Discussion for article #228257
Over the remaining five to six weeks in this campaign season, I think it would be a very large jump in the reporting quality on TPM (and everywhere else) if the “likely voter” models for each pollster were released, either with each story or as a one time cheat sheet. That is, what is the party breakdown being used for these “likely” voters or is it simply self identified 'likely"?
Don’t get me wrong, I’m not trying to “unskew” the polls, but I think its pretty obvious if one polling firm is using a 2010 turnout model and one is using a more long term model (or perhaps some “enthusiasm” based model) they will end up with different results. Thus it is significantly more informative to us as readers to understand the meat (cross-tabs) behind each poll…
" A recent Des Moines Register poll seemed to predict doom for Rep. Bruce Bradley…"
It’s easy to tell that Daniel Strauss reported for Politico and The Hill before coming to TPM.
No one would ever accuse the Des Moines Register as being DEM leaning.