Underneath most scientific findings may lie a hidden 'universe of uncertainty'

Researchers looking at the same data set in social science produce wildly different results, and very little of that variability can be explained by bias.