e-Articles (Chapter 9)

Here are some full-length research articles in which statistical tests were conducted on correlation coefficients. To view any article, simply click on its title. (NOTE: No claim is made that these articles are perfect in all respects. By carefully reviewing them, you will hone your skills at being able both to decipher and to critique statistically-based research reports.)

Program and School Characteristics Related to Teacher Participation in School Health Promotion

In this research report, Tables 2 and 3 each contain several Pearson correlations. Each of these correlations was tested to see if it was significantly different from 0. Can you tell whether these tests were conducted in a 1-tailed or 2-tailed fashion? In Table 3, can you locate the r that shows up as being nonsignificant when it should have had an asterisk attached to it?

Interpersonal and Individual Factors in the Grandiose Fantasies and Threats to Self-Esteem of a Non-clinical Sample

Illustrates how Spearman rank-order correlations can be tested to determine if they are significantly different from 0. See the last paragraph of the "Data Analysis" portion of the article's "Method" section, the 2nd paragraph of "Results," and Table 2.

Internet Use and Child Development: The Techno-Microsystem

Shows how the hypothesis testing procedure can be used to see if a correlation coefficient is significantly different from zero. Such a test was performed 136 times, with results presented in Tables 3, 4, and 5.

Effects of Supplemental Fish Oil on Resting Metabolic Rate, Body Composition, and Salivary Cortisol in Healthy Adults

In this study, the researchers conducted tests on 6 values of Pearson's product-moment correlation, with each of these tests being two-tailed in nature. The presentation of the rs and the ps shows nicely that the p in such a test gets smaller as the value of r moves away from 0. See the final paragraph of the portion of "Results" dealing with "Salivary Corisol Concentrations." (NOTE: a comparison of the of the ps across the 2 sets of rs requires sample sizes that are equal or nearly equal, as is the case here with nFO = nFO = 22.)

 If you have seen or authored a research report that you think might help others understand tests on bivariate correlation coefficients, please contact me (shuck@utk.edu) and provide a link to what you have found or written. If I post the link on this page of my website, I promise to give you appropriate credit for first seeing/writing the item you share.