On Data Colada

Lessig

--

Three days ago, I wrote an essay about the accusations raised against Francesca Gino. Based on my judgment about who Francesca is, and evidence about the nature of the evidence against her, I suggested people reserve judgment about whether the charge made against her — that she fabricated or manipulated data in four of her research papers—was true. It’s my belief that it is not true.

I also commented directly in that essay about the Harvard Business School, and about the way it had handled the charges against Gino. That way, I said, was wrong. I’m happy to see others begin to raise similar questions. See, e.g., Stephen Carter (paywalled), John Byrne (Poets & Quants).

In this short essay, I want to add a word about Data Colada. I like Data Colada. I love the idea of people testing the claims of academics regularly and systematically. In the decade since they launched, DC has produced critiques of academic and non-academic claims and practices, using empirical tools to test truth. Their contribution is enormous.

In Gino’s case, Data Colada was clear that they had identified problems in the data supporting at least four of Gino’s papers. Those problems, they asserted, were “evidence of fraud.” Yet as they have rightly acknowledged, they have no way of knowing who was responsible for the manipulations they identified. And though the site states that it shares its analysis with authors before publishing (“When discussing research by other authors we share ahead of time a draft with them and ask for feedback.”), it is my understanding that DC did not share these four posts with Gino before publishing them to the public. Harvard may well have been responsible for that omission.

Gino’s lawsuit, however, names DC’s principals as defendants. That, in my view, is a mistake. The world needs more DC’s, not fewer, and the burdens of litigation are overwhelming for everyone, especially non-profit and non-commercial sites on the web. I do believe that anyone, for-profit or non-profit, has an obligation to correct mistakes or false statements once noticed. That failure, I believe (and my actions evince), is a distinct kind of wrong in the age of perpetual publication. But I have no reason to doubt that DC would correct any errors once shown.

I am hopeful Gino’s lawyers will drop their claims against Data Colada. DC may well have triggered the burdens Gino is now suffering. They are not the (legal) cause.

--

--