Hi, I wrote a quite extensive chapter on the kind of experimental designs that can and maybe more often than perhaps expected, should be used. Not a research paper, but a chapter for grad students: www.nber.org/papers/w33630

Congrats!!!

Congrats!

yes, that's the problem with making things up, people often overdo it, 20-30% of the effect sizes, and he probably would have still been able to publish it, and maybe noone would have found out...

I wonder how much is driven by immigrants? In many European countries those are non-negligible numbers, no? (E.g. Austria: More than a quarter has migration background). Not saying immigrants don't matter. But Germany may look like US for people without immigration background...

Brave new world is a good starter for such ideas…

In Martinez-Marquina, Niederle and Vespa (2019) we (re-)invented a computational circuit complexity control, what Ryan in his paper calls mirror, to show that only parts of constructing the lotteries corresponding to a problem are computational, though they are far from explaining everything.

I agree, of course some people will also suffer from computational problems, and some less so given how simple the computations are. I think it is a great and valid point that we are sometimes too keen to jump to some (sometimes convoluted) preference explanation rather than acknowledging mistakes.

Having had a look, I don't see how the current version can remain as is. While I hope Ryan Oprea has an answer for many of the points that were raised, I am not sure what that would look like.

I totally agree with these points, both 1. and 2. Here is a third. I think we should abandon our "goal" to have a unified behavioral theory. It won't happen, and I hope that as a field we have matured enough to acknowledge that many forces exist beyond the neoclassical model.

Give me a few examples of what you really like, makes it easier to give recommendations:)