statistics example: consider other factors

This article about running from today's Washington Post reports that older people (that's over 50) who run regularly report 25% less musculoskeletal pain that others. But is it the running that produces that effect?

There are several potentially confounding factors: People with higher pain thresholds may be more likely to become runners; the running group is less likely to include people with chronic joint pain; and running can produce psychological benefits, including those triggered by the much-ballyhood endorphins, that may result in lower pain reporting.