About "second choice" votes

12.06.2005

Technorati tag:

One last bit of poll analysis, this one focusing exclusively on the "second intention" voters (when asked "If [preferred candidate] were to drop out of the race, who would you vote for?"). There's also a graphic from El Deber. Nationally, the clear "second choice" winner is Samuel Doria Medina (29%). But a look at the regional breakdowns is interesting.

In La Paz, where Evo leads w/ 53.2% the second choice winner is Doria Medina (22.2%) followed by Tuto (14.8%).

In Cochabamba, where Evo leads w/ 38.4% the second choice winner is Doria Medina (24.1%) followed by Tuto (13%).

In Santa Cruz, where Tuto leads w/ 41.8% the second choice winner is Doria Medina (39%) followed by Tuto (16.7%).

In Oruro, where Evo leads w/ 42.5% the second choice winner is Doria Medina (22.5%) followed by Tuto (19.4%).

In Potosí, where Evo has a narrow lead w/ 29.3% the second choice winner is Doria Medina (29.7%) followed by Tuto (15.9%).

In Tarija, where Tuto leads w/ 48.8% the second choice winner is Doria Medina (41.3%) followed by Tuto (14.9%).

In Beni, where Tuto leads w/ 39.5% the second choice winner is Doria Medina (30.6%) followed by Tuto (19.3%).

Finally, in Pando, where Tuto leads w/ 47.7% the second choice winner is Doria Medina (28%) followed by Tuto (14.2%).

Nationally, Evo is a second choice for only 7% of voters, and breaks 10% only in Pando (10.9%) and Chuquisaca (10.7%). In four departments (Santa Cruz, Beni, Pando, Tarija) the MNR candidate has more "second choice" support.

This suggests a clear problem for Evo as he heads into the last two weeks before the election. Doria Medina seems to be the clear consensus favorite, if his team can capitalize on this to catapult their candidate into second place. This is crucial, because parliament can only select a president from between the first & second place candidates (so far that looks to be Evo & Tuto).

But the problem for Evo is critical, because while his base is solid. About half his voters wouldn't vote for any other candidate. But that amounts to solid support from only about 15% of the electorate (half of his 30% in the polls). Other candidates seem more appealing to a broader electorate. In short, Evo's support is broad but thin.

In contrast, Tuto's support seems to run deeper. More voters are willing to move towards him from other candidates, which could give him a boost in the final stretch. And while his support has declined across the three polls, the total decline has been less than 2 points, well w/in the statistical margin of error (meaning, for all practical purposes, Tuto's support has remained steady). If Doria Medina continues to decline in support, Tuto can more easily capture those voters than Evo can.

On the other hand, Doria Medina should not be counted out. He has enough deep appeal among the electorate to suddenly catapult into second (or even first) place, depending on how the undecideds break, or if voters flock to his banner to avoid a polarizing choice between Evo & Tuto.

So it's still anyone's contest. But it's unlikely that Evo can get the 50%+1 needed to be directly elected to the presidency. And Evo's "thin" electoral support could easily unravel in the next few days, if his opponents play their cards right.

Posted by Miguel at 04:09 PM

Comments

Very interesting, and thanks for linking it at Fruits and Votes.

If I understand the data you present correctly, MAS will win only 2 senators. Agreed?

I knew he was a regional phenomenon, but I had no idea how much!

On his support being "broad but thin," I do not want to parse analogies too much, but isn't it more like "deep but narrow"? That is, concentrated in one regional pocket of the country (narrow) but his supporters tend to really like him intensely, to the exclusion of all others (deep).

Posted by: Matthew at December 12, 2005 08:47 PM

Sorry for the confusion, but no. MAS is likely to win a number of senate seats. The "second choice" poll was about the question of who voters would vote for if their first choice w/drew from the race. Not who was bound to come in second.

I've posted a preliminary lower house breakdown of seats here:

http://www.centellas.org/miguel/archives/001502.html

As for the Senate, it looks like Evo is going to place first in 3 or 4 departments, which would give him 6 or 8 seats (the first place candidate gets two seats, the second place candidate gets the third seat). He'll likely come in second in 1-2 departments, giving him another 1 or 2 seats for a grand total of between 7-10 seats. Tuto is expected to win a majority of them, by winning in the other departments and placing second in most of the rest. BTW, the MNR is expected to get 1 or 2 seats in the senate.

Posted by: mcentellas [TypeKey Profile Page] at December 12, 2005 10:53 PM