When trying to assess a validity of a claim relying on statistics, I was taught (in the school of epidemiology) that the scale to use is “the pyramid of evidence

alt text

However, when conducting a discussion in the field of economics or political science, it is often very difficult (to impossible) to recreate a political situation so to allow experimenting.

Since this is the case, my questions are:

  1. can this be mitigated in any way?
  2. In what way can (or does) political/economic disciplines build their arguments (using statistics) in a valid way?

The motivation for my question started from reading a discussion about the privatization of the academic system in Israel. e.g: Should the charging of students be the same for all disciplines or different. Should the running of the university be performed by the academic staff or by outside managers - and so on.

One things that appears to happen in the discussion is that the people supporting the privatization (economists mostly) seems to be using statistics of all sorts to support their claims. While the people from the other side of the argument are not as equipped with them (usually people from the humanities).

The economists seems to blame the humanities of not using numeric data. While the humanities accuse the economists of being to speculative - and that the data they bring is open to too many interpretations.

I am trying to understand if these two groups of disciplines can create a more fruitful dialog, or is it just not possible due to the complexity of the subject matter, and the limitation of our control over it.

The following discussion has a nice link in it, but it was stopped quite early:

Also a bunch of good relevant discussion where introduced here:

p.s: Due to the subjectivity of the topic - I am marking this as community wiki.

  • $\begingroup$ Can you explain why this is a pyramind and not a triangle? Or a triangle and not a ladder? I'm not being snarky -- I'm wondering if I'm missing something significant that this visualization is trying to convey. $\endgroup$ – ars Sep 28 '10 at 5:03
  • $\begingroup$ Hi ars, it could well be the triangle of evidence. The reason for the word pyramid is probably historical, I am simply using it as it was taught to me. BTW, it shouldn't be a leader, since the width of the triangle also reflects the abundance of such an evidence in practice (Or at least, that's my guess :) ) $\endgroup$ – Tal Galili Sep 28 '10 at 21:09
  • 1
    $\begingroup$ I hadn't come across this before, so just curious for any insight into what it was saying. Your point about abundance of evidence makes sense. Thanks. :) $\endgroup$ – ars Sep 28 '10 at 21:32

Way back in 1965, Sir Austin Bradford Hill wrote a great essay about something very akin to the Pyramid of Evidence, where he discussed how the piling up of evidence can increase our confidence in hypotheses of causality in Medicine.

Most of the factors he discusses can be applied to Economics and political sciences.

  • 1
    $\begingroup$ +1 for linking to the full essay rather than one of the all-too-common oversimplifications - it's not long and is still relevant today and well worth reading in full. Hill's criteria for causation are about much more than just pyramid / piling up of evidence though. $\endgroup$ – onestop Oct 19 '10 at 15:39

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.