These days, instead of discovering new things through instruments like microscopes, researchers seem disturbingly more likely to infer something's existence from statistics. What’s inferred by such methods can prove embarrassingly ill-founded.

A case in point is the hue and cry that went up in a number of reputable periodicals in the mid 1990s about a purported link between left-handedness and shortened lifespan, with left-handedness being called everything from an effect of prenatal brain trauma to a genetic marker for a variety of infirmities and defects

Eventually somebody took a second look at the data on which all these claims were based, and found a simpler explanation:  the original study drew conclusions about lefties' lifespan from the low proportion of left-handed folks found among older people.  In actuality, the reason fewer old people showed left-hand dominance was that many of them had been forcibly trained as children to use their right hands.