The transitivity misconception of Pearson's correlation coefficient

Although correlational studies are of great relevance for most research domains, many students and researchers alike hold misconceptions concerning the interpretation of the Pearson product-moment correlation coefficient. One of these, the so-called transitivity misconception, has not been documented in a systematic way so far. This study provides the first empirical evidence from 279 university students and tests the effect of moderating factors (size of correlations, type of representation, and contextualization). In addition, the relation between the presence/absence of the misconception and the level of confidence in the answer is studied. Almost fifty percent of the participants presented the misconception. Moreover, we found that the occurrence of the misconception depends on the size of the correlations and the type of representation.

Publication year:2008