Harvard Teaching Hospital To Retract Papers As Falsified Data Scandal Unfolds

Weeks after Harvard President Claudine Gay resigned amid a plagiarism scandal, the university’s teaching hospital is set to retract or correct dozens of papers authored by four of their top researchers, following claims of data falsification.

Six papers have already been selected for retraction, and 31 others are in the process of being corrected by the Dana-Farber Cancer Institute in Boston, according to the Harvard Crimson, citing the hospital’s research integrity officer, Dr. Barrett Rollins.

The corrections come amid claims of data manipulation against DFCI President and CEO Laurie H. Glimcher ’72, Executive Vice President and COO William C. Hahn ’87, Senior Vice President for Experimental Medicine Irene M. Ghobrial, and Harvard Medical School professor Kenneth C. Anderson. The allegations of misconduct were first compiled and publicized in a Jan. 2 blog post by data sleuth Sholto David.

In the statement, Rollins wrote that David contacted DFCI with allegations of data manipulation in 57 manuscripts. According to Rollins, 38 were articles in which DFCI researchers “have primary responsibility for the potential data errors.”

Four of the researchers at the Dana-Faber center have faculty appointments with the Harvard Medical School.

The data manipulation claims came to light after investigator Sholto David published a blog post several weeks ago alleging irregularities among 57 papers. He had previously alleged “data forgery” – which included discrepancies in images consisting of duplications of blots, bands and plots, according to the report.

David, who holds a doctoral degree in biology from Newcastle University, alleged that three papers authored by Glimcher, 12 by Hahn, 10 by Ghobrial, and 16 by Anderson contained “data forgery,” including five co-authored by both Anderson and Ghobrial. As is typical for scientific research, all of the papers referenced by David have several co-authors, though his post focused on the four DFCI researchers.

The papers, published between 1999 and 2017, most commonly have duplications of blots, bands, and plots within images, David alleged. In a Saturday interview, David said he used a combination of artificial intelligence image analysis software ImageTwin and manual detection to look for errors in the papers. -Harvard Crimson.

In short, AI is going to absolutely wreck a lot of ivory tower copycats…