Through a lab darkly

Cognitive scientists should be explorers of the mind, forging a path through the chaotic world of everyday life before even thinking of retreating to the lab, according to a critical article in the latest edition of the British Journal of Psychology.

Cognitive science often works like this: researchers notice something interesting in the world, they create a lab-based experiment in an attempt to control everything except what they think is the core mental process, they then test the data to see if it predicts real-world performance.

A new approach, proposed by psychologist Alan Kingstone and colleagues, suggests this is fundamentally wrong-headed and we need to completely rethink how we study the human mind to make it relevant to the real world.

The authors suggest that the standard approach relies on a flawed assumption – that mental processes are like off-the-shelf tools that do the same job, but are just assembled by the mind in different ways depending on the situation.

But imagine if this isn’t the case and mental processes are, in fact, much more fluid and adapt to fit the environment and situation. Not only would we have to change our psychological theories, we would have to change how we study the mind itself because the assumption that we can isolate and test the same mental process in different environments justifies the whole tradition of lab-based research.

The authors suggest an alternative they call ‘cognitive ethology’ and it focuses the efforts of cognitive scientists on a different part of the research process.

Let’s just revisit our potted example of what most cognitive scientists do: they notice something in the world, they create a lab-based experiment, they test to see if it predicts real-world performance.

The first part of this process (noticing -> lab-experiment) is often based on subjective judgements and rough descriptions and isn’t validated until the lab-based experiment is tested.

Kingston and his colleagues argue that scientists should be applying the techniques of science to the first stage – measuring and describing behaviour as it happens in the real world – and only then taking to the lab to see what happens when conditions change.

They give an example of this approach in an interesting driving study:

A Nature publication by Land and Lee (1994) provides a good illustration of a research approach that is grounded in the principle of first examining performance as it naturally occurs. These investigators were interested in understanding where people look when they are steering a car around a corner. This simple issue had obvious implications for human attention and action, as well as for matters as diverse as human performance modelling, vehicle engineering, and road design.

To study this issue, Land and Lee monitored eye, head, steering wheel position, and car speed, as drivers navigated a particularly tortuous section of road. Their study revealed the new and important finding that drivers rely on a ‘tangent point’ on the inside of each curve, seeking out this point 1–2 seconds before each bend and returning to it reliably.

Later, other researchers used a lab-based driving simulator study to systematically alter how much of this ‘tangent point’ was available to see what caused abnormal driving.

The authors also make the point that this approach is much better at helping us understand why something happens the way it does, because it ties it to the real world and helps us integrate it with the our knowledge of personal meaning.

It’s an interesting approach and meshes nicely with a recent article on cultural cognitive neuroscience in Nature Reviews Neuroscience. It looked at a number of fascinating studies on cultural influences on mind and brain function and discusses how we can go about understanding the interaction between culture and the brain.

If you want to skip the theoretical parts, Box 1 is worth looking at just for a brief summary of some intriguing cultural differences in the way we think.

The piece was also rather expertly covered by Neuroanthropology who cover the main punchlines and discuss some of the claims.

Link to ‘cognitive ethology’ article.
Link to PubMed entry for ‘cognitive ethology’ article.
Link to ‘cultural neuroscience’ article.
Link to PubMed entry for ‘cultural neuroscience’ article.

6 Comments

  1. Posted September 1, 2008 at 3:36 am | Permalink

    For a book-length treatment of this topic, Ed Hutchins’ Cognition In The Wild (1996) is worth a look: http://www.amazon.com/Cognition-Bradford-Books-Edwin-Hutchins/dp/0262581469
    He concludes right where Kingstone has just ended up: that studying cognitive processes must include the physical and social environment in which they take places, because so much cognition is externalized in the form of tools and customs.

  2. Posted September 1, 2008 at 3:37 am | Permalink

    Er, “place” not “places”.

  3. Posted September 1, 2008 at 9:44 am | Permalink

    I was thinking about this as I cycled to work this morning. The experimental method investigates one factor at a time, but to translate research into application (advice for driving, or mental health, or efficiency or performance or whatever) you need to understand the interactions between all the factors involved — something you’ve deliberately excluded from your experimental investigations. This seems like a bit of sticky issue to me.
    From the other end of things, if you don’t carry out experimental investigations then your ‘applied’ investigations will only report what was true in the specific circumstances that you investigated; you won’t necessarily find out what *could* be if you changed some factors (because you haven’t established causality). So your knowledge becomes limited in generalisability, or perhaps it becomes embodied in the intuition of the practioneer, which is unscientifically supported (not necessarily a bad thing) but also possibly wrong and/or uncommunicable too.
    Vaughan — as a scientist-practioneer, what do you think?

  4. Posted September 1, 2008 at 12:18 pm | Permalink

    One of the difficulties with the ‘cognitive ethology’ approach is that it works much better for measuring psychophysical variables in the real world (as illustrated in their example).
    Many of the things we’d like to measure (e.g. suicidal thoughts) are just bloody hard. The paper goes on to discuss ways of doing this and integrating personal and cognitive explanations but it’s clearly not as straightforward. One way this is being approached though, is ‘experience sampling’ techniques.
    To answer your question Tom, I’m a big fan of ad-hocism (‘converging evidence’ if you want to sound a bit academic) in combination with the blunt but effective tool of randomised controlled trials.

  5. tom
    Posted September 1, 2008 at 3:02 pm | Permalink

    Sounds fair enough to me. We can all be epistemological anarchists together ;-)

  6. Posted June 23, 2010 at 6:09 pm | Permalink

    everything is relative


Post a Comment

Required fields are marked *
*
*

Follow

Get every new post delivered to your Inbox.

Join 22,492 other followers