Thread by @chris_de92

My first publication also deserves my first proper tweet. Our paper on economic incentives and p-hacking in clinical trials just came out in @PNASNews
pnas.org/content/early/… (1/6)

We analyze p-values from nearly 5000 phase II and phase III drug trials reported to ClinicalTrials.gov, the largest registry of clinical trials. Do financial conflicts of interest undermine ethical standards and incentivize manipulation of results? (2/6)

Unlike previous studies of academic publications in many disciplines (including economics), we don’t find patterns suggestive of p-hacking--reassuring in times where the whole world looks hopefully at trials for #COVID__19 medications/vaccines. (3/6)

However, selective continuation of only the promising phase II trials cannot fully explain the increased number of significant results in phase III trials by smaller industry sponsors. The residual could be due to selective reporting. (4/6)

Bottom line:
- Transparency through result repositories helps--maybe also an idea for other fields?
- Enforcement of disclosure requirements should be more stringent, especially for smaller sponsors for which the disciplining effect of reputational concerns might be weaker. (5/6)

Neat and accessible summary of our results by @Unibocconi

https://twitter.com/Unibocconi/status/1268816337943568384

(6/6)

@threadreaderapp unroll