Measuring distrust (Score 1)

The actual paper this story is based off is quite interesting and worth a read (PDF). There's some really weird beat frequencies occurring in figure 2 that make me wonder if their data is valid. From the wording of both author's and Slate's article it appears that the collection of self-censorship data was started as a research project. If a user types 5 characters into a comment box, a boolean value is sent back to facebook if they decide not to continue, so it looks like its not sending what you type. I guess it's a good metric for measuring how much you distrust a person or system if you carefully revise statements before posting. Of course, perhaps if the self censorship data is interesting enough, Facebook's policy might change...