Ryan Armstrong
Change management. Learning. Maps.

RRREIC in the news

30 October 2019

From a recent Guardian article, here is a realist account of science if I have ever seen one, delivered from now Nobel Prize winning economists:

But good economics is much less strident, and quite different. It is less like the hard sciences and more like engineering or plumbing: it breaks big problems into manageable chunks and tries to solve them with a pragmatic approach – a combination of intuition and theory, trial and acknowledged errors. Good economics starts with some facts that are troubling, makes some guesses based on what we already know about human behaviour and theories that have been shown to work, uses data to test those guesses, refines (or radically alters) its line of attack based on the new set of facts and, eventually, with some luck, gets to a solution.

I would say we could clarify a few of their points (and I have no idea where they stand on these):

  1. All good science, not just economics, starts with some facts that are troubling, and therefore represent a deficiency in the way things are being done.

Like botany developed from a lack of knowledge about plants. The “troubling fact” could have been that people kept eating plants that were making them sick.

  1. The approach is pragmatic in its criteria of evaluation, which is that the theories are judged by their ability to address the deficiency.

They are judged by their explanatory power, not their statistical significance, though statistics may be useful in evaluating data.

  1. Intuition and theory are at play as we “make some guesses” in a process that has been called “abductive redescription”.

Trial can and perhaps should be sometimes separated from the guess making process, so that the intuition and theory can be made explicit (and subsequently put to the test).

  1. “Acknowledged errors” is really “acknowledging errors”, in that good science involves questioning what we know, and actively recognizing the limits of our own knowledge, in a sometimes painful process that may or may not involve Reviewer 2 but ultimately, with some luck, involves a solution that represents a marginal improvement to the way things are done.

We should not mistake the real world with the our knowledge of it.