quarta-feira, 14 de setembro de 2011

Brazilians and Apes: a Criticism about a The Economist Reading

The Economist, the weekly bible of the liberal world goes a bit uncomfortable with the relative loss of international status of their country of origin as well as of their peers in developed countries , and is striving to make people believe that the dominant perception of a world economy driven by BRICs, (acronym for Brazil, Russia, India and China) is not what it seems. Hence it has published in its current issue (link below) attractive infographic that seeks to demonstrate what it sees as the fallacy of emphasis given to the Brazilian economy in the world press.
The trick is simple: the magazine's infographic allows the reader to compare with a simple touch of the computer "mouse" the per-capita GDP (gross domestic product per capita and an assumed indication of well-being of populations ) of each Brazilian State to other poorer countries but with equivalent GDP. It turns out that the most developed state in Brazil, Sao Paulo, appears compared to Poland in such way that the continued declining ranking goes on until compares other fewer income brazilian states to the poorest countries of Africa and Asia.
 As always happens to the numbers, you can relate them in every way, extracting from them the results you may desire. The magazine editorialists rely their expectation in doing so kind of magic exercise upon the crude knowledge of readers about statistical techniques that, while it is true, does not suppress the bad faith of the magazine in promoting the distortion of reality that the confrontation of the numbers produces indeed.
 Needless to say that the material was hailed as "brilliant" or "revealing" by the cult boys whose Blogs Brazilian newspapers keeps hold as an appendix of their own. This was true in the case of the pseudo blogger Jose Roberto de Toledo, of the "O Estado de Sao Paulo" journal that spread among the national that infographic. Gossip aside, it is about the lack of science and especially the clumsy design imprinted by of the graphic piece what I'd like to comment.
 There are two ways of comparing a man and a gorilla (by the way, the depreciative word with which some Europeans like to name latin americans). The first is to compare the weight of a significant group of individuals of both species to verify if we can find a distinctive pattern of comparison that allows an individual to be categorized as a man like or a monkey. The surprise is that on this question, men and monkeys look very similar: there are human beings who weigh the 150 pounds of gorilla and apes also weighing the 80 or 90 kg of an adult male.
 In order to avoid push too far the comparison, we could still to draw upon another attribute to differentiate the species, height. But this criterion also present common outcomes to both categories since there are men with 1.50 m and monkeys 1.70 m high. Despite these dissimilarities, it would be possible to show a profile demonstrating that in a certain place on the planet residents have very similar characteristics to those of their ancestors of the savannas.
 Yet there is more honest way to relativize the Brazilian economy (which just went up 5 positions in the ranking of the world's most competitive displacing the UK itself which came down three steps in the same classification) not by quartering and comparing each piece with what it is not a part too.
 The most ethical and scientific comparison would be by the means of what economists call a principal component analysis. In the comparison between Brazilian and primates could be considered the main components, the characteristics of one species and another that mattered to differentiation, such as weight and height.
 In the case of economies itself would matter significant indicators of performance (for instance, some of them catched from those contemplated in the elaboration of ranking of global competitiveness) seen not one by one, but in sets against sets of the same type of information. What would produce a third variable, intended to eliminate the largest divergences inside each isolated group of characteristics and making possible a status comparison more valid technically.
Complicated? Not at all. To better illustrate the greater relevance of such a kind of thinking over that used by the liberal Yuppies magazine, imagine a big house in Morumbi (upscale neighborhood of São Paulo city), which was occupied by members of a same family. If the intention was to assess the level of wealth of the family living there would make no sense to divide the house in every of its dependencies, and compare it, in extension and average income of its occupant, with the same variables of a two floor house in Vila Formosa ( a poor neighborhood in the same city) .
The right way is to compare the entire big house with all por ones which would be equivalent, grouping the variables of different questions (such as light availability, water and average household income) in a single measure and then submit such homogeneous measure to comparison.
From my humble gorilla position reader I will write to the editors of rag to warn them about that innocent oversight. Go there the link of the near-trick wich transforms an emergent country in a lot of Zambias
If you read portuguese come to my site at Brazil That Goes

Nenhum comentário:

Postar um comentário