-relate if you just look at the raw scores
Need to standardise the distributions (Z scores) - how many SDs above or below the mean any value is
Makes it easier to compare
What do you need from the SPSS output?
r score
p score
df = N - 2
How do you report the results of a correlation?
There was a significant positive correlation between X and Y, r(df) = [r score], p = [p value], such that those with more _____ held more _____.
If there are multiple correlations, report the results in a table
What are the assumptions underlying Pearson’s correlation coefficient?
- The two variables must be measured on interval scales
- The two variables must be linearly related
If violated a non-parametric correlation should be used which uses ranked scores
When can a Spearman’s Rank correlation coefficient be used?
When the scatterplot indicates that the relationship is monotonic, but non-linear (and when there are outliers)
What would happen if a Pearson’s correlation was used on a non-linear relationship?
It would underestimate the strength of the relationship by trying to draw a straight line.
When you rank the scores it creates a linear relationship getting rid of the curves.
What happens in non-monotonic relationships?
The ranked scores would change them into V shapes, but you can’t run correlation on these relationships
How might outliers affect the relationship?
By ranking the scores it brings the outliers in, so that the correlation doesn’t overestimate the strength of the relationship
How would you report a Spearman’s Rho?
You still have to work out df (N - 2)
Due to the presence of two outliers, the strength of the relationship between X and Y was tested using Spearman’s rank correlation coefficient. There was a significant positive correlation, rs(df) = [r value], p = [p value], such that those with more _____ held more _____.
When is a partial correlation used?
When you want to control for the effect of a third variable
It measures the strength and direction of a linear relationship between two continuous variables whilst controlling for the effect of a third continuous variable
How would you report a partial correlation?
A partial correlation was conducted to test the strength of the relationship between X and Y after controlling for the effect of Z. A significant positive correlation was found between X and Y, r(df) = [r value], p = [p value], whilst controlling for Z.
