It would be interesting to see if the data from the poll match the log data. I'm always curious about the accuracy of self-reporting in polls like this, with the ingoing assumption that people tend to be more estimators of frequency either with an aspirational bias (how often they think they should do something) or just a recent history bias (if I have done something a few times recently, I likely overestimate how often I do it).
Lots of reported conclusions depend on polls like this. Since in this case the real data may be available in the logs of HN, it would be more interesting to me to see the comparison of real v. reported than it is to see the actual answers.
