Residual confounding

A statistical bias related to unmeasured or insufficiently controlled factors that may distort the results of an observational study. In soy studies, women who consume a lot of isoflavones often have a globally healthier lifestyle, which may bias results.