Talk:Guido Imbens

How exactly was LATE used in the Massachussetts lottery paper?
The article currently reiterates the following claim from a press release:



It's not clear to me how the Local Average Treatment Effect concept was used in this study. It seems that the term isn't even mentioned in the paper (nor does it cite the 1994 Imbens & Angrist paper that introduced the concept). And who exactly are the compliers, always-takers, never-takers and defiers in this situation? Non-compliance in the sense of a lottery winner not accepting their winnings doesn't seem to have been an issue. The paper discusses some other reasons why the selection of lottery winners and "losers" might not be perfectly random (starting with the fact that some players buy more tickets than others), but skimming section 5, it appears that they were handled with different methods, not LATE.

Press releases authored by PR staff (i.e. laypeople) are usually not considered to be the most reliable sources on scientific topics - especially if they might have been written in haste to react to a Nobel announcement within a few hours. But apparently User:Ktin was also drawing from the actual research paper when adding the corresponding text to this article (at least they cited it alongside the press release). Ktin, did you look at where and how the paper uses the LATE?

Regards, HaeB (talk) 12:09, 11 October 2022 (UTC)