December hasn't been a great month for Harvard President Claudine Gay — not that October and November reflected particularly well on the academic either. The newly minted Ivy League president fumbled the initial reaction and was slow to condemn the grossly antisemitic statement issued by several Harvard student organizations in the immediate aftermath of Hamas' horrific October 7 attack on Israel. Her attempts to tidy that reaction up weren't impressive. 

Then, she testified before Congress. And if there was any benefit of the doubt previously granted Gay, she managed to squelch that with her ridiculous response about the necessity of context to determine whether calling for the genocide of Jews violates Harvard's code of conduct. Since her disastrous visit to Capitol Hill, the academic has faced heightened scrutiny over her qualifications and body of work. 

READ MORE:

REPORT: Harvard President Claudine Gay Plagiarized Portions of Her Ph.D. Dissertation

Scholars Find Dozens More Examples of Plagiarism by Harvard President Claudine Gay

Now Even CNN Is Calling Out Harvard President Gay's Plagiarism Problem

Now, even more questions are being raised regarding Gay's scholarship. In addition to charges of plagiarism, Gay is being challenged over the analytical methods she employed in support of her 2001 PhD thesis and a related paper which helped secure tenure for Gay at Stanford University in 2005. 

The harsh spotlight has spread to her academic record, with accusations of plagiarism - and on Tuesday, a data scientist challenged her analytical methods. It was then revealed that she had refused to share her data, raising eyebrows in academia.

Jonatan Pallesen, a Copenhagen-based data scientist working for the Confederation of Danish Industry, tweeted that he had examined her use of data in her PhD thesis, and a 2001 American Political Science Review (APSR) paper.

The 2001 paper -  'The Effect of Black Congressional Representation on Political Participation' - was one of four peer-reviewed political science articles which secured her 2005 tenure at Stanford.

I am not a political scientist, but I am pondering about the whole approach of her study.

The claim is that having a Black representative in a district causes lower White voter turnout. But what is being measured in the study are correlations between the two.

Having a Black…

— Jonatan Pallesen (@jonatanpallesen) December 26, 2023

In a two-part tweet, Copenhagen-based data scientist Jonatan Pallesen observed: 

Claudine Gay wrote two times about the hypothesis that Black representation causes lower voting turnout among Whites. 

There are odd differences between the two works.

Originally she looked at it in her PhD thesis, and then she redid the analysis in a 2001 APSR paper. In the thesis, the change in White turnout looks like the image below. 

At the bottom of the table, there are three data points which go against the hypothesis, and instead have larger White turnout in the districts with Black representatives.

But then in her APSR paper:

Illinois disappeared! Why were the results from Illinois not included, when they were included in the PhD thesis?

If we look again at the table for the results from the thesis, two out of the three results that go against the hypothesis are in Illinois.

The ASPR paper doesn't mention any specific reason for why Illinois was excluded. All it says on the topic is this:

I do not find this to be strong reasoning. The correct scientific approach is quite obviously to include all 21 states. Perhaps it was a lot of work to get the data, and this was not practically feasible. But we know that she had already collected the data for Illinois, so that reason doesn't hold in this case.

I am not a political scientist, but I am pondering about the whole approach of her study.

The claim is that having a Black representative in a district causes lower White voter turnout. But what is being measured in the study are correlations between the two.

Having a Black representative is not a random event. It possibly happens more often in inner city districts, for example. And maybe inner cities have lower vote participation. This would result in a correlation without any causation. There are countless such possible factors.

It would make more sense to me to compare elections turnouts in the same district, in two different election years, where one is with a Black representative, and the other is without. Then you are holding most things constant except the variable of interest.

Yet Gay has refused to share the data ostensibly supporting her conclusions. 

Christopher Brunet, a contributing editor at The American Conservative, then followed up on Pallesen's concerns in The Dossier, and found that Gay had refused to share the data which informed her conclusions.

Two professors - Michael C. Herron, a quantitative social science professor at Dartmouth, and Kenneth W. Shotts, a professor of political economy at Stanford Graduate School of Business - told a 2002 conference of The Society for Political Methodology that she would not share her data or code with them.

'We were, however, unable to scrutinize Gay's results because she would not release her dataset to us (personal communication with Claudine Gay, 2002),' they noted.

The professors said the statistical practice Gay used in her research often leads  'logical inconsistencies.'

But, Brunet finds, their criticism of Gay was removed from the website.

Brunet said that Gay's refusal to share her data was 'shameful'.

Thus far, Gay has resisted calls for her to resign, unlike U Penn President Liz Magill, who stepped down following the intense backlash to her disturbing testimony (even without multiple claims of plagiarism and academic dishonesty). She's even been defended by none other than former President Barack Obama. Whether this latest knock on her credentials will prompt her (or Harvard) to reconsider remains to be seen.