The “what percentage do you think” vs “what percentage are” would be much better either as a scatterplot with a y=x line, or as a ratio (think/are) vs actual percent

Here’s a rough plot of the second thing. There’s a number of issues with it but I think it more clearly conveys the crucial information

The first plot might arguably be better but you couldn’t really see the way the ratio of overestimation decreased with more immigration as clearly as this one.

edit: Damn that came out way smaller than I expected. It’s readable if you zoom in

Swiss: about a third?

Real: 40%

https://www.bfs.admin.ch/bfs/de/home/statistiken/bevoelkerung/migration-integration/nach-migrationsstatuts.html

how do people in the US think Muslim folk make up 22% of the population!? My guess was like 4-5% and I still overshot by a lot.

These polls tend to fumble their own methods, by trusting people to read definitions. They should be asking directly: how many people in your country were born in another country? The word “immigrant” literally means that… but that’s not the only meaning people envision, when they hear that word. To some extent you are always measuring that disconnect.

On the other hand, what fucking lunatics think 22% of America is Muslim?

I think the issue is that most people live in cities, where populations tend to me more diverse. Then most polls probably also end up disproportionately asking cityfolk. So the polls ask people who live in areas with disproportionate numbers of immigrants (relative to non-urban parts of the country), and they forget how many non-immigrants are outside the cities.

How do people in Japan think that 10% of the population is foreign!?

I guess Argentina makes a bit more sense - except that not many people are trying to get to Argentina. That sounds like Argentina though.

90% or so of people in the USA are immigrants

I dont get these graphs