Not a bug Poll percentages incorrect for multiple choice polls

[Paul B]

I'm not sure this is how the percentages should be displayed for multiple choice polls.



Surely the total should still be 100%?

 

[Brandon_R]

Nice catch

 

[Disjunto]

someone's calculating from the number of user's that have voted not the number of votes....

 

[Paul B]

I put this in the discussion forum but I'm fairly sure it's a bug.

 

[Disjunto]

I'm not sure if it is, ~66.7% of people would start a new community. this is correct and is a useful bit of information. If it said ~33.3% of the votes, it would be less informative really.

 

[Mike]

It's not a bug. It's how multi choice voting works. If you ask 10 people if they like A, B, C, and D (allowing them to choose multiple values), and they all chose each option (40 "votes", 10 voters), it doesn't make sense to report 25% for everything. 100% of people voted for each option.

 

[Paul B]

I suppose there's an argument either way but I think this could be quite confusing.

 

[majdi]

It's not a bug. It's how multi choice voting works. If you ask 10 people if they like A, B, C, and D (allowing them to choose multiple values), and they all chose each option (40 "votes", 10 voters), it doesn't make sense to report 25% for everything. 100% of people voted for each option.

your logic is correct bu its confisung for the average Joe like me to see a poll with a global result higher than 100%

 

[Kier]

I think this could be quite confusing.

The truth may be confusing in many instances, but it doesn't make it any less true.

 

[Enigma]

I agree with Mike on how this works. I guess the confusion comes from the fact that in multiple-choice polls, we each define "percentage" in two different ways:

1. Percentage as the ratio of how many voters out of all voters voted for an option-- so if all voters choose option A as well as other options, then option A has 100% (and other options have other percentages).
2. Percentage as a comparison to the other options in the poll, or popularity in comparison to the other options. (In this case you can't really call it "percentage" anymore, you'd have to call it "popularity ratio" or something)

For example there are 10 voters and 2 options. All ten people voted for option A, and five people voted for option B. With definition #1, this would be:

A. 100% (100% of the people voted for this)
B. 50% (50% of the people voted for this)

With definition #2, this could be:

A. 66.6% (10 of 10 people voted for this, thus it is twice as popular as option B)
B. 33.4% (5 of 5 people voted for this, thus it is half as popular as option A)

For people who want the percentages to add up to 100%, the second definition makes sense. For those of us who think of percentage as the ratio of how many people out of the total number of voters voted for a particular option, definition #1 makes more sense.

Personally, I'd keep it as-is.

 

[Mike]

FWIW, it's effectively this system: http://en.wikipedia.org/wiki/Approval_voting Though that is brought up in a political sense.

 

[Nick]

I suppose there's an argument either way but I think this could be quite confusing.

This method only seems confusing because it's a multiple-selection poll. If you look at a poll that only allows voters to select one option (e.g. http://xenforo.com/community/threads/handedness.1649) you will see that the percentages add up to 100%.

 

[Peggy]

Note to self - never use multiple choice poll!

 

[Paul B]

The truth may be confusing in many instances, but it doesn't make it any less true.

True

 

[Raitsa]

Somewhat offtopic but still:
do i have to count the total given votes myself?

Number of total votes cant be seen anywhere?