Common Errors in Statistics (and How to Avoid Them), 4th Edition

EXPERIMENTAL DESIGN

If a study is allegedly double-blind, how was the blinding accomplished? Are all potential confounding factors listed and accounted for?

WHAT IS THE SOUND OF ONE HAND CLAPPING?

Gonzales et al. [2001] reported that Maca improved semen parameters in men. A dozen men were treated with Maca. But no matched untreated (control) subjects were studied during the same period. Readers and authors will never know whether the observed effects were due to a change in temperature, a rise in the Peruvian economy, or several dozen other physiological and psychological factors that might have been responsible for the change. (Our explanation for the reported results is that 12 men who normally would have their minds occupied by a score of day-to-day concerns spent far more time than usual thinking about sex and the tests to come. Thus, the reported increase in semen production.)

The big question is not why this article was published with this absence of a control group, but why the human-uses committee at the Universadad Peruna Cayento Heredia in Lima permitted the experiments to go forth in the first place. The tests were invasive—“semen samples were collected by masturbation.” A dozen men were placed at risk and subjected to tests, yet the final results were (predictably) without value.