Poll Math

[eponym]

I admit that sometimes I don't know math from a hole in the ground, but isn't it standard for poll percentages to add up to 100? I think this was addressed a couple of years ago but I searched to no avail.

-Bryan

 

[*dave*]

These stats are obviously generated by a golf pro.

 

[eponym]

Or maybe SB reserves one-sixth for rock bottom opinion.

-Bryan

 

[tc]

That's not a bug- it's a feature.

 

[redacted]

Clearly round off error. In the first case, the sample size of 15 was rounded of to 13. In the 2nd case, the sample size of 79 was rounded off to 66. Maybe it is the new math.

 

[vladimir]

I admit that sometimes I don't know math from a hole in the ground, but isn't it standard for poll percentages to add up to 100? I think this was addressed a couple of years ago but I searched to no avail.

-Bryan

They are multiple choice polls. If you and I were the only respondents, and we both selected item one and I selected item two as well, item one would show 100% and item two would show 50% for a grand total of 150%. Single choice polls will add up to 100%

 

[HowardE]

They are multiple choice polls. If you and I were the only respondents, and we both selected item one and I selected item two as well, item one would show 100% and item two would show 50% for a grand total of 150%.

Exactly.

Multiple Choice polls... the percent of respondents is not necessarily equal to 100% as users can pick more than one response.

In polls where there are only 1 choice per user, totals will equal 100% (within rounding off limits of +- 1%)

 

[gypsyjim]

A lot of fun polls on here have many possible right answers!

Which often appeals to those of us with seriously warped senses of humor!

 

[redacted]

Exactly.

Multiple Choice polls... the percent of respondents is not necessarily equal to 100% as users can pick more than one response.

In polls where there are only 1 choice per user, totals will equal 100% (within rounding off limits of +- 1%)

I'm not sure that the percent of respondents is a more useful percentage than the percent of responses but I'll have to think about that. In any case, is it possible to add the average number of responses per respondent to multiple response polls to avoid this confusion. Even something as simple as a count of responders would work.