"We find the standard deviation of teacher effects on student height is nearly as large as that for math & reading achievement, raising obvious questions about validity." https://twitter.com/MatthewAKraft/status/1198965387012067328

i.e., researchers predicted variation in student height attributable to teachers and altho they shouldn't find any impact, they did. Which is a red flag for validity issues in teacher effects models more broadly, again.

Which reminds me of this 2014 statement on the use of "value added models" to determine teacher effects by the American Statistical Association. https://amstat.tandfonline.com/doi/full/10.1080/2330443X.2014.956906#.XdvlGlNOlaw

Great thread explaining this recent research from one of the authors! https://twitter.com/spcorcor18/status/1198994835098808320?s=19