Pearson's r and Linearity

The most popular technique for assessing the strength of a bivariate relationship is Pearson's product-moment correlation. This correlational procedure works nicely if the two variables have a linear relationship. Pearson's technique does not work well, however, if a curvilinear relationship exists between the two variables. [In the latter situation], Pearson's correlation underestimates the strength of the relationship that is present in the data. Accordingly, you can place more confidence in any correlation coefficient you see when the researcher who presents it indicates that a scatter plot was inspected to see whether the relationship was linear before Pearson's r was used to summarize the nature and strength of the relationship. Conversely, add a few grains of salt to the rs that are thrown your way without any statements concerning the linearity of the data.

(From Chapter 3 of the 6th edition, p. 64)

Copyright © 2012

Schuyler W. Huck
All rights reserved.

| Book Info | Author Info |

Site URL: www.readingstats.com

Top | Site Map
Site Design: John W. Taylor V