We’ve received several reader comments about how we rank the candidate matches in the election quiz. Based on those constructive comments we are making improvements to the ranking system.
On Oct. 20, 2014, we changed the quiz to rank candidates based on their number of pro and con matches with you vs. their percentage of matches.
Reader Daniel Westcott emailed us on Oct. 14, 2014 with a comment about our election quiz. He also posted his comment on Facebook which another ProCon.org reader kindly forwarded to us. We thank Daniel for his feedback which we have reposted here so our readers can see more of the context which prompted an improvement to one of most popular resources.
"Flaw. If anyone knows anyone at procon.org, please pass this on. Not sure, but I think maybe Frank Gruber only answered 4 out of the 15 questions -- apparently these are safe questions -- the ones where he would agree with most voters (not sure though). The table at procon.org indicates that I agree with Frank 100% of the time and he is listed 2nd on my customized candidate list. I believe the bar graph and the percentage should show that if the candidate did not answer, there is not a match. If that was changed it would show my agreement with Frank Gruber at 26%. Nothing against Frank, but if someone wants to be informed about voting, this does not work. I hope this comment is encouraging critical thinking and I hope you fix it. Thanks."
We greatly appreciate this constructive criticism from Daniel. As a result of his initiative, we will now rank candidates based on their number of pro and con matches with you vs. their percentage of matches.
Under our prior system, if Candidate A answered "Pro" or "Con" to only 3 out of 10 questions (with the rest of the responses being "Not Clearly Pro or Con"), and if you agreed with Candidate A on those 3 questions only, you would see Candidate A as first in your ranking list with a 100% (3/3) match with you. By contrast, if Candidate B answered pro or con to all 10 questions and you agreed on 8 out of 10 issues, then using a percentage measure would indicate an 80% match. You would have more in common in Candidate B, but because of the percentage ranking system, Candidate A would appear higher in the rankings. With our new ranking system, you are matched with candidates based on the number of pro and con answers you and the candidates have in common – a much better system.
