Monday, November 22, 2010

Tale of Two Polls

November's news from Honduras has brought us piecemeal reporting on the results, though not all the data, from two different opinion polls in Honduras. These show an interesting evolution in opinion in Honduras, though you couldn't tell that from the press coverage.

The first poll, carried out by Vanderbilt's Latin American Public Opinion Project (LAPOP), is covered in recent stories in El Heraldo and La Prensa. Press coverage was sparked by the public release of the LAPOP report Cultura politica de la democracía en Honduras 2010, part of their Americas Barometer series. [While the link to the PDF appears to be working erratically, the technical information page establishes the information used here.]

This report is based on data from a poll carried out in January and February of 2010, just after Porfirio Lobo Sosa was inaugurated. LAPOP describes it as a stratified and clustered sample of 1596 respondents, with a margin of error of 2.75%. Previous commentary on the poll appeared in LAPOP's Insights series in August.

LAPOP has been issuing press releases based on analyses of the data since at least April. We covered an earlier one in this post.

On November 8, El Tiempo reported the results of a Centro de Estudios para la Democracia (CESPAD) poll carried out in September, 2010. This poll, carried out by a European polling agency and funded by OXFAM, surveyed 800 respondents and bears a reported margin of error of 0.07%. Their report can be downloaded here.

Polls measure sentiment at the time they were carried out, which can stay relatively stable or change in response to changing knowledge or conditions.

Poll results can also be influenced by the way questions are asked, which is why analysts always need to see the specific wording of a question. At its extreme, so-called "push-polling" involves asking questions in ways that, while purporting to solicit opinion, actually seek to change opinion. This verges from spreading outright falsehoods to simply framing an issue in ways likely to give people a bad impression and raise negative feelings.

Push-polling is considered a political strategy, a way to distort people's ideas. But there is a broader literature on polling that asks whether polls themselves can influence opinion, either through "bandwagon" effects (or contagion-- moving to the winning side), or through what in elections is called "strategic voting": making a choice that is not the one you really most support because your assessment is that what you most support has little potential of happening. Research results do support the existence of such strategic opinion shift.

So polls need to be considered historically and with attention to the specific questions asked.

In the wake of the inauguration of Porfirio Lobo Sosa, the LAPOP poll from January/February question PN4 asked "In general, would you say you were very satisfied, satisfied, unsatisfied or very unsatisfied with the form of democracy in Honduras?" La Prensa reported that 53.6% of their respondents were satisfied with Honduran democracy, and a further 12% were greatly satisfied. 31.6% told LAPOP they were unsatisfied, and a further 2.9% were very unsatisfied.

In its poll, taken in September, CESPAD obviously did not ask precisely the same question. Instead, it included two questions that generate comparable opinion; and those opinions completely reverse the picture of satisfaction LAPOP's earlier poll presented.

CESPAD found a huge level of dissatisfaction with the condition of democracy in Honduras. In answer to the question, "Are you satisfied with the current democracy or think that things should become more democratic?", only 14% of respondents reported that they were satisfied. A full 86% reported that they were unsatisfied and felt things should become more democratic:

Satisfied/Very Satisfied

Unsatisfied/Very Unsatisfied







In September, only 15.4% of the respondents felt that Honduras had overcome the crisis set off by the 2009 coup, with a further 8% thinking Honduran democracy was functioning normally. The remainder either felt that Honduran democracy was in crisis (50.0%) or felt Honduras had no democracy (21.4%) or weren't interested (5.1%).

Compare this to the January/February LAPOP survey question HONCRSPOL7: "How satisfied were you with the solution to the political crisis of 2009?". Note that this question contains the assumption that the political crisis had ended, presumably with the election of Lobo Sosa in November 2009. In early 2010, LAPOP reported that 59.6% of respondents were satisfied with whatever they saw as the solution to the political crisis.

If we take these two questions as measuring about the same thing, then sentiment about whether the coup of 2009 has been resolved has eroded from 59.6% in January/February to 23.4% in September. A much higher proportion of the population in September thought the democracy was in crisis or there was no democracy in Honduras (71.4%) than believed in February that the crisis had been resolved (23.4%).

In January/February LAPOP followed their question about how satisfied with the "solution" to the crisis people were with another (HONCRSPOL8) that asked respondents what their preferred solution to the crisis was. The questionnaire provided a predefined list of answers that were to be scored, along with the instruction that the list of potential answers was not to be read to the respondents. The list of preferred solutions to the crisis did not include holding elections.

Here, 42.2 percent of the LAPOP respondents preferred reinstating Zelaya in some way for a period of time (31.2 % until the end of his term in January 2010, while a further 11 % said restore Zelaya and allow his reelection).

29.8% thought Micheletti should continue in power, another 11% thought a third party should be appointed to head the government as a solution.

Finally, 8.1% thought the solution to the crisis of 2009 involved trying those who broke the law.

All of the other proposed solutions were favored by under 2% of those surveyed.

In contrast, CESPAD found that in September, 76% of their respondents felt that a solution to the crisis would require an agreement among all the actors in the crisis. 65% felt that this was the only way to solve the crisis, and a further 11% felt that it was necessary but would not resolve all the underlying issues.

These respondents didn't feel the crisis was over in September, and it is likely that they would define "the crisis" differently than respondents did in January and February. Then, the proposal being put forward politically by everyone from the newly elected Honduran government to the US State Department was that the November 2009 elections had put the coup behind. Now, half a year further on, people can see that has not really happened.

Changes in popular opinion between the two polls suggest the campaign to promote a constitutional assembly by the Frente Nacional de Resistencia Popular (FNRP) is bearing fruit.

In January/February, LAPOP asked "Are you in favor of forming a constitutional assembly?" (HONCRSPOL5). A largely majority of their respondents were not in favor of a constitutional assembly (70.5%) with only 29.5% in favor.

In September, CESPAD found that things had turned around, so that when they asked "Are you in favor or against convening a constitutional assembly?", a majority of their respondents, 55%, were in favor, with 45% opposed.

As we have noted in other posts, in recent months politicians from all the major parties have begun to single willingness to talk about changing Article 5 of the constitution to allow polling the public to find out their opinion, and Lobo Sosa is willing to talk about holding a constitutional convention.

Public opinion changes.

LAPOP's press releases this week were based on old data. There is newer, and therefore, more relevant, opinion research available, which tells a different story.

That's the tale of these two polls.


Tambopaxi said...

LAPOP is well-known outfit with years of experience in Honduras.

I googled CESPAD by its full name and acronym, and cannot find any info on who/what this outfit is; google keeps referring back to El Tiempo, the HCP blog and other sites which refer to the El Tiempo article. Does CESPAD have its own site with info on its work in Honduras?

RAJ said...

CESPAD is a Honduran NGO located in Tegucigalpa.

The principal researchers are a political economist, Gustavo Irias, and Eugenio Sosa, a sociologist. Both previously worked as researchers for SNV, the Netherlands Development Organization. Sosa has published academic work on civil society in Honduras.

In June of this year, Sosa took part in a program on democracy and governability at UNAH. The program described his qualifications and experience as follows:

José Eugenio Sosa Iglesias (hondureño): Es Licenciado en Sociología por la Universidad Nacional Autónoma de Honduras UNAH, candidato a Maestro en Ciencias Sociales por la Facultad Latinoamericana de Ciencias Sociales FLACSO-Guatemala. Coautor de varios libros sobre sociedad civil y procesos electorales con el Centro de Documentación de Honduras (CEDOH), ha sido columnista de diarios nacionales y colaborador de la Revista Política de Honduras. Conferencista sobre temas de democracia y gobernabilidad.

In the report on their survey (which is linked to our post), they described the objectives of CESPAD as

To generate knowledge and promote dialogue and debate for the construction of a more democratic, inclusive and participatory society.

This is the first survey undertaken by CESPAD, and follows on the heels of important analyses previously published of the coup and its aftermath.

While they do not have the long history of conducting surveys that LAPOP does, they do have the credibility that comes with research experience. Their work has been financially supported by a variety of NGOs, in the case of this survey, OXFAM.

As we note here, the apparent contradictions between the two polls only exist if you assume that people's opinions don't change, and ignore the chronology involved. CESPAD is not sampling the same opinion; it is sampling opinion after months of the Lobo Sosa administration during which there has been significant public coverage of the rationales for a constitutional assembly.

phoenixwoman said...

I don't think the margin of error is 0.07%, which is improbably small for a sample of 700. This is stated to be for a confidence limit of 93% (unclear why they chose this instead of the usual 90 or 95%), with an uncertainty, as stated, of 0.07, i.e. 7%.

RAJ said...

Seems likely that you are right.

boz said...

At the 95th percentile, margin of error on a sample of 800 people for a voting population of 4.6 million is about 3.5%.

Thanks for the followup on Tambopaxi's question about CESPAD. I wasn't familiar with them either.

dianabarahona said...

Checking the original LAPOP question on the "solution to the crisis" the wording is ¿Qué tan satisfecho quedó usted con la solución a la crisis politica del 2009?

This is grammatically correct, meaning "the particular solution" to the crisis that was implemented. However you are correct in that there is a bias in that the question implies that the crisis was in fact "solved." It could easily be misunderstood as meaning "How satisfied are you with the fact that the crisis is over." Beyond that, to ask people who have never lived in a true democracy to rate how democratic their system is, is push polling.

The very fact of taking a poll represents bias, because each of the social classes wishes to demonstrate certain tendencies. The ruling class is trying to show that it is winning legitimacy--that it is gaining in attempts to dominate by consent. On the other side of the class war, the subordinate class wishes to show that the ruling class is losing hegemony.

The coup showed once again (as if it were not already obvious) that the ruling class supports "democracy" only so long as "democracy" keeps it in power. It supports elections only as long as it wins. The minute it begins to lose its grip on reins of the political state, meaning control over economic policies, "democracy" goes out the window and CIA experts disguised as ambassadors go to work. And to keep the this option available, the United States is beefing up all of the militaries of the region with training and funding. Just in case.