My results are mixed with polls and quizzes. I had great results with quizzes with fun articles of humor with a hint of an opinion. I don't know if those articles really drive Google traffic. I dun'no . . . I did use it with informational too.
The quizzes were designed offering a chuckle inside the article the first type. Perhaps that created more interest with what followed? I would use hints like read further or more following kinda' thing with both types. Sometimes I would offer a poll at the end based on that quiz. I haven't tried that strategy for quite sometime.
With polls today my results related to traffic is maybe around 5+%. Today, I tend to use two. One about 1/3-1/2 along. The other at the end. I get best results with the first. I ponder how far the article is read. Maybe checking with Google Analytics the normal reading time on an article will offer a hint if exit occurs with where the poll is?
My participation with polls and quizzes? I always look at the results of a statistic orientated poll first though don't answer all. Is that cheating? In other words when seeking a range of something if there are already many answering my vote will not be relevant. So, an early on one I will more than likely vote. As intent I most likely will not return to see if the result changed. I fancy offering an opinion type poll without looking first.
Quizzes depends on the article type and what is offered with the question. I run from quizzes orientated toward 'How much did I learn'. I don't want to know if I am smart or not.