Unnamed comments on 2011 Survey Results - Less Wrong
You are viewing a comment permalink. View the original post to see all comments and the full post content.
You are viewing a comment permalink. View the original post to see all comments and the full post content.
Comments (513)
It looks like about 6% of respondents gave their answers in decimal probabilities instead of percentages. 108 of the 930 people in the data file didn't have any answers over 1 for any of the probability questions, and 52 of those did have some answers (the other 56 left them all blank), which suggests that those 52 people were using decimals (and that's is 6% of the 874 who answered at least one of the questions). So to get more accurate estimates of the means for the probability questions, you should either multiply those respondents' answers by 100, exclude those respondents when calculating the means, or multiply the means that you got by 1.06.
=IF(MAX(X2:AH2)<1.00001,1,0) is the Excel formula I used to find those 108 people (in row 2, then copy and pasted to the rest of the rows)
Nevermind.