Chris Blattman

Search
Close this search box.

Me, myself and I

self has essentially no meaning except in distinction to some other, and so the relevant question becomes where to draw the borderline between the two. In social psychology, studies of group identity have suggested that sometimes the notion of self extends beyond that of the individual. Here, we suggest the reverse: in one specific but important way—namely across time—the notion of self is considerably narrower than the individual.

That is Julian Jamison and Jon Wegener in a new behavioral economics paper.

To understand decisions made across time, economists often model individuals and their future selves as separate persons. Just a modeling convenience? Maybe not. Jamison and Wegener argue that brain systems and decision processses work exactly that way.

Worth reading.

4 Responses

  1. Julian, I’m also curious you jokingly mention collecting another 500 people. Is this for your own studies? I didn’t notice any new data in the working paper and a quick online search of your name doesn’t show any either. Did you collect your own data or are you planning to collect data? It’s really hard to fully understand the scientific and statistical limitations of these types of data without collecting and analyzing some yourself.

  2. Julian, your study seems to be making a very basic logic error that is common in this type of neuroscience theory work. Brain regions X are involved in two seemingly separate tasks of Y and Z. Therefore, Y and Z are actually the same thing. Each brain region has multiple roles and even combinations of brain regions have multiple roles. Talking about a region at THE theory of mind region and saying anything that also happens there must relate to theory of mind is simply false.
    For the most basic example, primary auditory cortex will have an fMRI signal increase if a play a series of random beeps or a complex piece of music. This doesn’t mean that primary auditory cortex interprets music the same way as random beeps. It merely means you need to do more advanced studies to tease apart how each type of stimulus activates each brain region separatesly.

    Also on the issue of N=16, both Julian and John are wrong. Few neuroscientists will look at a single N=16 and pronounce an area of inquiry closed. The study will be replicated with the same or different parameters and the findings that survive repeated examination become established knowledge. (For example the many ToM papers that Julian references that point to commonalities across studies) This isn’t to say that a single N=16 isn’t significant or newworthy, but its just one step in the process.
    Also, for healthy volunteers, N=16 is really on the border of a sample size needed to produce a cognitive neuroscience result that would be deemed reliable. Sample sizes in the 20-25 range have been shown to give reliable results for some basic paradigms.

  3. John — this confuses a lot of people and is a critique often heard by neuroscientists; however it’s like saying: “You did the evaluation in 16 different countries, and found similar results every time, but that really only counts as 16 observations.” Just remember that we have hundreds of trials for each individual, and thousands of datapoints per trial. We only considered results significant if the corrected p-value was below 0.05, and there is actually rather a lot of statistics involved. Of course, if you want to give me a grant to run another 500 people through the magnet, I’ll be happy to accept!

  4. Seeing “n=16” in papers with such broad policy implications will never not bother me. Neuroscience, this is statistics; you two should hang out more.

Why We Fight - Book Cover
Subscribe to Blog