One additional problem w/ “proximate” explanations—like explanations in terms of our beliefs and preferences:

It leads you to splice up the world in rather artificial ways. (Which can make it harder to spot patterns and develop useful interventions).

See below 🧵 for examples.

https://twitter.com/Moshe_Hoffman/status/1526893031416377350

(Here’s another thread laying out additional problems w/ proximate explanations.)

https://twitter.com/Moshe_Hoffman/status/1167288599735361536

Hi y’all.

I wanted to do a 🧵 summarizing our book that came out last month.

(Thoughts, questions, more than welcome.)

smile.amazon.com/Hidden-Games-S… The basic idea is that game theory is super powerful.

Consistently yielding counter-intuitive insights, w/ minimal assumptions.

And has the bonus of being very precise; you know exactly what assumptions are needed, and what implications come out. (The benefit of mathematics.)

New paper (👇).

We present+test a functional explanation for why we are “ineffective altruists”.

Allow me to summarize the paper. And use it to highlight what we think is “the right way” to test ‘functional’ explanations & rule out purely ‘proximate’ accounts.

1/

https://twitter.com/naturehumbehav/status/1315699744514158592

By ‘proximate’ we just mean an explanation in terms of our psychology. An explanation of what we do in terms of what we think, feel, or want.

2/

Really liked recent mini-series “why we hate”.

👍 summary of psych lit+ historical examples.

But...🔑 thing I thought missing, as usual: No discussion of structural problems/underlying incentives.

Let me explain what I mean+why I think that’s 🔑.



(🧵) As the docu-series nicely summarizes:

Milgram showed people do bad when authoritative figure tells them. Informative. Important.

But, umm, what determines who authorities want us to harm?

2/

Historical discourse, & discourse surrounding current events, often conflate ‘justifications’ w/ ‘causes’.

Let me give some examples illustrating the conflation, why common methods for de-conflating don’t work, some methods that might, and why it matters.

(Thread) Of course, sometimes stated reasons are actually the the true reason.

If you asked me why I was carrying an umbrella, I would probably tell you the truth: cause I thought it would rain.

2/

Recent editorial 👇 (by Kahneman & Renshon) offers a readable summary of many interesting ‘judgment and decision making’ findings pertinent to trigger-happy leaders.

But imo ALSO exemplifies a key limitations of jdm-style explanations.

(Thread.)

foreignpolicy.com/2009/10/13/why… So the article talks about -overconfidence
-reactive devaluation
-sunk cost fallacy
-the fundamental attribution error

etc.

(I’ll explain.)

2/

A common debate that shows up in historical analysis 👇.

My 2¢:

Incentives shape ideas far more often than we think. And hence the independent causal role of ideas is vastly over-estimated.

Here’s 3 reasons why I think that.

(Thread)

https://twitter.com/scholars_stage/status/1215360428168994817

1) incentives don’t just shape our conscious strategic behaviors, but also which ideas we generate, spread, and come to believe.

2/

FWIW:

A 🧵 of 🧵’s, on my 🔑 criticisms of social-science research.

👇 My 🔑 criticism of evolutionary psychology:

(Focused on evolutionary past, overlooking adaptations to *contemporary* context, enabled by learning processes.)

https://twitter.com/Moshe_Hoffman/status/1073270050809937923?s=20

Podcast 👇 provides 👍 summary of happiness lit.

Imo: a lot of nice anecdotes & empirical results.

But the lit has many problems.

Below, I’ll summarize a 🔑 conceptual-flaw, a bundle of methodological-flaws, and a few perverse-incentives at play.

(🧵)

podcasts.apple.com/us/podcast/the… The 🔑 conceptual problem:

The lit nicely documents 😀 returns to baseline after some time. Almost regardless of improvements/decrements in material circumstances.

(And baseline seems to be largely genetically determined.)

A common problem w/ experimental “tests” of social science models:

Literally reproducing the model in the lab. As opposed to testing the key assumptions and implications of the model.

Let me explain what I mean. And why this is problematic.

(Thread) I am, in general, a big fan of using lab experiments to test theory.

But careful thought needs to be given to the question: what exactly, about the model, needs to be tested.

And what’s the ideal experiment to test that.

And what does this experiment actually test.

The social-sciences has a real problem w/ causal inference.

(Thread) I am not referring to problems w/ the way we write down regressions.

Or confounds in our experimental designs.

But another endemic problem: conflating “proximate” causes w/ “ultimate” causes.

Allow me to expound. W/ examples. And describe how this holds science back.

Two claims I often rail against:

1) Our beliefs are driven by evidence.

2) Our good deeds are driven by caring.

Both false.

—-

Let me explain why I believe this. And why it’s fundamental for the social sciences.

(Thread.) By understanding why these claims are false:

-we can gain a deeper understanding of how preferences & beliefs *actually* work

-we can see the problems w/ trusting our intuitions on these topics

-we can see what social science needs to do differently to make *real* progress

Thread:

What causes disasters, like Chernobyl, Deepwater, Challenger, & Titanic?

Claim:
-Perverse incentives <— cause
-Corprorate-Culture, deception, ... <— mechanism, not cause

Lessons:
-How beliefs, culture work
-How to think bout causality in such contexts

(Details ⬇) What perverse incentives?

- Rewarded for haste
- Rewarded for cutting costs
- Insufficiently penalized for causing harming

➥ Incentivizes risky behaviors

A really beautiful model👇, imo.

Explains so much about the human experience—our thoughts, feelings, & behavior—with so little.

Let me summarize the model, what it explains, and what it teaches us about good modeling in the social sciences.

(Thread)

shorturl.at/mvM57 Some puzzles the model helps explain:

Bonobos, like us, spend an inordinate amount of time playing—hanging out and doing ow seemingly wasteful, but “fun” activities. Some might be learning.

But much is best described as “bonding.”

Best criticism of social-psychology I have seen 👇.

(By @PsychRabble et al)

Allow me to summarize the key points.

And two important takeaways.

(Thread)

dropbox.com/s/elimpj5u0aae… The key point:

Fundamental problem in social psych (in addition to p-hacking):

Researchers, motivated by the “wow factor,” read into their (usually cool & interesting) results, the most shocking interpretation that’s *plausibly* consistent w/ their data.

Why did the Nazis kill off the Jews?

(A thread.) One of the key incentives at play was most definitely the ability to loot. Both $ & jobs. Given to Germans, the Nazi leadership. And their allies. $ and jobs especially valued in time of depression. And war.

(See above image summarizing “Hitler’s Beneficiaries”)

A legitimate criticism imo of many laboratory experiments in behavioral lit.

(Lab experiments often measure immediate effects w/o considering how we will adapt.)

https://twitter.com/page_eco/status/1137004785939361792

One reason might be misleading:

Measuring effect of a proximate mechanism. When other mechanisms are available. And is a deeper driver.

Thread:

Why do people have ridiculous beliefs?

Like climate denial, creationism. Or that DJT is a good president.

Many think such beliefs can be reconciled with a “rational,” “bayesian” account.

I don’t.

Here, I’ll walk through all the evidence against. I’ll list 5 counterarguments

1) accurate beliefs would be quite *counterproductive,* in these domains

2) beliefs coincide *too well* with what we are incentivized to believe

3) disagreement is * known* to be *systematic* & *persistent*

Vestigial traits.

Some of the most straightforward evidence for human evolution.

How can anyone see this & deny human evolution? So weird.

But also, how clever of naturalists (Darwin?) to have spotted these & known to present as irrefutable evidence. So clever.

https://twitter.com/DorsaAmir/status/1085227486869377024

(Note to social scientists:

This evidence, once again, isn’t experimental. And yet it’s more irrefutable than any experimental evidence for evolution I can think of.

How come? What makes it so compelling? When is it better to look for evidence like this vs experiment?)

Hmm interesting question.

My take on cause of/lesson of replicability/p-hacking issues:

Classic motivated reasoning

People have strong incentive to 1) cheat, but to do so in ways that are 2) culturally acceptable and 2) plausibly deniable. AND to 3) avoid self-reflection

No?

https://twitter.com/DinaPomeranz/status/1084881496459132929

(Thus replicability issue doesn’t say *much* about the individual (albeit is clearly a moral hazard component and individual differences to susceptibility therein), but about problems w/ institutional incentives+human nature.)

Hmm, reading the wiki article Tage links to below (in reference to recent work by @datingdecisions).

Which provides a cost benefit analysis of relationships.

Except the costs are characterized there in terms of pleasure and pain. “Psychic utility,” if you will.

1/3

https://twitter.com/tage_rai/status/1083850526922629122

Whereas a functional/evolutionary approach to relationships (what is done in my literature) would do something similar

*But* where costs/benefits are described not in terms of how we feel but *material* considerations

Like help, financial security, mating opportunities...

2/3