Which non-parametric test is used for assessing correlation?

• A. Spearman's
• B. Pearson's
• C. Kendall's tau
• D. Chi-square

Answer: (A)

Correlation can be assessed without assuming a normal distribution by using a non-parametric, rank-based measure. Spearman's correlation does this by ranking the data for each variable and then assessing how well the ranks align. It captures monotonic relationships—situations where as one variable increases, the other tends to increase or decrease, but not necessarily in a straight line. Because it relies on ranks rather than raw values, it doesn’t require normality or linearity, and it’s robust to outliers. This makes it the typical non-parametric alternative to Pearson's correlation, which assumes linearity and normal distribution of the variables.

Kendall's tau is another non-parametric option that uses concordant and discordant pairs, often with good interpretation and sometimes greater robustness in small samples, but it can be more conservative and less familiar to some students.

Chi-square tests are designed for categorical data to test independence, not for measuring correlation between continuous variables.

So the best non-parametric approach for assessing correlation here is Spearman's correlation.