4 ms·
This doesn't surprise me. People tend to self-segregate. All else being equal, people will tend to live near other people of similar backgrounds. Now let's say
by mikeash 12y ago
This doesn't surprise me. People tend to self-segregate. All else being equal, people will tend to live near other people of similar backgrounds.
Now let's say you're a purple person (fictional race chosen to avoid invoking any unintended implications) moving into a town that's 50% purple and 50% green. There are lots of places to live near other purple people. Some of those places will probably be places you like and can afford. So you'll naturally join your co-purpleists.
Now let's say you move to a different city that's 5% purple and 95% green. There's probably only a couple of places where you can live among other purples. Odds are pretty good that those places don't suit you. You'll be much more likely to choose to live among the greens.
- unabridged 12y agoHere is an agent based simulation that illustrates this concept: http://nifty.stanford.edu/2014/mccown-schelling-model-segregation/ http://nifty.stanford.edu/2014/mccown-schelling-model-segreg...
- DanAndersen 12y agoOne interesting thing of note about the Schelling segregation model is that you still get segregation/clumping even when the threshold is relatively low: >For the remainder of this explanation, let's assume a threshold t of 30%. This means every agent is fine with being in the minority as long as there are at least 30% of similar agents in adjacent cells. In addition, when the threshold is too high, there's a large amount of churn that happens, due to constantly dissatisfied agents moving around and having "undesirable" agents coming in. I wonder if there has been work on adjusting the Schelling segregation model to give different threshold/"tolerance" values to different types of agents. I also wonder if there have been attempts at applying this model to real-world racial data. Given the several most recent recordings of census data for race, could there be an analysis that concludes, "the arrangement we have is, if assumed to be a Schelling segregation model, most likely to have arisen as a result of a tolerance threshold of X%"?
- lotharbot 12y agoOne thing I found fascinating is that, with a sizable majority (I set red to around 80%), the minority actually becomes less stable if you set a moderately low threshold for moving. With a high threshold (like t=67%) you get very rapid clumping as both reds and blues try to move into areas where they're in the majority. But with a low threshold (like t=25%) the few majority people who are isolated move, minorities quickly fill the open spots, and then the only remaining locations for minorities are completely surrounded by the majority (while no more majority members ever move). So the minorities just shuffle around between a bunch of unacceptable locations, none of which give them the 2 minority neighbors they require. Then there's a sudden transition when t gets low enough, at around 16% -- the minority will tolerate having only one similar neighbor as long as there's also an empty spot, and therefore they stabilize very quickly. So when the threshold is high, there's a lot of churn early but then stabilization. When the threshold is low, the tolerant majority stops churning quickly and therefore stops opening up spots the minority finds desirable.
- mxfh 12y agoanother interactive version of the Schelling model: http://ncase.me/polygons/ http://ncase.me/polygons/
- vacri 12y agoInstead of purples and greens, let's use whites and yellows. We can see in the real world, whether it's Europeans in Asia, or Asians in Europe, the foreigners would be lucky to make 5%, yet tend to live in the same neighbourhoods in whatever city. When you're that level of minority, clumping together gets you better services you can access - in languages you can speak fluently, from people who are familiar with your cultural background and needs.