This is the title of a new article by Philip Stark posted on arXiv today. I’ve only skimmed it so far but the point he’s making aligns with the comment I made in the Week 01a lecture that statisticians often don’t take model uncertainty all that seriously and that there are negative consequences for this.