Enduring Risks to Take the Pulse of the People

Some of the interviewers were robbed in broad daylight while surveying people on the streets of the Mexican capital. Others spent hours on boats, buses or burros to reach remote villages with no roads or telephones.

In all, it took 80 interviewers from Mexico City’s Reforma newspaper more than a week to do on foot in Mexico what 90 Times pollsters did in four days by telephone in the United States in a joint poll by the two newspapers.

The task for the pollsters south of the border: interviewing 1,500 people in 25 of Mexico’s 31 states in a polling process that differed as radically from U.S. methods as many of the responses did from opposite sides of the border.

“It’s a lot harder to poll scientifically in Mexico than in the United States,” said Reforma’s poll director, Rafael Gimenez Valdes, of his young but growing industry south of the border. “The biggest difference, of course, is the telephone system.

“In the United States, more than 90% of the country is covered by telephones. So you can get a representative sample by phone. Here, it’s about 40% in Mexico City. And in the rest of the country, only about 25% have telephones. So, to poll properly here, we must do it face to face.”

As a result, when Gimenez conducts a nationwide survey as ambitious as this one, done last month, he pays his pollsters to hit the bricks, dirt roads, waterways and mountain trails that lead to a representative sample of the population.

The resources needed for such a poll--Gimenez said his portion of the Times-Reforma poll cost about $14,000--are among the chief reasons that independent polling is a new industry in Mexico.

“During the last 10 to 15 years, there has been more and more polling in Mexico,” said Nancy Belden, of the U.S. polling firm Belden & Russonello, based in Washington, D.C. She has extensive experience in surveying south of the border.

Belden said the most significant change came during Mexico’s highly competitive 1994 presidential election. For the first time, the private sector commissioned several polls that came within a few percentage points of predicting the election results. Belden’s own poll that year for a consortium of banks came within 4 percentage points of the results.

“What 1994 proved for the first time is that if you have the resources you need to follow the rules of good data collection, you can, in fact, ask people what they think and get a true reading,” she said.

Gimenez agrees. His 10 years’ experience in polling here has shown that most Mexicans--once deeply suspicious of such things--are now even appreciative of interviewers’ long list of questions.

“For many years, we had no tradition of doing independent polls in Mexico,” Gimenez said. “People didn’t feel confident telling a stranger about their problems, their income and especially which party they wanted to vote for. Fortunately, in recent years we have had many changes.”

Elections are more open now, he said, and independent polls are much more frequent. Both trends have reinforced public confidence. “Now, especially outside urban areas, there are places where people are anxious to tell us their opinions,” he said.

But the risks to interviewers have not declined. Besides the difficulties of reaching remote villages, Reforma’s pollsters face an increasing threat to their physical safety in the cities.

Urban crime has soared since Mexico’s economic crisis began in December 1994, and Gimenez said he forbids his interviewers to work the streets at night. Even so, he said, 10 of them have been robbed in the past four months.

One in five Mexicans interviewed in the capital reported that they or a family member had been victimized by crime in the last 12 months, so those attacks on the pollsters were very nearly a representative sample themselves.