Wednesday, September 12, 2012

The Embryology Of Spin


Yavchitz et al. looked at which factors correlated with the presence of "spin" in the reporting of medical randomized control trials. Spin is emphasizing the benefits of a treatment more than is appropriate on the basis of the data. They cooked up a multivariate regression with the explanatory variables of journal type, funding source, sample size, type of treatment (drug or other), results of the primary outcomes (all nonstatistically significant versus other), author of the press release, and the presence of “spin” in the abstract conclusion.

In their sample (N = 41), the only factor that correlated significantly with spin in the press release and news article was spin. Spin in the abstract conclusions of a study leads to a 5.6 (95% CI 2.8–11.1) times higher relative risk of there being spin in the press release and news reports. So, to the extent that we care about curbing vicious information cascades, it's essential for authors and editors to be conscientiousness about word choice and framing in the abstract. 

Tuesday, September 4, 2012

The Trade-Offs Of Publicizing Your Goals

In Ben Casnocha's reflections on writing his book The Start-Up Of You, he mentions this tidbit:
When you embark on a project that’s going to take awhile, you have to decide how much to publicize the fact – on your blog, to your friends — that you’ve started. ... When you publicly announce that you’re starting toward a goal, you can benefit from the self-fulfilling prophecy effect, you can collect feedback from your network, and be held accountable to lots of external people tracking your progress. On the flip side, when you announce a goal, you risk tricking your mind into believing you’ve already partially accomplished your it when in fact you’ve done nothing. Derek Sivers says: “Keep your goals to yourself.”) Plus, external accountability of the wrong kind can add unhealthy pressure.
This is a complicated and thus interesting trade-off. It involves an interaction between managing your own psychology and providing others the context they need to offer you help. The best-case scenario (not necessarily possible) would be for others to know what you are attempting to do without you knowing that they know. On the other hand, the worst-case scenario (much more plausible) is that you think others know when they don't.

In science, there are often norms against sharing too much of your project with outsiders, to prevent it from being "scooped". It strikes me now that these norms serve the dual purpose of preventing you from taking mental credit for something that you haven't yet done the grunt work to accomplish.

Anyway, I certainly don't have any general solutions to this trade-off, and it is something I worry about too.

I am probably slightly biased about the book because Ben is a friend, but I recommend it highly.