I am raising a question of the much better attuned to polls, readers of Daily Kos with a question I have been thinking about this afternoon. I read a piece in the Washington Post this afternoon about how a poor field operation for a campaign can cost it significantly on election day. The evidence for the study was interesting, examining areas that received the same media coverage, but where one section was considered a battleground and the other wasn’t. Since the media message was the same or very similar, the difference in the election results seemed to point to better turnout where fieldwork, GOTV, etc. was conducted, in the battleground sector of the study. This raised the question if there was any way to account for this in the poll analysis that is done by FiveThirtyEight or the like? Is it possible to have this factor as a variable in the “Polls Plus” analysis or is it just too ambiguous to utilize in the statistical model? Is it too early to really predict how bad the Trump field work is going to be closer to election day to accurately build anything into the model for this variable? Is it already built into the model perhaps and I just don’t know it?
Not being a polling expert, by a long shot, my own opinion, is that since poor fieldwork can mean a difference of multiple percentage points in the final vote count, that the difference between the Clinton organization and Trump organization at this point in time could translate into a difference from the current polling by at least a couple of points if not more in some key states and areas. Would I be widely off the mark on this?
Thanks in advance for any thoughts someone might want to share on this idea. I have also included the link to the Washington Post if there is someone that might not have seen it.
www.washingtonpost.com/...
