Replies

Hummm... their last post on the website sez:

"Santorum holds a small edge there with 29% to 27% for Romney, 22% for Gingrich, and 19% for Paul."

Also, quote: "Minnesota looks like a toss up... with 29% to 27% for Romney, 22% for Gingrich, and 19% for Paul."

But, who knows - polling is very often indicative, very often dead wrong.

that’s an old article from the other day talking about their last poll. This tweet is fresh off the press and they haven’t even released the details yet.

go to tweeter and search Public Policy Polling, and you will see all their tweets

Polls on caucuses (as opposed to primary or general elections) are notoriously difficult because you are polling a subset of a subset.

Comparing 2008 turnouts is a fairly good idea since the base year is one of the few in recent history where both parties start out without an incumbent president or vice president as the perceived front runner. You have to go back to 1952 to get a comparable situation.

Good laboratory states to pick for comparison are New Hampshire, Iowa and Minnesota. Not only because they are early in the process when the interest is the highest and the field completely winnowed down, but also because they are very high voter turnout states.

Here are the statistics from 2008:

Iowa:
Republican Caucus participants 118,696
Democrat Caucus participants 13,485
General Election voters 1,537,123
Caucus participants as percent of General Election voters 8.6%

New Hampshire:
Republican Primary voters 239,793
Democrat Primary voters 287,556
General Elections voters 710,970
Primary voters as percent of General Election voters 74.17%

Minnesota:
Republican Caucus participants 62,828
Democrat Caucus participants 214,066
General Election voters 2,910,369
Caucus participants as percent of General Election voters 4.51%

You can run this type of math on any of the other states, but a few general rules hold sway:

Primary voting or caucus participation is largely a function of general election participation of registered voters.
Caucus participation, even in the most ideal conditions, seldom will reach even 10% of those voting in a General Election. As the process continues, the participation rate drops even in high voter participation states.
Same trend for primary states. In ideal situations (e.g. the publicity whores in New Hampshire), participation rates in the primary may approach the same high level of participation with respect to general election voters as do general election voters with respect to total registered voters. But the overall trend is still downward as the nominating process moves on.

Just to verify, look at the latest labatory ideal primary state in 2008 (also a high turn-out state).

South Dakota:
Republican Primary voters 60,964
Democrat Primary voters 97,797
General Elections voters 381,975
Primary voters as percent of General Election voters 41.56%

Bottom line is that while Primary polls tend toward accuracy because you have a large participation rate, Caucus polls can be all over the place because even in a high turnout state like Minnesota, you have less than 1 in 20 voters who ends up participating in the general election actually turning out to caucus.