More polls than ever? Huh... no.

This year, opinion polls appear to be everywhere. You can’t browse a news website without stumbling on an article about polls for president, for the Senate, for governors. There are surveys about rejection rates, how each candidate is faring in different regions, genders… You get the picture.

Despite this sensation of opinion poll saturation, the 2018 general election hasn’t actually seen an increase in surveys carried out, it just feels that way. Every time institutes carry out a poll, they must register their surveys with the Superior Electoral Court. Failing to do so is an electoral crime. If the poll has not been previously registered, its results cannot be published. With this data, we can say for sure that no, there haven’t been more opinion polls in 2018.

Let’s go to the numbers:

Four years ago, 485 surveys were registered by 103 companies between January and September 15. In 2018, the number dropped to 300 surveys by 69 institutes.</span></p> <p><span style="font-weight: 400;">There is a higher concentration of surveys from the top polling institutes, more than ever before. While in 2014 the top 10 pollsters were responsible for 45 percent of polls, their share of the market jumped to 58 percent this time around. Small institutes that had conducted one or two polls in previous electoral cycles are not absent from the picture.</span></p> <hr /> <p><img class="alignnone size-large wp-image-9204" src="" alt="There haven't been more polls in 2018 - we're just paying more attention" width="1024" height="580" srcset=" 1024w, 300w, 768w, 610w, 1412w" sizes="(max-width: 1024px) 100vw, 1024px" /></p> <hr /> <p><span style="font-weight: 400;">But business is not so rosy among the top players, which tend to conduct much more expensive surveys. Ibope did 57 presidential polls this year, in comparison with 86 in 2014. The number of Datafolha polls was cut in half in four years: from 18 to 9. Those firms are usually hired by media companies, which are in a </span><a href=""><span style="font-weight: 400;">much bigger crisis</span></a><span style="font-weight: 400;"> now than they were in 2014.</span></p> <h2>Disruptive players</h2> <p><span style="font-weight: 400;">Traditional institutes are also facing more competition. Out of the 10 biggest pollsters of 2018, five did not exist or conducted no first-round polls in 2014. The disruptive players include Ipespe, a traditional institute located in São Paulo that was hired by XP investment bank to provide them with presidential polls. With 23 surveys as of September 15, they are only behind Ibope. </span></p> <p><span style="font-weight: 400;">As </span><b>The Brazilian Report </b><span style="font-weight: 400;">has shown, these new polls have been challenged from all sides: big pollsters </span><a href=""><span style="font-weight: 400;">question their methods</span></a><span style="font-weight: 400;"> &#8211; telephone interviews (sometimes through automated calls) &#8211; and economists question who is publishing them. </span><a href=""><span style="font-weight: 400;">Polls impact currency trading</span></a><span style="font-weight: 400;">, as the markets quickly react to the latest scenario. And banks who sponsor them have access to the results beforehand. Economist Monica de Bolle, for instance, questions if this could be an excuse for insider trading.</span></p> <p><span style="font-weight: 400;">Another controversial player is Paraná Pesquisas, which conducted 17 polls in the first round. In 2014, they were <a href="">scorned</a> for releasing a second-round poll that was wrong by almost 6 points (way outside the margin of error). Most of their surveys are registered as being self-funded &#8211; a traditional way to hide who is paying for the poll. </span></p> <p><span style="font-weight: 400;">Paraná Pesquisas has an interesting strategy to gain relevance. They usually offer their results to media companies as scoops &#8211; and short-note political columns publish them acritically. They seem not to know that, if you’re not paying for a service (in this case, for the poll data), of course, you are the product.

Read the full story NOW!

PowerSep 29, 2018

Tags: - -

BY Marcelo Soares

Marcelo Soares is a Brazilian journalist specializing in data journalism and reader engagement.