4 ms·

Something seems off about these graphs. Montana's data looks particularly suspicious. The author claims to have gotten the data from Stanford's open policing da

by swebs 9y ago

Something seems off about these graphs. Montana's data looks particularly suspicious. The author claims to have gotten the data from Stanford's open policing dataset which can be found here: https://openpolicing.stanford.edu/data/ https://openpolicing.stanford.edu/data/ Opening the Montana dataset in Pandas you can see: df.driver_race.value_counts() White 752032 Other 38782 Hispanic 16055 Black 8805 Asian 6699 Native American 2 58 1 60 1 Asian / Pacific Islander 1 25 1 So we can already see that there were 85x more whites being pulled over than blacks, which is the opposite of what their chart was suggesting. That is not very surprising since according to the census, whites made up 87.8% of the Montana population while blacks made up 0.4%. When accounting for population difference, you get numbers similar to what they claimed, though they fail to mention this in the presentation. "Traffic stops for white and black Americans per 100 driving-age people, 2015 (where data is available)" How much of this article can be trusted? It seems to be filled with a lot of appeals to emotion, and misleading charts.

Y_Y 9y ago
"Black motorists are pulled over by police at rates exceeding those for whites" Do you think this isn't true? I don't know one way or another, but I think if National Geographic is going to make such a claim they'd have made some effort to verify it. How is someone who is not able (or maybe not bothered) to assess this claim themselves from raw data, supposed to know what to believe? Is there a good way to compare these claims? (For the record I did check the numbers in the parent post myself)
mikeash 9y ago
There are ~220x more whites than blacks in Montana, yet only 85x more whites are pulled over. How is it your conclusion that whites are 4x more likely to be stopped when accounting for population? Just from those numbers it looks like blacks are 2-3x more likely to be stopped.
- swebs 9y ago
  You're right, I multiplied by .04 instead of .004. I'll edit that out.
  - mikeash 9y ago
    You still have the part where you say the numbers aren’t surprising and you don’t trust the claims, which also doesn’t seem to add up.
- calvinbhai 9y ago
  I wonder if the %age of infractions (across a range of severities) tend to be the same across all races. Because if it is, the white people should be pulled over at least 3x times more. Another thing I dot understand: when a cop pulls over a vehicle, how is it possible for the cop to know if the car is being driven by a black person or not? so, it is possible that of all those pulled over, a greater %age of black drivers get ticketed for smaller infractions.
  - mikeash 9y ago
    Maybe I’m misunderstanding your last paragraph, but it’s pretty easy to observe the race of the driver from the outside in most cases.
kthejoker2 9y ago
If blacks are 4 out of 1000 Montanans and whites are 878 out of 1000, then if there are 8,000 stops of blacks there should there should be 1,756,000 stops of whites.
- maxerickson 9y ago
  Yeah, there's ~2 stops of a black motorist per black person living in Montana. Kind of drives home the point.
mjg59 9y ago
> which is the opposite of what their chart was suggesting No it's not. The chart says that a black driver has approximately a 3x greater chance of being stopped than a white driver, which seems consistent with your numbers.
Feniks 9y ago
"It seems to be filled with a lot of appeals to emotion, and extremely misleading" You heard it minorities of America. Go back to bed, everything is fine. Nothing to see here.
tzs 9y ago
> That is not very surprising since according to the census, whites made up 87.8% of the Montana population while blacks made up 0.4% I'm not sure that going by overall population is useful. There are at least two problems with it, when looking at being pulled over while driving. 1. Age distributions need to be taken into account. I don't know if this applies to Montana, but I believe that in some states there are noticeable differences in age distribution between races. At the very least, you want to limit it to just people who are old enough drive. 2. Wealth distribution needs to be taken into account. In most states blacks are on average less wealthy than whites. I would expect that means that white families are more likely than black families to have more than one car. I would expect that to result in disproportionately more white driven car trips than black driven car trips. You really need to go by what police are seeing on the roads: what percent of the cars that a police officer sees are white driven and what percent are black driven? That's loosely correlated with general population race distribution, but there are several filters including those described above between that and cars on the road that change the distribution.

by swebs 9y ago

Y_Y 9y ago

"Black motorists are pulled over by police at rates exceeding those for whites" Do you think this isn't true? I don't know one way or another, but I think if National Geographic is going to make such a claim they'd have made some effort to verify it. How is someone who is not able (or maybe not bothered) to assess this claim themselves from raw data, supposed to know what to believe? Is there a good way to compare these claims? (For the record I did check the numbers in the parent post myself)

mikeash 9y ago

There are ~220x more whites than blacks in Montana, yet only 85x more whites are pulled over. How is it your conclusion that whites are 4x more likely to be stopped when accounting for population? Just from those numbers it looks like blacks are 2-3x more likely to be stopped.

swebs 9y ago
You're right, I multiplied by .04 instead of .004. I'll edit that out.
- mikeash 9y ago
  You still have the part where you say the numbers aren’t surprising and you don’t trust the claims, which also doesn’t seem to add up.
calvinbhai 9y ago
I wonder if the %age of infractions (across a range of severities) tend to be the same across all races. Because if it is, the white people should be pulled over at least 3x times more. Another thing I dot understand: when a cop pulls over a vehicle, how is it possible for the cop to know if the car is being driven by a black person or not? so, it is possible that of all those pulled over, a greater %age of black drivers get ticketed for smaller infractions.
- mikeash 9y ago
  Maybe I’m misunderstanding your last paragraph, but it’s pretty easy to observe the race of the driver from the outside in most cases.

kthejoker2 9y ago

If blacks are 4 out of 1000 Montanans and whites are 878 out of 1000, then if there are 8,000 stops of blacks there should there should be 1,756,000 stops of whites.

maxerickson 9y ago
Yeah, there's ~2 stops of a black motorist per black person living in Montana. Kind of drives home the point.

mjg59 9y ago

> which is the opposite of what their chart was suggesting No it's not. The chart says that a black driver has approximately a 3x greater chance of being stopped than a white driver, which seems consistent with your numbers.

Feniks 9y ago

"It seems to be filled with a lot of appeals to emotion, and extremely misleading" You heard it minorities of America. Go back to bed, everything is fine. Nothing to see here.

tzs 9y ago

> That is not very surprising since according to the census, whites made up 87.8% of the Montana population while blacks made up 0.4% I'm not sure that going by overall population is useful. There are at least two problems with it, when looking at being pulled over while driving. 1. Age distributions need to be taken into account. I don't know if this applies to Montana, but I believe that in some states there are noticeable differences in age distribution between races. At the very least, you want to limit it to just people who are old enough drive. 2. Wealth distribution needs to be taken into account. In most states blacks are on average less wealthy than whites. I would expect that means that white families are more likely than black families to have more than one car. I would expect that to result in disproportionately more white driven car trips than black driven car trips. You really need to go by what police are seeing on the roads: what percent of the cars that a police officer sees are white driven and what percent are black driven? That's loosely correlated with general population race distribution, but there are several filters including those described above between that and cars on the road that change the distribution.