Fake Checking - What is the margin of error accepted by "investigative" agencies?

It is not uncommon to see "investigative" agencies or even traditional journals publishing what they call data-checking. In many cases, screening is used to promote or prejudice the investigator. Especially when the check is about presidential candidates.

It is a fact that most of the presidential candidates walk around Brazil giving interviews, giving lectures, and attending events day and night in the pre-campaign period. At a rate like this it is at least natural that small deviations occur when presenting some data (unless the presidentialists go around with powerpoint presentations and reading scripts and manuals).

Even so, it is curious to note that the data checks seem to pick and choose the "facts" to be verified. In the last check made by the Public Agency, Geraldo Alckmin had eight sentences verified, being five true, two false and one exaggerated. This time came the time of Ciro Gomes. Of his eight sentences verified, they were five false, one exaggerated, one impossible to prove and one debatable (later they ran behind and a false turned true).

Eight sentences were chosen to be verified based on two events: an interview with Radio France International (duration of 23 minutes) and a lecture at the University of Sorbonne (lasting 70 minutes). Of a total of 93 minutes of speech, eight sentences were handpicked - first let's look at the chosen phrases, then come back to the choice of phrases.

The biased title of the story is: When talking about Brazil, Ciro Gomes uses false and exaggerated data! This title in itself is already false and exaggerated, let's face it. Let's talk about this:

First and foremost, we note that the evaluation considers the following categories: FALSE, EXAGEDED, DISCUSSIBLE, IMPOSSIBLE, PROOF, TRUE . At the same time, the agency has no parameters to categorize the exaggerated, does not present a tolerable margin of error, and nothing like it.

"[Brazil had] 64.700 homicides in the last twelve months."

The agency says it is impossible to prove. In the article she states that the actual data for homicides for the year of 2017 point to 61.238. Ciro missed here by 5,3%. Ah, but the data is 2017 and Cyrus speaks of the last 12 months.

Friends readers, really? How much would be a tolerable margin for you? If only they had classified as "TRUE, BUT ..."

"The escalation of organized crime and criminal factions produces an impunity of almost 92% - only 8% of homicides are cleared up."

The agency says it is false, despite being a fact published in an official document of the National Strategy of Public Security (Enasp). The average rate of reported homicides, according to the reference document for the National Public Security Strategy for five capitals is from 0 to 1%. The Sou da Paz Institute, a reference of the Public, stated in a recent study that approximately 2% of crimes are not resolved in Brazil.

As much as the data presented by Cyrus is below the 10 to 15% presented, does this in no way diminish Cyrus's argument that in Brazil the crime is unpunished?

"Absolutely explosive debt today has R $ 1,184 trillion, that is 26% of public debt, maturing in 4 days."

The agency says it's fake. The R $ 0 trillion indicated by Ciro Gomes, which expires in 4 days and corresponds to more than 1 / 4 of the public debt, whose maturity may lead Brazil to a banking crisis, is debt contracted / rolled by the Central Bank in Committed Transactions. Such a type of operation is not accounted for by the National Treasury (which issues debt itself), but by the Central Bank, which holds such an instrument, in order to regulate the liquidity of the Financial System (Banks). However, the repurchase agreement is a debt with a commitment of payment / rollover, in very short term, by the Central Bank, which belongs to the Union.

The information is in the Independent Fiscal Institute of the Federal Senate, in a document published by Josué Alfredo Pellegrini. Of false, that has nothing.

"The Welfare Deficit, after deducting its resources from the DRU [Union Unbundling], reached R $ 180 billion last year."

The agency says that the information informed by Cyrus is false. In fact, instead of 180 billion, Cyrus should have claimed that they were 192 billion. Ciro missed here by 6,67%.

Once again the question remains, what is the tolerable margin of error for the Public Agency?

"Brazil is de-industrializing like no other country in the world."

The agency says that this is debatable. First agrees with Ciro when doing the analysis from 1960 to 2016 where industry share fell from 29,6% to 11,7%. He then states that this is not so evident as we observe an increase in industrial output in 2012 and 2018. Is it intellectually honest to compare periods so different? Cyrus was talking about a period of 30 years, not a period of 5 years.

After that, the agency still shows the UN Report that shows Brazil with the largest industrial decline in all the countries studied. And it presents a lot of other data that more confuses the reader than they clarify. They call it debatable.

Everyone who studies industrialization knows that physical production is not the indicator used. Indicator of industrialization is value-added industry participation in total added value or share of employment in industry or total employment. Moreover, the "experts consulted by the agency" are even experts in the field? Or are those that already have the expected answer to the subject?

"When Lula takes possession of the exchange rate at today's value it was R $ 9,20. And Lula delivers to Dilma at a constant value with an exchange rate of R $ 1,75. That is, in the constancy of the Lula government the people increased by 4 times the capacity to consume. "

The agency says it is exaggerated. Considering the indexes used by the agency, the values ​​would be R $ 8,54 and not R $ 9,20 as stated by Ciro Gomes. And in the Dilma Government the rate would be R $ 1,65 and not R $ 1,75. Ciro missed by 7,2% in the first and 5,8% in the second.

The appreciated currency increases the purchasing power of the population in the short run, as it has in recent years, but in the medium term it increases imports and harms local businesses. Everyone is happy with falling import prices and cheaper trips abroad, but in the medium term everyone loses their job. And that's what happened in Brazil.

Even assuming that the indicators used by the agency are the only ones, how can 7,2% be exaggerated, but in the other items above with errors less than 7,2%, the statement is considered false?

"The federal government makes an intervention in Rio, but it embarrassed Rio de Janeiro, in a recent agreement for the fiscal collapse, not to replace retired police officers. It is forbidden to make a contest in Rio. "

The agency says it's fake. In fact, Ciro has slipped in saying that Rio can not replace retired police officers. But then to say that the whole sentence is false? Who will not agree that Rio de Janeiro was embarrassed? It became clear that the state can not make contests in general. If one considers the entire sentence, perhaps the classification here should be exaggerated. Or you could just select the part of the retired and classify as false.

"Brazil has the most perverse distribution of income from all the organized societies I have observed. Look at the number: five people have a fortune equivalent to 100 million nationals, which makes up half of our population. "

The agency had considered the false data not to be 5 people but 6. But now, after thousands of hits, it has updated. Now they say the data is true. The only true. Even so, how do you reverse the impact of thousands of people who actually read that it was false?

We return now to the selection of phrases to be checked.

When critics study Ciro's speech and can only point to these "exaggerations", it is a sign that Ciro Gomes is, at the very least, walking alongside the truth. Imagine that, according to the agency, two moments were selected for investigation. The two together total about 90 minutes of speech.

The ninety minutes of speech consists of important and profound speeches about:

  1. The position of the PSDB in relation to corruption in Brazil
  2. The trial of former President Lula and his possible candidacy
  3. The escalation of organized crime and criminal factions
  4. The bankruptcy of institutions in the state of national law
  5. Procedural times in Brazilian courts and the law of the clean sheet
  6. Political history of Brazil and recent events of Brazilian politics
  7. Current electoral situation and political polarization in Brazil and the world
  8. Attributes of Bolsonaro and analysis of its supporters
  9. Legal protagonism, institutional "anarchy" and collapse of political power
  10. The Conservative Project of Policy Abomination
  11. The redemocratization agenda
  12. Exchange devaluation after the re-election of FHC
  13. Movement Out FHC
  14. The purchasing power in the Lula government
  15. Brazilian elite class bias
  16. Devaluation of the exchange rate in the Dilma government
  17. The 80 lawsuits for moral damages
  18. Party loyalty x coherence of ideas
  19. Ceiling PEC
  20. Denial of Constitution
  21. Schizophrenia of the Brazilian bourgeoisie
  22. Commodity price
  23. Ministry of Finance and Central Bank
  24. Happiness regarding consumption
  25. Brazilian foreign policy
  26. The conflict in Venezuela
  27. The design of the National Development Project
  28. The social protection network in Brazil (minimum wage and social expenses in Brazil)
  29. ProUni, FIES, tax waivers
  30. The crisis of 2008 in the USA
  31. The pedagogical paradigm in Brazil and the international rankings of education
  32. The overall financing conditions
  33. Bilateral relations between the US and China and their conflicts
  34. Population growth in Brazil and job demands in the country
  35. Brazil's tax system in the world
  36. The fiscal situation of the Brazilian states
  37. Human Rights and Identity Appeals
  38. Government of Dilma Rousseff and the Impeachment
  39. Brazil's prison system
  40. Universalization of basic education in Brazil
  41. Employment problems in developed countries
  42. Governance without the PMDB
  43. Socioeconomic conditions of the Northeast
  44. Credit expansions and retractions in Brazil
  45. What is being left?
  46. History of the Real Plan

That's it. All of the above points were debated in 90 minutes of conversation. Of all these ninety minutes, the agency chose 8 phrases to point to "exaggerations and lies".

However, since the publication of the matter the agency has already made two updates correcting errors in the verification of data. And we here, we try to show that the verification of data made by the agency is not impartial, has no defined methodologies, and most importantly: it is wrong. Wrong when searching little in your data check and wrong when taking context sentences.

In addition, it is striking that they talk only to people who are not in the debate and have opinions contrary to those of Cyrus.

From 8 phrases pinched by the agency amid a sea of ​​debate, absolutely all were checked against Cyrus. An "IMPOSSIBLE PROOF", five "FALSE", a "DISCUSSIBLE", a "EXAGERATED". However, the presidential candidate had a margin of error of less than 10% in four. Possibly he missed a little more than 10% in another, had a fact erroneously classified as false (later corrected), had one fact hardly investigated and actually slipped into only one.

And it is good to point out one thing: Ciro Gomes will not be the only victim of this. This will come on top of all the presidential candidates. We have to be careful about what lies ahead.

Instead of the FALSE, EXAGED, DISCUSSIBLE, IMPOSSIBLE, TRUE, and TRUE thermometer; how about using a thermometer with the following ratings:

  1. True
  2. True majority
  3. Exaggerated
  4. No Fundamentals
  5. Sluggish
  6. FALSE
  7. Misleading advertisement

It would also be worth presenting a classification methodology for each of the items, presenting acceptable error margins and other basic requirements so that they can do the categorizations in a transparent way. We say this because, although it may not have been the intention, for some it may seem that the latest verifications are being used as propaganda tools.
 

Leave an answer

18 − dois =