Seems far fetched to me but I'd imagine it's testable. One could test people about their knowledge of poll data and then see whether a higher degree of interest in polls correlated with a lower awareness of these undemocratic feature. I would imagine you'd find your thesis refuted in such a test.

Anyhow, it's likely you're trying to convey something that I'm not quite understanding so please feel free to correct me.

Speaking personally, I would posit that the media puts more focus on polls because they are both interesting and timely to viewers, where as discussions about the flaws of first-past-the-post voting are more dry and are not timely -- timeliness being a crucial determiner of what gets reported as news. So unless there are newsmakers actively making a spectacle about the issue, you're not going to here about it in mainstream media.