Global Courant
SUBSCRIBERS EXCLUSIVE
To make a focus group sample, some consultants give away electrical appliances.
Pollsters already know that out of every ten automated phone calls made by their computers, only one person will answer a pre-election survey. If the call is made by an operator, the number of responses increases a little: one in 7 responds. The most effective, and most expensive, way is to do surveys the old-fashioned way: go out and ask people. In that case, five out of ten people respond to the survey. Still, the response rate remains the lowest ever.
To try to improve that performance, a large consultancy -and which has the contracts that allow it to increase its budget- decided to tempt the respondents. In the nationwide sample that is being carried out these days, the pollsters offer each person who says they do not want to answer a 1,500-peso voucher from a supermarket or business in that area so they can answer. The most remarkable thing is that of those five initial rejections, only one gets the voucher. The other four leave. To make it better understood: in the midst of a very serious economic crisis, four out of ten Argentines reject a $1,500 gift in order not to spend a few minutes of their day talking about politics and saying who they will vote for next month.
Payment for responding was also extended to qualitative analyses, the other means used by politics to find out about the vagaries of public opinion. The typical scenario of these studies is the focus group, a meeting of people called by a consultant in a closed office who talk and answer slogans proposed by a moderator. To be useful, this group must be representative of the whole of society, or of the sector of society that the client who pays for the study wants to know. Thus, if half of the register is made up of women, half of the group will be women, and the same will occur with respect to the income strata or levels of education of people of voting age. In this case, several consultants detect that it is very difficult to make a good representative sample if they do not reward those who participate in the study with an appliance.
Do these difficulties speak of anger with politics, with pollsters, or with both? Be that as it may, politicians know that they are groping. They know that the lens that allowed them to listen to the dreams, anger, opinions and preferences of citizens has a black spot that occupies almost half of its surface and that the efforts made by pollsters to reduce the defects in their survey tools measurement have a limit.
The big question today is who makes up that invisible sector. Do they vote the same as the sector that can be surveyed? Can the opinions of those who answer be projected onto the rest who do not want to answer, or is the hidden group so different that any projection is fallible? Does the group that answers do so because it is more politicized and has a more extreme opinion or do centrist voters also respond with the same disposition?
Pollsters try to spy on the dark side of the Moon by comparing polls from previous years with comparable levels of rejection with the results of the votes of those years. It is a way of trying to control the reliability of the method, but, as the mechanism has already had several electoral turns with failures, sometimes these attempts do not even satisfy those who put them into practice.
