@Gargron It should be based on the amount of votes. Each option represents an isolated true/false question.

@gargron If users can choose more than one option, I would expect the totals to add up to MORE than 100%.

Otherwise you lose information.

If the totals add up to 100% I assume everyone voted exactly once.

@Gargron

On multi-choice poll it's way more clear to know which percent of voters chosed each option

But maybe it would help to distinguish easily between multi-choice and single-choice polls

Thank you the answers are making me more confused than before

@Gargron Yeah I feel like it's wrong too. Some of us can barely add so seeing a larger percentage than 100% in a poll is confusing.

@Gargron it's not wrong, because the aim is usually to know how many people has chosen that option. But here, who knows...

@Gargron no, that's fine... in a multiple choice poll the options aren't exclusive, so they don't have to add up to 100%. the actual poll is a yes/no question for each, so the only constraint is that each option has between 0-100% votes.

@Gargron Always should be 100%.

@Gargron the question assumes that "percentages" already have a meaning before you try to get "100%". Percentages of what? If you want to show the relative strengths of each option, then the number of users who voted are irrelevant, and it's a matter of showing 100 * optionVoteCount / totalVoteCount for each option. And then that result tells you how to rank the options, but nothing about the _popularity_ of each option. For that, you'd need to weigh votes based on how many things a user voted.