One of my members has pointed out an anomaly in the results of Polls. I'm going to assume this is NOT a "bug" but perhaps due to a lack of knowledge or perhaps it is something which cannot be coded. For the sake of clarity I'm simply going to include this members actual observation and request below:
Polls, whether single- or multiple-answer, are intended to show the percentage of RESPONDENTS who give a particular answer. For some reason the percentage function here, when a poll is configured to allow multiple answers, shows the percentage of ANSWERS instead. Now my guess is that the same function is being used for single-answer polls, but it doesn't result in an error there because in that case #RESPONDENTS==#ANSWERS, resulting in the correct percentage even though an incorrect calculation is done.
So someone needs to change the variable in the denominator from ANSWERS to RESPONDENTS. Until that is done, all these polls are going to have really scrawny, meaningless numbers, which in turn might inhibit some folks from using them.
And you might want to add, when talking to the engineers/coders, that the suggestion was not just from a peevish programmer, but a peevish programmer in the Market Research industry whose days consist of reporting ... answers to Poll Questions!
