Something stood out to me in a number of academic presentations I attended last week: the assumptions and lack of nuance when using measures of cognitive function.

A shift-change has occurred in researchers’ usage of IQ or cognitive function measures. Countless papers have shown links between low socioeconomic status and low IQ scores, or low scores on other cognitive ability measures at various point along the lifecourse. A recent paper in Science was widely quoted in the media (Mani et al 2013), the shorthand for their findings being forged into media-pleasing sound bites that make me cringe: “How poverty makes people stupid “ an example on one page. This does not reflect what the authors found.

The problem is that academics are now using shorthand terms like ‘intelligence’ & ‘stupidity’ when discussing pathways between socioeconomic conditions, cognitive ability, education and outcomes in later adulthood. They are skipping over underlying issues: these variables are heavily intertwined and difficult to disaggregate. Even if a researcher is able to line-up the variables in a time-sequence, a variable collected at one time point does not necessarily reflect a construct at that time point alone.

So how do we discuss these potentially controversial findings?

Symbolic variables

In her keynote address at a conference last week, Melanie Bartley referred to many of these variables we use as symbols. She gave the example of birth weight, which has been used to symbolise the complexities of the early life environment: inherited qualities, social conditions, biological exposures etc.

Quantitative research is teaming with these symbolic variables. Sometimes we can attempt to pluck from a symbol one of its key characteristics – but most often we cannot. What we must not do though, is use symbolic variables to represent something they are not – just because it makes our life easier, and results in a controversial sound bite picked up by the media.

Cognitive ability is neither intelligence nor stupidity. It is a symbol of complex processes and adaptations. Like so many symbolic variables, it is an outcome, a process and a determinant all in one.

We cannot do much about our work becoming distorted in the media, but let’s not do it for them. We can remain clear and nuanced about our work, so let’s avoid the sound bites.

By Michelle


Anandi Mani, Sendhil Mullainathan, Eldar Shafir, and Jiaying Zhao. Poverty Impedes Cognitive Function. Science 30 August 2013: 341 (6149), 976-980, DOI:10.1126/science.1238041