Tuesday :: Feb 12, 2008

Looking Ahead to the General Election - Part 4: Caucuses and the Clinton Loss/Win Margin

by eriposte

[Also see: Part 1 (The Democratic Nomination and Superdelegates), Part 2 (Polls, Favorability and Electability) and Part 3 (Primaries, Caucuses and Electability)]

NOTE: This post was updated on 2/16/08 with data on New Mexcio where Sen. Clinton was recently declared the winner.

In Part 3, I briefly mentioned that I had not seen a compelling explanation for why Sen. Obama has been dominating the Democratic caucuses. I still don't have a very good explanation, but as I discussed in Part 3, turnout has something to do with his success - but I don't mean high turnout! I hadn't put together all the data in my previous post but I hinted at something:

This is suggestive - and not entirely surprising - that Sen. Clinton tends to do better overall when a much larger percentage of the voting population is voting in the primary or caucuses (and indicates that her campaign has not done a good job of getting their voters to the caucuses and making them stay till the end of the caucuses).

Now that I've had a little bit more time to analyze the data, it appears we may have an interesting trend on our hands. Based on a construct that I saw in a post by MyDD diarist DaveOinSF, I decided to look at the ratio of the 2008 Democratic caucus voter turnout to the number of voters (in the same state) who voted for the Democratic ticket in the 2004 general election. This ratio - expressed as a percentage - is a useful indicator of the fraction of the 2004 Democratic base in each caucus state that turned out to vote in the 2008 Democratic caucuses. When I compared this percentage to Sen. Hillary Clinton's loss/win margin in the caucuses against Sen. Obama, there appears to be a strong trend that broadly confirms my previous observation (above). Here is the chart (raw data with links provided at the bottom of this post) - note that the trend line (added 2/16) is really just a guide to the eye to show that there is some indication of a positive correlation between higher turnout and Sen. Clinton's greater competitiveness in the caucus.


So, the main observations I'd like to make on this data are as follows:

1. The voter turnout in the 2008 Democratic caucuses might be record-breaking compared to the 2004 Democratic caucuses (some of which were not really contested strongly), but the turnout has generally been small compared to the 2004 general election (GE) turnout of voters that supported John Kerry. This is not a surprise in itself since the turnout in caucuses tends to be small.

2. Sen. Clinton appears to have generally increased her vote share vis-a-vis Sen. Obama as the percentage of the 2008 Dem caucus/2004 GE Dem turnout increased.

3. The closest caucus races in 2008 were three caucuses that were contested most aggressively by Sen. Obama and Sen. Clinton - namely, Iowa, Nevada and New Mexico. It is therefore no surprise that these caucuses had the highest voter turnout as a percentage of the overall 2004 Dem voter turnout.

4. The huge victory margins that Sen. Obama enjoyed in the other caucus states are real, but it is almost certain that those margins were significantly inflated because of the lower turnout of voters in those states in comparison to the size of the total Democratic voting population in the same states. In other words, it is only the most motivated voters who generally show up at caucuses and Sen. Obama did a better job of motivating his base to show up to the caucuses and vote.

5. A corollary to #4 is that it is quite likely that in a primary or general election setting - with much higher turnouts - Sen. Obama's victory margins might have been substantially reduced or possibly even reversed (i.e., he may have lost to Sen. Clinton) in some cases.

In a nutshell, Sen. Obama's caucus victories are impressive but it would be foolish to extrapolate from these victories and assume that they reflect better electability in the general election. I see no compelling evidence for this in the turnout data - if anything, the data suggests that Sen. Clinton might have even done better than Sen. Obama in some or many of these states if we had much higher voter turnouts more typical of a primary or general election.

APPENDIX: The turnout data were obtained from various sources as summarized below (win/loss % subject to rounding).

Caucus State
Approx. Voters in
2008 Dem. Caucus
Total Voters Voting for
John Kerry in the
2004 General Election
2008 Dem caucus voters/
2004 GE Kerry voters (%)

Clinton's Win/Loss
Margin v. Obama (%)

North Dakota
New Mexico


eriposte :: 11:03 PM :: Comments (29) :: Digg It!