Point-by-Point Response to @sndurlauf's comments
2) It is not clear why Steve is discussing county data. Precincts are much smaller units. Fulton County had 384 precincts in 2020 (404 in 2016, but consolidated some because of the Coronavirus). By looking at adjacent precincts...
2) It is not clear why Steve is discussing county data. Precincts are much smaller units. Fulton County had 384 precincts in 2020 (404 in 2016, but consolidated some because of the Coronavirus). By looking at adjacent precincts...
https://twitter.com/sndurlauf/status/1344342593241554945
...across the street from each other, I compare places that are much more similar to each other than county comparisons. In addition, controlling for the difference in Trump’s share of the in-person vote & the difference in race & gender demographics, you are picking up a...
...lot of any remaining differences between the precincts.
3) If Steve has a suggestion for a more “imaginative” identification strategy, but I think that my approach is a pretty imaginative.
3) If Steve has a suggestion for a more “imaginative” identification strategy, but I think that my approach is a pretty imaginative.
4 or 5) Why wouldn’t you expect two precincts adjacent to each other and have similar voting patterns and demographics to be similar? Steve suggests that “politicization” would make the two precincts differ in that dimension. But why wouldn’t factors such as the difference...
...in Trump’s share of the in-person vote and racial and gender demographics pick this up? Why would two similar precincts respond differently to the politicization that is equally available to both?
6) No Voting Model? The model is very simple. If claims of vote fraud are true, we should see certain patterns in terms of either the differences between absentee or provision ballots & in-person votes or in terms of voter turnout rates.
7) Steve suggests that comparisons with earlier elections are useless because “exchangeability requirements” are not met because Trump was a candidate. However, Trump was a candidate in 2016, making this election comparable to 2016 in at least one dimension. ...
...Are there new voters in the precincts? Sure, but if 5% or 10% of the voters are new, that still means 90% or 95% of the voters are the same, & some comparisons can be made. Every election has candidates with differing voter perceptions, which can be considered random. ...
... This randomness means that that exchangeability requirements are met. Public choice has lots of papers that make comparisons across elections.
8) "grotesque irresponsibility of releasing the paper under guise of scholarship"
This is not a substantive claim.
This is not a substantive claim.
9&10) Comparison to wage discrimination.
The analogy is specious. The paper takes adjacent precincts in different counties, where one county is alleged to have voter fraud. If there is no voter fraud, one might expect similar behavior. If there is voter fraud, one can expect...
The analogy is specious. The paper takes adjacent precincts in different counties, where one county is alleged to have voter fraud. If there is no voter fraud, one might expect similar behavior. If there is voter fraud, one can expect...
... predictable unusual behavior in the precinct whose only characteristic is that it is in the county with potential fraud. The predicted unusual behavior was found.
Others who responded to @sndurlauf's posts are cordially invited to respond here. While Steve has strong feelings on this, I don't think his comments were useful.
@KarlOakes1 @RichardHanania @GarettJones @jadler1969
@NAChristakis @pnikolov @indigenalysis @Chris_Auld @JohnMullahy
@KarlOakes1 @RichardHanania @GarettJones @jadler1969
@NAChristakis @pnikolov @indigenalysis @Chris_Auld @JohnMullahy
A copy of my paper is available here so that people can check out my points.
papers.ssrn.com/sol3/papers.cf…
papers.ssrn.com/sol3/papers.cf…
• • •
Missing some Tweet in this thread? You can try to
force a refresh
