Comment by like_any_other

> For example, failing to control for obvious confounders in observational data is likely to produce biased results. If we like the direction of this bias, we can do less adjustment for confounders.

For example, the study showing that having a white doctor increased mortality of black babies didn't correct for birth weight - once that was done, the effect disappeared (and media interest waned): https://www.wsj.com/opinion/justice-jacksons-incredible-stat...