Content
No exercises available for this concept.
Spearman's rank correlation coefficient rs measures how close data is to be monotonic - either solely increasing or solely decreasing. Spearman's coefficient is calculated by ranking x- and y-values from least to greatest and performing a linear regression on the ranked lists.
If rs=1, then the data is perfectly monotonic increasing: as one variable increases, the other does too.
If rs=−1, then the data is perfectly monotonic decreasing: as one variable increases, the other decreases.
Spearman's coefficient is useful because it captures all monotonic relationships - unlike Pearson's coefficient, which only measures linear monotonic relationships. Consequently, Spearman's coefficient is not as sensitive to outliers as Pearson's coefficient.