October 07, 2005

Evo Morales and MAS Ahead in the Polls

MABB © ®
Three different companies very recently conducted surveys trying to take the pulse on Bolivian voter's preferences. Equipos Mori, Encuestas y Estudios and Apoyo, Opinion y Mercado Bolivia, asked a representative pool of voters during September and the first week in October for whom would they vote in the upcoming general elections. The results seem to point to a clear lead by presidential candidate and leader of Movement Toward Socialism, Evo Morales. But I think, a closer look at the results reveals big diverse results which are by no means a clear indication of who will win next December 4.

The first thing that jumps out right at first look is that on the three surveys, Evo Morales leads Tuto Quiroga by at least 3 percentage points (the Mori survey) and as much as 6 points (the Apoyo and EyE surveys). At national level, this is a respectable lead, though if the lead is more like 3 points, it should not be insurmountable. Currently, there is no a historical pattern to compare these results with prior surveys. That is why the significance of these results are somewhat hadicapped, but useful nevertheless. The lack of historical references does not permit us to especulate on how easy voter's preferences change in Bolivia in terms of speed or magnitud. So, we need to take the outcomes at face value and conclude that Evo Morales is leading the race to the Bolivian presidency, but that lead is not assured.

It is when we start digging into the numbers when more and more questions start to emerge. For example, whe we look at the results per departments (the political units in which Bolivia is divided), we immediately notice the big differences in the percentages. In the Mori survey, Evo leads Tuto by 33 percentage points, while in the EyE poll, the lead is reduced to 25 points. That is a difference of some 8 points between them. But what is more striking is that in the Apoyo poll Evo only leads by 6 points. That is a considerable difference of some 27 points with the Mori survey and 19 points with the EyE study. Why such a diverse set of outcomes? I mean, in polling numbers 2 or 3 percentage points are significant. What are we to make of differeces in the order of 27 or 19 points among polls?

Continuing with the same type of observations, when we turn to the Santa Cruz results, we can observe that in the Mori and EyE polls Tuto leads Samuel by 13 points. However, the Apoyo people report a lead by Samuel over Tuto of some 5 points. How can that be? Who did the Apoyo people ask? We are talking about the same department (Santa Cruz) placing two different candidates at the top of preferences at the same time. The same situation happens in Cochabamba where in the Mori and EyE polls Evo leads Tuto but in the Apoyo survey Tuto is leading Evo.

From these observations I tend to conclude that the Mori and EyE surveys are more or less comparable, while the Apoyo poll is conspicuously at odds with the prior two. It suggests that the Apoyo people asked a completely different group of voters than the Mori and EyE people. In fact, this is not the first time I make such observations to Apoyo polls. In this post I argue that Apoyo's numbers are not reliable because their sample population outright ingnores the country side. They concentrate on what they call eje troncal (something like core axis) which consists on the major cities plus major representative rural towns.

Granting that this sample population might be statistically representative of the whole population as numbers is concerned, but the oppinions and preferences in the very rural areas and the population centers are notably different. I just think that these population (the rural population) cannot be excluded for statistical purposes in Bolivia.

In stark contrast, the polls conducted by Mori and EyE included the major cities as well as a significant number of rural towns and small villages. The Mori survey was conducted in 58 major and intermediate cities as well as rural town and small villages in the 9 departments of the Bolivian territory. The EyE survey was conducted in the 9 capital cities of each department plus the city of El Alto. Additionally, the EyE people went to 161 intermediate, rural towns and small villages around the country.

I think it is clear why the two (Mori and EyE) polls are closer in results. The two companies not only included ALL the major (capital) cities in each department, but also went to the country side where about, according to the 2001 census, 38% of the population still lives.

Nontheless, one partial fact emerges of these three polls: Evo Morales is closer and closer to reach his goal. The only thing he needs to do is to bring Samuel Doria Medina to his camp and convince him to vote for Evo in Parliament (where the Bolivian presidency is decided).

More links on the topic:

Here is a link to Ciao!'s post on the issue.
Here is another link to Miguel Centella's (Author of Ciao!) flikr data graphic stream which has a collection of La Razon's graphics on the Bolivian elections.

2 comments:

mcentellas said...

Good catch, I was in too much a hurry this weekend to get a close look at that. But another difference between the three polls is the time they were taken, the Apoyo poll was a few weeks earlier than the others, I think, and those two were almost done the same time (last week/first week of Sept/Oct). That could make a difference. Maybe?

It's good that there are now more polls, which allows to compare across polling companies. Apoyo has long had a near-monopoly. Now we can start comparing it's results to other polls, to see which ones are better at predicting.

MB said...

I have recently seen an article which talked about the polls conducted by the parties themselves. There were three polls carried out by Podemos, UN and MNR and put Tuto in front by 5 or 6 points.

What's funny is that all of the polls showed Tuto in front. I got the impression that UN was getting closer to MAS and away from Podemos.