The graph going around showing that people overestimate the proportion of people in minority groups? It doesn't mean much. If you look at the original source you will see that this is nothing except the usual pattern that *every* probability elicitation curve shows. today.yougov.com/politics/art... - ThreadSky

djnavarro.net • 93 days ago

The graph going around showing that people overestimate the proportion of people in minority groups? It doesn't mean much. If you look at the original source you will see that this is nothing except the usual pattern that *every* probability elicitation curve shows. https://today.yougov.com/politics/articles/41556-americans-misestimate-small-subgroups-population

Comments

margrouse.bsky.social•19 days ago

Does this hold when asked to estimate smaller proportions?

djnavarro.net•19 days ago

I think so, but there's some weirdness that arises when people have to think about extremely large numbers or extremely small numbers. @david-landy.bsky.social is probably better placed than I am to speak to that, since he's done a lot of excellent work in this area

david-landy.bsky.social•92 days ago

This.

drussellmrichie.bsky.social•93 days ago

Glad to find you/see you posting on here. You were one of the reasons I liked old Twitter.

djnavarro.net•93 days ago

I have mixed feelings, to be honest. I am a little unhappy to find myself posting threads like this one. It’s not really what I want to be doing, but I haven’t seen anyone correcting the misrepresentations spreading like wildfire about this graph: people who should know better are getting it wrong

david-landy.bsky.social•92 days ago

This is exactly the mix of feelings I think we have all felt doing this work! It's not exactly groundbreaking, but it seemed important.

drussellmrichie.bsky.social•93 days ago

I can understand that.

I look forward to seeing your other posts. Like your generative art!

djnavarro.net•93 days ago

I cannot stress enough that this is extremely typical. Almost every study asking people to judge probabilities or base rates produces this pattern. It's been found repeatedly since the 1970s or so. It shows up across a wide variety of topic areas. Please don't read too much into this.

omriroden.bsky.social•91 days ago

Yeah makes sense. I didn’t look at the source and enjoyed the graph for some loose interpretation 🙃. Are you aware of studies looking at people’s ability to predict that you would recommend for their methodology?

djnavarro.net•93 days ago

I am as concerned as anyone about the rise of false beliefs that people hold about minorities. However, this isn't one of them. This is a fundamental feature of how people respond to questions about probability. No more, no less. Scientists should refrain from spreading misinformation.

jwolondon.bsky.social•93 days ago

Useful and insightful reminder. Not sure about 'No more, no less'. I think it tells us to be critical of political positioning relying on "too many of group X". Especially as many promoting it cynically rely on overestimating group X numbers rather than structural critiques:
https://bsky.app/profile/theguardian.com/post/3lbquxljq232w

zabong69.bsky.social•93 days ago

What is the reason? People just being very bad at percentages?

djnavarro.net•93 days ago

Mix of factors. Some part of it is that people don’t interpret the question literally (e.g. “50%” often means “I don’t know and I don’t want to guess”). Some of it is a statistical artifact (if true value is 1%, noise will be push responses up rather than down), some of it is people being cautious

jonathancorbin.bsky.social•92 days ago

Log-odds probability transformation plus Bayesian updating.

In people's defense, I still have to read this paper again and again whenever I see this making the rounds.

https://link.springer.com/article/10.3758/s13423-017-1360-2

tdawry.bsky.social•93 days ago

Every few years, someone polls Americans on how much money they think NASA gets, and how much it should get.

The results tend to be similar, with people wanting to increase NASA's funding... while already thinking it gets ~20x more funding than it actually gets!

oleksandrsharp.bsky.social•93 days ago

How is "this is the usual pattern" defies the fact that people often overestimate the proportion of people in minority groups? If else it proves the initial statement.

tjmahr.com•92 days ago

people are bad at estimating frequencies in general is a more benign interpretation than other possible (political) interpretations.

Comments

Posting Rules

Reply