Exceeding data

Undergraduate psychologists attending statistics courses are told of many heinous errors: traps that lie before them in their future analyses.  They must not conflate correlation with causation, they must be wary of false positives and negatives, and they should interpret their data in light of experimental and null hypotheses, drawing no further inferences.  Watch out for multiple comparisons, and keep your Bonferroni close to hand. Beware collinearity! And so on, and so forth.

This is admirable and sound.  However, the other product in Psychology - written argument - is rarely treated to the same excoriating scrutiny. Indeed, much of the formative and summative feedback that is delivered to students is about grammar and the communicative impact of their piece. To some extent this is because a large part of what is written accompanies a laboratory experiment, and the formalisms of the experiment and its statistical analysis are thought to lend themselves to the structure of the report: here is the statistically expressed result, and here is a verbal description of it.  But the current curriculum, as endorsed by the British Psychological Society, does not seek to train logic.  The closest it comes to this is a recommendation that conceptual and historical issues are introduced across the curriculum, either as a stand alone module or embedded within other content.

To be fair to the BPS, I think they are trying to be quite relaxed in their view, and not seek to be prescriptive to the nth degree.  None the less, in my experience I have not encountered a department that sees the introduction of logical reasoning as a component of this provision.  At best students are likely to be exposed to an account of clashing views on epistemology that revolve around tired arguments about positivism based on readings by people such as Bruno Latour.

All this concerns me because science is about formal arguments, where clauses contain data from which you can draw only legitimate conclusions. The literature in behavioural economics and decision making tells us that we have deep biases in our reasoning skills that make us irrational in certain senses of the term.  Much hay is made of this in cognitive psychology classes, and yet few think to expend effort providing the tools to tackle this issue.  Some have written books to help us - such as Human Nature After Darwin and Methods of Theoretical Psychology - but as yet I have seen no evidence of them in use in the classroom.  Note to self - I must change this.

This is not just a pedagogical suggestion.  I think that the evidence for our irrationality also suggests that logical argument cannot simply be learnt by example, it has to be taught or at least modeled.  One reason for believing this is the constant recurrence of technically meaningless debates such as the nature versus nurture discussions that plague undergraduate seminars and review commentaries in esteemed journals.  Nurture is something that happens in nature and to force a debate on their relative contribution to a trait is to make a category error.  The discipline as a whole would benefit from rethinking its educational practices in this arena and extending our view of research methods.

If I do not write more on this topic in the future I will be very surprised.