When the underlying distribution of interval scores is violently skewed, the appropriate correlational technique should be
A) the Pearson r
B) the t test
C) the Spearman r
D) none of these, since correlations cannot be computed for skewed

Question

Quizplus · Accepted Answer

The appropriate correlational technique for violently skewed interval scores is the Spearman r, which measures the strength and direction of a monotonic relationship between two variables, regardless of the shape of their distribution. This is because the Spearman r is based on the ranks of the data rather than their exact values, and is therefore more robust to outliers and non-normality. The Pearson r, on the other hand, assumes a linear relationship between normally distributed variables and can be biased by extreme scores. The t test is not a correlation technique, but a hypothesis test for a single sample or two independent samples.

When the Underlying Distribution of Interval Scores Is Violently Skewed