You know what’s even worse than spending six months of your life crunching statistics on human mortality? Getting your study picked up by the national press and nobody reads it because you called it “Low Protein Intake Is Associated with a Major Reduction in IGF-1, Cancer, and Overall Mortality in the 65 and Younger but Not Older Population“.
Mass Media to the rescue
Ignore most of the statistics, find one that can be massaged a little bit, and presto! All of a sudden you’ve got “Too much protein in middle age ‘as bad as smoking’.”
Except that’s not really what your study said. It said in some groups there is a statistical correlation between reported protein intake and mortality … assuming the self-reporting is accurate … and allowing for the fact that the definition of “high” vs. “low” protein was defined after-the-fact … and ignoring that the correlation went in the opposite direction for half the study population.
It doesn’t even matter if you say right in the summary that this only points to interesting questions that must be verified in a prospective study. Most reporters don’t know what that is. But they know what will get people to read their story.