Friday Links
1. On doping &c. I like this protocol:
More than one (usually two) questions are presented to each respondent, who picks which of the two to answer based on a simple probability mechanism that he/she controls, like the roll of a die. The interviewer does not know what the result of the roll was, and thus, which question the respondent answered.
For example, if a 1, 2, or 3 comes up you answer question #1, which is innocuous and has a clear right answer, for example, Does the ocean contain water?. But if instead a 4 through 6 comes up, you then answer the sensitive question of interest, in this case whether you've used [Performance Enhancing Drugs] in the last 12 months.
2. My Bioinformatics paper will make the September 15 issue.
It has been available since early July in Advanced Access, now it is available with a September 15 date. Probably it matters for the people who still read the paper version, but I never met anyone who still subscribed for the paper edition.
3. Dominant Assurance Contracts
I love the logic behind this type of market design and to see that it works out in practice is nice.