4 Comments
Jan 13Liked by Sarah Constantin

My first issue with this sort of thing is in the measurement and philosophy. Like your elevator pitch starts with "You know how people get freaked out sometimes and can’t think straight?". And sure, I sort of know that, though I'm unsure whether it's One Thing. Like first of all, there's a distinction to be made between the interpersonal dynamic (in situation S, upsetting thing T comes up and makes person P unable to think straight), specific relationship to that dynamic (person P has a uniquely strong tendency to be unable to think straight when thing T comes up), the general factor underlying that relationship (person P has a uniquely strong tendency to be unable to think straight when various upsetting things come up), and various specific factors underlying that relationship (person P has a uniquely strong tendency to be unable to think straight when their political outgroup comes up).

It's not even clear to me which of these things you're trying to measure. If we do pick one of them, as you say it's unclear whether the fMRI method measures it well enough. But note that this uncertainty primarily exists because we don't know which thing we're trying to measure, as otherwise one *could* check whether the fMRI method gives similar results to what you'd get in the thing you're looking at.

This is one of the reasons I obsess about measurement so much. Measurement forms the foundation for all the other research that people do, one the one hand it makes it clear why one can just dismiss vast swathes of research out of hand, but on the other hand it also unveils tons of productive research questions. Admittedly those research questions start from a more sociological point of view (what do we mean by resilience and for which purposes do we care about it), but that also yield the benefit of making it easier for people to engage in, as it doesn't require expensive equipment, just some attention to the uses of the research.

Expand full comment
Jan 13Liked by Sarah Constantin

Ooops, I skimmed from the elevator pitch to the next section and missed the part where you basically talked about the definitional issues too. (As in "But I don’t even have a really good definition, let alone a good measurement, of “doing the wise-sane thing instead of the dumb-freakout thing.”") I guess you already talked about the definitional issues.

I think I have the philosophy of how to solve these sorts of definitional issues very well worked out, so if you want to create a good definition of it, I can help you do that. It requires a bunch of empirical work though.

Expand full comment
author

Oh I’d love to hear more about your approach! I’m at srconstantin@gmail.com if you’d like to take this to email or a call.

Expand full comment

I'll write a general response here, but if you're interested in using it for researching this question I'd also be interested in assisting via calls and emails.

So according to my philosophy, the reason people want to study these sorts of issues is generally that they observe places in life where they want to apply it. This might for instance be because they themselves have episodes where they struggle with emotional resilience, or because they want to support friends that struggle with emotional resilience, or because they see Their Outgroup and decides that this outgroup is bad because they struggle with emotional resilience.

And the first problem is, if one immediately abstracts things into "emotional resilience", then one is going to lose any details that are relevant to the application, but it is quite possible those details were critically important, and that one ends up studying something else if one abstracts away those details. Pretty much the only way to avoid this is by explicitly thinking about what the application is, all the time, while doing the research. For instance one could start by writing it down explicitly, and one should make sure to collect a bunch of potentially-relevant details about it to work it out properly.

Of course quite likely it's not Exactly One Situation one is referring to, rather there are multiple different ones that one is hoping are related in some way. But there are, as I mentioned in my previous comment, quite a lot of different ways that things can be related. The minimal form of relatedness that I think your post implicitly requires is analogous mechanism (emotions in some situation makes people act irrationally) with analogous consequences (the irrational behavior is counterproductive in whatever they are doing). There are other forms of relatedness that might also be possible, namely shared causes (e.g. the thing that causes the emotions is a property of the situation and/or of the person, and therefore also causes the emotions in different situations) or shared effects (e.g. the irrational behavior contributes to social isolation). Whatever the case, by explicitly listing the different applications, there is clarity to gain by thinking about where exactly the analogies are, like creating a shared schema for them, or noticing that there are important disanalogies that makes it worth paying attention to the differences.

The goal is basically to know the causes of this, and for this, an important indicator is the ability to predict it. It is easier to know how well one can predict something if one has a detailed feedback loop about it, where one regularly gets new information about the phenomenon of interest. If one wants to set up toy models of the phenomenon (which might be helpful for brain studies as then maybe it becomes easier to induce within the machines), one needs to be able to predict it well enough to know that the toy models are relevant. Given a sufficiently good model of the phenomenon, the model will most likely contain a direct term for "being dumb because of bad feels", so by reading off that model, one can create a test of that.

It's kind of hard to say more in the abstract, without specific information about how you're researching this. But basically most of the research questions get automatically answered when considering the applications, as long as one knows how to ask the right research questions. But then there's a lot of work in getting enough information about the applications to answer the tougher parts of the research questions.

Expand full comment