The correlation coefficient can be larger than 1 (r>1).

[DreamSciTech]

In recent discussion, the correlation coefficient has been mentioned several times. This led me to remember a buried article published years ago in Eur J Drug Metab Pharmacokinet 2003; 28: 85-92, http://www.ncbi.nlm.nih.gov/entrez/quer ... med_docsum.

This article, entitled “An extremely strange observation on the equations for calculation of correlation coefficient”, shows that the correlation coefficient can be larger than 1. A reprint is available by emailing [email protected].