Categories
Elections Federal Politics Iowa Political Analysis Political Campaigns Political Polls United States White House

My attempt to analyze a junk Iowa Democratic caucus poll that was done for Breitbart

Earlier today, an obscure polling firm called Neighborhood Research and Media released a poll of Iowa Democratic voters in the lead-up to the February 3rd Iowa Caucuses. You can view the results of the poll here if you wish to do so.

Anyways, the poll was enough to move the FiveThirtyEight poll aggregate to show a narrow Joe Biden lead in Iowa; the aggregate prior to the poll showed a narrow Bernie Sanders lead in Iowa. FiveThirtyEight lists Neighborhood Research and Media as a B/C-graded pollster, so the pollster is deemed by FiveThirtyEight to have not conducted enough polling to receive a proper grade.

I will now make several observations about the recent poll out of Iowa.

The poll was conducted for the far-right website Breitbart

Any attempt by a right-wing organization, such as the far-right, pro-Donald Trump website Breitbart, to poll Democratic voters about an upcoming Democratic nomination contest is immediately suspicious in my view. Breitbart may be trying to drive a Democrats-in-disarray political narrative, a narrative that benefits Republicans politically, in more mainstream media outlets, and, at the very least, has certainly influenced FiveThirtyEight’s polling aggregators, by paying Neighborhood Research and Media to poll Iowa Democrats.

The poll sample was DRASTICALLY older than the 2016 Iowa Democratic caucus electorate

The poll sample was drastically older than recent Iowa Democratic caucus electorates, including the 2016 Iowa Democratic caucus electorate. In 2016, entrance polls, the caucus equivalent of exit polls, showed that 37% of the Democratic caucus electorate in Iowa was 44 years of age or younger, with 28% older than 65 years of age. The pre-2020 caucus poll for Brietbart had a sample with less than 19% of likely caucus participants under the age of 50, with more than 58% of likely caucus participants over the age of 65. The median age in poll for Breitbart was likely more than a decade older, maybe as much as a decade and a half older, than the median age in the 2016 entrance polls!

The poll included Donald Trump as if he were a Democratic presidential candidate

Donald Trump is seeking re-election as a Republican presidential candidate. However, the poll for Breitbart listed Trump as if he were a Democrat and a candidate in the Iowa Democratic caucuses, which he is not. While Iowa Democratic caucus participants may have an uncommitted option (I’m not sure if the Iowa Democratic caucus rules allow for that), Trump will not be a candidate in the Iowa Democratic caucuses. The poll had Trump, who is not a Democrat and is not seeking the Democratic presidential nomination, at 4.7% among likely Iowa Democratic caucus participants.

The poll was in the field during the January 14 Democratic presidential debate

The poll for Breitbart was in the field from January 14 to January 17. This means that the poll was in the field during the January 14 Democratic presidential debate that was televised by CNN and sponsored by the Des Moines Register. One day of sampling, January 14, was pre-debate, with the other three days being post-debate. The poll had Biden losing a ton of support from the pre-debate sample to the post-debate sample, with Amy Klobuchar and Elizabeth Warren, respectively, being the largest beneficiaries of an apparent loss of support for Biden. The pre-debate sample was a one-day sample that included 103 likely caucus participants, whereas the post-debate sample was a three-day sample that included 197 likely caucus participants, or an average of 66 likely caucus participants sampled per day. Given Iowa’s population, samples of less than 200 individuals have a very high margin of error.

Conclusion

This poll that was conducted for Brietbart is complete junk, in part due to the poll’s sample being drastically older than what an Iowa Democratic caucus electorate would be expected to be.

Advertisement