I am a huge fan of the working paper culture. As you said, all work must be read carefully and evaluated for how convincing the claims are. So, why ignore a massive amount of information that is available in working papers that have been released.

Big fan of Mellisa Kearney's new book The Two Parent Privilege. Have focused on family structure for years in my Econ of Poverty class, but most books view that issue as tangential when for the U.S. it is likely one of the central issues.

Great thread. If balanced already, only reason for DiD is efficiency. If unbalanced, enter world of strong assumptions. Many top DiD/event study papers have many trend controls beyond baseline 2 way FE that influence outcomes and few show stability of estimates as such controls are added.

By itself, post-registration of RCT is a really weird policy unless the journal requires disclosure of the fact that the pre-analysis plan was post-reg. Otherwise, just misleads readers while making the journal look good for requiring PAP.