2016 Postmortem
In reply to the discussion: Hot Off The Press Marist Iowa and South Carolina Polls-IA HRC 48% - SBS 45%/ SC HRC-64%-SBS 27% [View all]CoffeeCat
(24,411 posts)has just interjected a bias into the poll.
The poll should reflect the age/gender of the population being polled, not some random opinion of what the pollster thinks people will do or not do.
With that said, most polls probably do not do that. I'm not seeing that this NBC/Marist poll engaged in what you're describing.
However, there is this: There are 450 likely Republican caucus-goers and 426 likely Democratic caucus-goers defined by a probability turnout model which determines the likelihood respondents will participate in the 2016 Iowa Republican/Democratic Presidential Caucus based upon their chance of vote, interest in the election, and past election participation.
This is an issue. In Iowa, they polled "likely Democratic caucus-goers". They used past election participation as a criteria for defining a "likely Democratic caucus-goer." This is flawed methodology that will skew toward Clinton because: The polling excludes first-time caucus goers of all ages and younger voters.
Ann Selzer has discussed this method of polling Iowans--and how it is flawed and unscientific.
http://maristpoll.marist.edu/wp-content/misc/IApolls/IA160124/NBC%20News_WSJ_Marist%20Poll%20Iowa%20Tables%20of%20Likely%20Republican%20Caucus-Goers_January%2028%202016.pdf#page=1