June 11, 2015

Slipping Down the Fluid Slope of Ethical Integrity

This post will focus in on the slippery slope of research ethics, particularly the consequences of strange things happening in the course of pursuing one's best intentions. Cheeky images and puns will help to accentuate the story.

I have been leery of the start-up Uber ever since I heard stories about their varied ethical breaches [1], but now I'm even more deeply skeptical. Uber is showing exactly what can be accomplished when the "technically not illegal" ethos runs amok. Apparently, the ridesharing service entered into a research partnership with Carnegie Mellon, only to poach massive amounts of staff at-will [2]. As I understand it, the reason for this is largely superfluous. Uber wanted to possess expertise in Artificial Intelligence and automation, but did not want to go through an intermediary. Generally speaking, academic-private sector partnerships are not supposed to work like patent troll litigation. But Uber is a wildly-successful startup, so some non-zero percentage of the population are sure to overlook the ethical lapses.

Not illegal, not illegal....the Uber business model? REFERENCE: Family Guy.

Another example of the ethical slippery slope comes from the open-access journal troll and Alan Sokal wannabe John Bohannon. As a feature writer for the journal Science, he once did an updated version of the Sokal hoax where nonsensical papers were sent to a large number of open-access journals. The catch is that some of the journals published the articles for little more than a publication fee (and minimum editorial oversight) [3]. More recently, a new hoax involved an intentionally shady study that touted the health benefits of chocolate [4]. The thing is, the popular press picked up on the paper before it could be retracted. Awesomely delicious stuff (pun intended). The chocolate study has ignited a debate in the world of internet opinion, both in support and as criticism. Aside from the rhetorical point that provocative results can often result from p-value hacking [6], the most obvious problem is that you are intentionally drawing questionable conclusions and having them published. Even though an important point is made, the purposeful dissemination of false findings can lead to serious unintended consequence [7]. In any case, I suppose if the pages of Science ever see a high-profile paper retraction [8], Bohannon's team will get right on the case.

The iffy ethics of sting operations against questionable peer-review processes and journalistic hype machines. REFERENCE: Willy Wonka and the Chocolate Factory.

[1] Newton, C.   This is Uber's playbook for sabotaging Lyft. The Verge, August 26 (2014).

[2] Lowensohn, J.   Uber gutted Carnegie Mellon's top robotics lab to build self-driving cars. The Verge May 19 (2015).

[3] Alicea, B.   Fireside Science: the Consensus-Novelty Dampening. Synthetic Daisies blog, October 22 (2013).

[4] Bastian, H.   Tricked: the ethical slipperiness of hoaxes. Absolutely Maybe blog, May 31 (2015).

[5] Gelman, A.   John Bohannon’s chocolate-and-weight-loss hoax study actually understates the problems with standard p-value scientific practice. Statistical Modeling, Causal Inference, and Social Science blog, May 29 (2015).

[6] Kassel, M.   John Bohannon's Chocolate Hoax and the Spread of Misinformation. Observer.com, June 6 (2015).

[7] Data “were destroyed due to privacy/confidentiality requirements,” says co-author of retracted gay canvassing study. Retraction Watch blog (2015).

No comments:

Post a Comment