EYES ON IOWA: TRUMP AND CLINTON SURGE 10 DAYS BEFORE THE CAUCUS. CRUZ AND SANDERS FALTERING AT FINI

[JackRiddler]

17. It's a robocall poll.

I know anytime a robot calls, I hang the fuck up, as do higher proportions of people than with humans.

The results fluctuate by more than 10% per day!

The "finish" is ages away, the all-caps headlines on the site are a tipoff on the intellectual tenor of the outfit.

Emerson College-based, but a for-profit outfit (.com, there's a page for putting in bids) that exploits student labor.

"Methodology" at http://www.theecps.com/#!methodology/c1iwz.

Through the use of a screening question, we usually seek out "likely voters" which has been shown to give a much more representative sample of the populations for most political polling.

Our first step in weighting is to survey more than enough people. This allows us to then be able to systematically reject individual surveys from demographics that are over represented. Next, survey data is weighted with a 3 point decrease in Conservative opinion and a 3 point increase in Liberal opinion to offset the bias in land line only telephone polls.

For the Presidential Polling Initiative the following Methodology will be employed. First, all respondents who did not finish the survey is eliminated, then anyone who finished the survey in under 4 minutes is eliminated (based on 2 standard deviations and length of survey). If respondent said they did not vote in 2012 they are eliminated because prior voting behavior is the best predictor of future voting behaivor. Finally, using a rim weighting the results are weighted based on 2012 final election returns either nationally or by state depending on the sample and then by region (percentage of electoral vote for each region or total vote if it is a statewide poll).

All of which is arbitrary, especially given the extraordinary circumstances of 2016.