What is correlation?

Correlation refers to the relationship between two variables.

The correlation can be:

1. Positive - This is when both variables increase or decrease together.
2. Negative - This is when one variable increases as the other decreases.
3. Non-existent - This is when there is no clear relationship, indicated by a correlation coefficient close to 0.

A correlation coefficient of 1 indicates a perfect, positive linear correlation, where all points lie on a straight line.

Assessing correlation using Spearman's rank correlation

Spearman’s rank correlation coefficient is used to measure the strength and direction of association between two continuous variables that are not normally distributed.

The process for calculating Spearman's rank correlation is:

1. Convert the raw data values of each variable into ranks from the smallest value to the largest value.
2. If two values are the same for one variable, give them an average rank (e.g. if the two smallest values are 6 mm, they would both have the rank 1.5, and the next smallest value would have the rank 3).
3. Determine the difference in ranks $(d)$ for each pair of values.
4. Square these rank differences $(d^2)$.
5. Sum up all the squared rank differences.
6. Calculate Spearman's rho $(\rho )$.

Use the formula (this will be provided in an exam):

$\rho =1-\frac{6\sum d^2}{n(n^2-1)}$

Where:

• $\rho =\text{ Spearman’s rank correlation coefficient}$
• $\sum d^2=\text{ sum of the squared differences of the ranks}$
• $n=\text{ number of pairs of data}$

You will most likely be given critical value tables that use the sample size, n, for this test, rather than needing to calculate degrees of freedom.

Then, compare $\rho$ to a critical value at 5% significance level to determine the significance of the correlation:

• A $\rho$ value near +1 indicates a strong positive correlation.
• A $\rho$ value near -1 suggests a strong negative correlation.
• A $\rho$ value around 0 implies there is no correlation.

It can be helpful to carry out the calculations for steps 1 to 4 in a table, for example:

Worked example - Correlation of study time and test scores using Spearman's rank

The table below shows the study time and test scores of 10 students. The critical value at the 5% level for n = 10 is 0.600.

Using Spearman's rank correlation coefficient, assess whether there is a significant correlation between study time and test scores.

Step 1: Equation

$\rho =1-\frac{6\sum d^2}{n(n^2-1)}$

Step 2: Rank the data to calculate d²

rank the data for both variables from smallest to largest, using the average rank when two values are identical, and then calculate the rank difference (d) and d²

Step 3: Substitution and correct evaluation

$\rho =1-\frac{6\sum d^2}{n(n^2-1)}$

$\text{sum of the squared differences of ranks }(\sum d^2)=4$

$n=10$

$\rho =1-\frac{6\times 4}{10({10}^2-1)}$

$\rho =1-\frac{24}{990}$

$\rho =1-0.0242$

$\rho =0.976$

Step 4: Determine significance and interpret result

the significance of the correlation coefficient is determined by comparing it to a critical value at a chosen significance level

the critical value at the 5% level for n = 10 is 0.648, so the correlation coefficient of 0.976 is much larger than the critical value

this means the correlation is significant at the 5% level, indicating a very strong positive correlation between study time and test scores that is unlikely to be due to chance