ingres comments on 2016 LessWrong Diaspora Survey Results - LessWrong

32 Post author: ingres 14 May 2016 05:38PM

You are viewing a comment permalink. View the original post to see all comments and the full post content.

Article Navigation

Comments (21)
Tags: fortforecast analysis empiricism data 2016 survey lesswrong

Comments (21)

You are viewing a single comment's thread. Show more comments above.

Comment author: ingres 15 May 2016 10:07:03AM *  5 points [-]

I'm somehow reminded of the scene in Logicomix where Russel figures that Godel should be given some kind of award for actually reading his Principia, though he's not sure what.

You went through the analysis files, closely enough to spot errors, and then actually went and reported them? Props.

Comment author: Emma_MK 15 May 2016 04:02:32PM *  7 points [-]

Thanks. I'll treat myself to a cookie. There's more:

In Calibration Question And Brier Score Analysis:

  • All questions that ask for an answer "on a scale from one to five" show two 5.0 answers, with different percentages. I guess one of the fives should be None. That's 3 questions for each Brier score group.

In Blogs And Media Demographic Clusters:

  • If None means "no answer", then I think it is incorrectly given as 0.0 for some questions. For example, let's sum up the percentages for ReligiousViews for Cluster 0, where 0.0 is for None: 0.027+ 0.055 + 0.112 + 0.414 + 0.0 + 0.019 + 0.061=0.688 The percentages do not sum up to 1.

  • After the questions about children, there are 9 items labelled WorkStatus_1..9. I think these should be Student/Full Employment/Self-employed/Unemployed, etc.

  • In EducationCredentials, these are listed as two separate items: Bachelor's and Bachelors

  • Not an error, but maybe sort the answers to the number-of-children and number-of-partners questions? Looks a bit weird when they are scrambled and go like 0, 4, 5, 1, Lots and lots, 2, 3.

Powered by Reddit Powered by Reddit