How Fast Can It Change? Iowa Caucus Polling - December 2003 and December 2007
The Iowa Democratic Caucus was held on January 19, 2004. On December 2, 2003 Howard Dean was leading the pack with 26% supporting him, with Dick Gephardt in second at 22%. John Kerry was in third at 9%, and John Edwards was fourth with 5%.
On January 7th, less than two week out, the numbers were different but the ranking was the same: Dean 30%, Gephardt 23%, Kerry 18%, Edwards 11%.
The actual results on January the 19th were: John Kerry 37.64% John Edwards 31.83%, Howard Dean 18.02%, and Dick Gephardt 10.06%.
Source Wikipedia: https://en.wikipedia.org/wiki/Opinion_polling_for_the_Democratic_Party_presidential_primaries%2C_2004#Iowa
The changes in how Democratic candidates fared between pre-caucus polling for the Iowa Caucus and the actual results on January 3, 2008 were not quite as dramatic, but they still were major. I found this data from ARG (American Research Group) for the Iowa Primary Contest. For point of reference, ARG was ranked tied for 5th out of 20 polling operations for their accuracy in predicting the final outcome of the 2008 presidential elections. Source: http://www.politisite.com/2010/08/06/poll-accuracy-in-the-2008-presidential-election-rasmussen-pew/
12/16 - 12/19/2007
Clinton 29%
Obama 25%
Edwards 18%
12/26 - 12/28
Clinton 31%
Obama 24%
Edwards 24%
12/31 - 1/2/08
Clinton 34%
Obama 25%
Edwards 21%
Actual Iowa Results on January 3, 2008
Obama 37.6%
Edwards 29.7%
Clinton 29.4%
Not to be misleading, the Real Clear Politics (RCP) composite polling data for Iowa that year was more accurate in the final days than ARG was for the Iowa contest (but harder to break out comparative figures out of), though composite figures for the weeks before those final days were similar.
For the period 12/26/07 through 1/02/07 RCP predicted:
Obama 30.8%
Clinton 29.2%
Edwards 26%
It should be noted that while both Joe Biden and Bill Richardson were only polling in single digits, they each still under performed by several points on the night of the actual caucus.
Here is something from the Gallup organization that is worth pondering, but while you do please note that it was published on January 6, 2004. In other words, it does not include data from the 2004 and 2008 races
"There have been 10 races over the last 50 years in which there was a significant contest for the Democratic nomination: 1952, 1956, 1960, 1968, 1972, 1976, 1984, 1988, 1992, and 2000. (The omitted years of 1964, 1980, and 1996 were ones in which a Democratic incumbent president ran for re-election with little or no opposition.)...
...In fact, in only 4 out of the 10 elections (Adlai Stevenson in 1952, John F. Kennedy in 1960, Walter Mondale in 1984, and Al Gore in 2000) did the front-runner in late December/early January win the Democratic Party's nomination. In all other instances, someone else came from behind as the primary season unfolded."
http://www.gallup.com/poll/10120/history-shows-january-frontrunner-often-does-win-democratic-nomination.aspx
One final point. National opinion polls can react strongly to he results of early caucuses and primaries. Let's go back to the 2004 Democratic contest again and look at how national opinion polls correlated to early states results. Remember, Iowa voted on January 19th and NH on January 27. Here is a snapshot from a CBS national opinion poll taken shortly before those contests, and one a month later. Source:
http://pollingreport.com/wh04dem.htm
Shortly before the 2004 Iowa Caucus, national polling:
1/12-15/04
Dean 24
Clark 12
Gephardt 11
Kerry 7
Edwards 5
One month later it looked like this
2/12-15/04
Kerry 53
Edwards 7
Sharpton 4
Kucinich 1
Other 12
It is a little bit early to be talking about the 2016 presumptive Democratic nominee.
= new reply since forum marked as read
online sometime Sunday/Monday.
