If article a from source A has 100 votes, and article b from source B has 10 votes, without simply saying that article a is better because it was voted more, how can I normalize the weight of these votes based on source popularity? Seen in that light, if source A is incredibly popular, 100 votes is actually not so much. If B is incredibly unpopular, 10 votes might be a game changer.
Is there any way to describe this mathematically?
Thanks in advance! :)
A simple way is to ask what fraction of the votes on source A went to article a and compare that to the fraction of the votes on source B went to article b. So if source A gathers a million votes for all the articles, and article a only got 100 or $0.01\%$ of them while source B got 100 votes and article b got 10 or $10\%$ of them, you could rate b higher. Maybe there are more articles on source A, so you should look at the fair share of votes for each article. There are many possibilities. Maybe the voters on source A don't give out votes as easily as those on B....