In order to conduct a clinical trial properly, scientists need to be able to minimize interference from confounding variables, things that influence final measurements, without being the variable of interest.
The effect was small but it remained significant even when the findings were adjusted to take into account confounding factors like years of education and socioeconomic status.
Laszlo Garamszegi, of the Do? ana Biological Station in Spain, studied patterns of malarial infections in birds to avoid confounding human factors in determining the epidemiology of a closely related parasite.