Find the revolution

I have a challenge for you. The chart above displays the evolution of infant mortality rate in Portugal. The years in the x-axis are not labeled on purpose. In one of those years there was a left-wing revolution. Left-wingers say that a sharp decline in the infant death rate is one of the “conquests” of the revolution, while right-wingers say that the trend was already there and the revolution didn’t change anything (for this specific variable). If you can read Portuguese here is the post I’m copying the idea from.

The original question was something like this: based on this chart alone, can you tell me when the revolution took place?

To this question, I’d like to add a few more:

  • Given the nature of the variable, should we see a dramatic change after the revolution?
  • Is this an innocent or a loaded/biased question?
  • Is there a way to answer the first question, perhaps using more data?

I have my own answers to these questions (hint: I hate this chart and I find it biased beyond redemption), but I’d like to hear from you first 🙂 Please share your thoughts in the comments below.

Here is an Excel file with the data for Portugal and the European Union (Euro17): infantdeathRatePortugal

[fusion_builder_container hundred_percent=”yes” overflow=”visible”][fusion_builder_row][fusion_builder_column type=”1_1″ background_position=”left top” background_color=”” border_size=”” border_color=”” border_style=”solid” spacing=”yes” background_image=”” background_repeat=”no-repeat” padding=”” margin_top=”0px” margin_bottom=”0px” class=”” id=”” animation_type=”” animation_speed=”0.3″ animation_direction=”left” hide_on_mobile=”no” center_content=”no” min_height=”none”][Update]

First things first: you may not even know when the revolution took place, while this is something I cannot avoid to know. This may explain different approaches to the data.

Here is the same chart, correctly labeled. The revolution took place in April 25, 1974, 40 years ago:

intant-mortality-ratio-2

Infant mortality rate declines since 1961, and the impact of the revolution is not visible at all. Is that surprising? Hardly. Population data changes slowly, and new public health policies usually have no immediate impact.

What really maters, from a political perspective, is not the decline itself, but the rate of convergence towards a commonly accepted goal. The ratio between Portugal and European countries (Euro 17) allows us to check if the rate before and after the revolution changed significantly. I superimposed a 2-year moving average to make things clear in this chart:

intant-mortality-ratio-vs-europe-2

As you can see, for 10 years (1962-1971) the rate of convergence remains flat at a very high level (I find this criminal, by the way). Then something happens in 1972. In 1973 the rate goes up again, but not to the previous levels. In 1974 it also declines, but I don’t think we should thank the revolution yet.

We would need more data regarding new public health policies to have a better understanding of what happened between 1972 and 1974. From this dataset alone, I would say that both leftists and righwingers are wrong. There is no long-term decline in infant mortality rate when compared to other European countries, no convergence whatsoever for far too long. On the other hand, apparently the sharp decline began in 1972/73, before the revolution. You can’t say that a lower infant mortality rate is a “conquest” of the revolution, you can only say that it kept the trend downward.

Yes, if you torture the numbers long enough, they will confess. But things are usually complex, and if you try to make them too simple you’ll draw the wrong conclusions (or you are trying to push your agenda).

Corollary: never trust a chart published in a political blog, or at least take it with a grain of salt.

 [/fusion_builder_column][/fusion_builder_row][/fusion_builder_container]

3 thoughts on “Find the revolution

  1. Jorge, as the author of the original post I would just like to make a correction. the original question was not “based on this chart alone”, but based on 5 different charts with 5 different variables. Obviously, the best way to assess the point in question is to map the convergence against developed economies which would pretty much convey the same message (although I couldnt find data for all the variables I wanted to show).

    Thanks for reading in any case.

    Carlos

  2. Carlos, thanks for your comment. I’ll remove “alone”, since in theory the analysis of the five charts taken together could indeed improve the chances of finding the correct data point (or reinforce the idea that nothing actually changed). I’m focusing on infant mortality rate because it’s your first chart and because, as you know and I say above, it has a special meaning for leftists.

    That said, you recognize that there should be an analysis of convergence against developed economies, but you dismiss it because it “would pretty much convey the same message”. That’s at the heart of your analysis, so I’d like you to show us that there was no change. The oposite is true, as you’ll see in the update.

  3. What I like about Jorges article is, that we need to think harder. Charts can be so dangerous, because we usually believe what we see. In business and public we lack of basic statistic and analytic thinking. A very dangerous situation for rather complex societies. Thanks Jorge.

Comments are closed.