When deciding whether and how to override an algorithm’s recommendations, we hypothesize that people are biased towards following naïve advice weighting (NAW) behavior: they take a weighted average ...