Welcome to the Utopia Forums! Register a new account
The current time is Thu Apr 15 22:52:35 2021

Utopia Talk / Politics / Sebs are bad at statistics
Rugian
Member
Tue Feb 23 11:30:54
A recent nationally representative survey commissioned by Skeptic Mag asked respondents to estimate the number of unarmed blacks killed by police in 2019. Overall, 44% of liberals guessed 1,000 or more as compared to 20% of conservatives (this calculation is based on the cross-tabs shared with me by the researcher [see link])

According to the Mapping Police Violence database, the actual figure is 27.

Furthermore, the average liberal respondent also thought that a clear majority of people killed by police in 2019 were black (in actuality, roughly a quarter were)

Link to full report https://t.co/O8CEkG3raZ?amp=1

Worth adding that this pattern of results is likely more general than specific to estimates of police homicides. For instance, in the study below, liberals significantly overestimated the number of black job resumes per callback

http://mob...932/status/1364024711592738817
Rugian
Member
Tue Feb 23 11:33:39
What a surprise, leftists have a skewed perception of reality that's warped by the media they consume. In other news, the sun will rise in the east tomorrow.
Dukhat
Member
Tue Feb 23 11:51:43
Normal person: is this question worth even asking in the first place?

No? Move on.

Cuckservative: lets keep clicking these social media links and find more “proof” of white genocide!!!111
Dukhat
Member
Tue Feb 23 11:52:07
Also your dumbass dad is still dead and you’re all alone.
Rugian
Member
Tue Feb 23 11:58:10
"is this question worth even asking in the first place?"

It very much is, since a lot of the support for BLM stems from a misguided idea that there is an "epidemic" of police violence against blacks in the US.

If you read even one leftist paper in the last year or so you'd know that.

Say hello to your wife for me, I know she misses having a real man that makes her feel something (anything).
Rugian
Member
Tue Feb 23 12:00:36
I invoked Seb's name specifically for a reason - he has previously said that the "epidemic" of police violence against blacks is such a settled science that it's not even worth debating.

In that regard, he's about as ignorant as you. In his defense though, his IQ is still higher than yours.
Nimatzo
iChihuaha
Tue Feb 23 12:31:08
Rugian
Well, the Rugians are also bad at facts, just on other subjects.

"his IQ is still higher than yours."

In Dukhat's defense, his recent behavior, is tacit admission that he knows the gig is up. A decade of lies about himself and his success, unraveled, with the passing of your father. I guess we can add The "Meltening" of Dukhat to your dad's CV?
Paramount
Member
Tue Feb 23 12:31:37
” the average liberal respondent also thought that a clear majority of people killed by police in 2019 were black (in actuality, roughly a quarter were)”

Maybe blacks are over represented as victims.



”A recent nationally representative survey commissioned by Skeptic Mag”

Skeptic Mag? Rofl what the fuck is that. Is that like The Onion?
Paramount
Member
Tue Feb 23 12:39:49
Does it say how the survey was conducted?
werewolf dictator
Member
Tue Feb 23 12:43:54
The Skeptics Society is a nonprofit,[2][3] member-supported organization devoted to promoting scientific skepticism and resisting the spread of pseudoscience, superstition, and irrational beliefs. The Skeptics Society was founded by Michael Shermer.. The stated mission of Skeptics Society and Skeptic magazine "is the investigation of science and pseudoscience controversies, and the promotion of critical thinking."[4]

Editorial board
..
Richard Dawkins
Jared Diamond
..

_____

maybe not my favorite type of people but hardly the onion.. and better people than 1984 and 1994 crime bill champion joe biden.. or prosecutor kamala harris.. who are the types of people who actually ruin lives of blacks in usa
Rugian
Member
Tue Feb 23 12:48:30
Nim

It wounds me that we must dislike each other these days. Can you please return to the light and reaffirm your belief in the superiority of Western liberal values already?

Paramount

Of course they're overrepresented. He issue is a) whether racism is the cause of that overrepresentation and b) how significant the problem actually is.

As it turns out, its not a huge problem at all. Certainly one worth addressing, but not worthy of the media attention it received in the last year.
Paramount
Member
Tue Feb 23 12:48:45
Michael Brant Shermer (born September 8, 1954) is an American science writer, historian of science, founder of The Skeptics Society, and editor-in-chief of its magazine Skeptic,

Shermer was once a fundamentalist Christian

Shermer was raised with guns.

he voted for John Kerry in 2004

http://en.wikipedia.org/wiki/Michael_Shermer


Well... I don’t know. I’m a bit skeptic.
Rugian
Member
Tue Feb 23 13:16:45
Would you have preferred that he voted for GWB? Never would have thought that would come from you, my Swedish friend.
Habebe
Member
Tue Feb 23 13:22:00
Dukhat,

"is this question worth even asking in the first place?"

Considering there are a lot of public policy proposals being put out there based off of this misconception , then why should it not be asked?

Paramount, Over represented in what manner?

http://youtu.be/Cv00hGUYYbY

Interesting watch.
BigDickNegro
Member
Tue Feb 23 14:30:52
"unarmed blacks killed by police"

You know how it works: put a gun on the dead body.
habebe
Member
Tue Feb 23 15:14:58
^Don't forget to sprinkle crack on them.
Nimatzo
iChihuaha
Tue Feb 23 19:17:48
"Rugian
Member Tue Feb 23 12:48:30
Nim

It wounds me that we must dislike each other these days."

tsk bitch... Who said anything about dislike you? Just because you happen to be in a cult now, doesn't mean I won't be your friend and we can't play.

You didn't kill anyone right? No raping or molesting? Words, right? Ok then.
obaminated
Member
Tue Feb 23 20:06:43
"The meltening of dukhat" is one of the best phrases this forum has come up with in ages. Good job nim.

And we can all be friends so long as none of you threaten to eat me for some bizarre and clearly mentally damaged reasons.
Seb
Member
Wed Feb 24 01:05:18
Rugian:

Firstly, epidemic is your language not mine.

Secondly, you are down filtering to situations where the outcome is death.

Thirdly, you are filtering to situations where the individual is reported "unarmed", and using that as the only situation where violence is unwarranted. There are numerous examples where police have lied about that.

Finally, you are using absolute numbers, not normalised.

And you say I don't understand statistics.
Nimatzo
iChihuaha
Wed Feb 24 03:35:18
Seb. He is speaking in memes. It isn't really "statistics".

Obaminated, thank you :) my greatest works since I came up with "Obaminated" ;)
Seb
Member
Wed Feb 24 03:48:21
Nim:

I've toyed with looking at the number of convictions from the insurrection that are police and using Samstistics to see if we can "prove" that US law enforcement officers are more criminal than African Americans that need to be deported.
Seb
Member
Wed Feb 24 03:48:57
"Deport the police" etc.
Nimatzo
iChihuaha
Wed Feb 24 04:04:53
Not totally sure what you meant, but I think his point is that liberals model of reality is off while conservatives nail it, in general... on this specific issue. I have posted across the political spectrum versions of this kind of data. People have biases and those biases correlate with political conviction.

I don't get why the issue, human biases, should be talked about, as something divisive.
Pillz
Member
Wed Feb 24 05:51:44
Step 1 - Seb says criteria are too narrow, sources can't be trusted. Broadens it to police interaction in general and declares it systemic because of the scope and his assertion that police lie

Step 2 - Seb attempts to switch goal posts, proposing a ridiculous counter argument to something that hasn't even been brought up in this thread

It's like you are entirely disconnected from what is going on at all times.
werewolf dictator
Member
Wed Feb 24 06:24:18
just because police supposedly killed someone doesn't really mean they actually killed someone.. for example george floyd clearly died of cardiopulmonary arrest from fentanyl

and just because someone is unarmed doesn't mean police killing is unjustified [at least according to seb and liberals].. her name was ashli babbitt and she was shot while unarmed point blank execution style to the neck by a cop and seb and liberals saw no real problem
Habebe
Member
Wed Feb 24 06:24:28
"I think his point is that liberals model of reality is off while conservatives nail it, in general... on this specific issue. "

That's what I got from it.Im sure we could find a roughly flipped situation, but on this particular situation Conservatives seemed to be far closer to reality than liberals, thats it.
Nimatzo
iChihuaha
Wed Feb 24 06:38:40
And I would add, getting it "right" like this, says nothing about how sound the process of arriving there is, as in how well it scales. Most often it doesn't scale. Is there an open season on black people? No. And perhaps something conservatives are inclined to search the answer for or find the answer that suits their bias, because otherwise it poses a challenge to their core beliefs of how great their country is. It misses the bigger issue though: does the police fuck more with (innocent) black people? The answer to that isn't a resounding no, explained with one graph, it is more complicated.
Seb
Member
Wed Feb 24 08:02:39
Pillz:

No, Rugian attributed to be a particularly narrow view which I don't subscribe to.

The fact my view is broader is an entirely reasonable response. If rugian didn't *mean* to encompass a broader view, he shouldn't have invoked me in the first place.

As for police etc. abuse of statistics is clearly the subject of the thread. It says so in the title.
Nimatzo
iChihuaha
Wed Feb 24 09:26:03
He also clearly referenced a plurality of "Sebs". I'm just saying if we are going to be anal about it, we should wait for the rest the of sebs. Presumably they will agree with you, since they are clones (?) of you.

There is no end to the silliness!
Seb
Member
Wed Feb 24 09:45:25
Nim:

"Not totally sure what you meant,"

Nothing really, just mocking the general weaponization of poor stats to prove bogus points.

I didn't really engage with whatever Rugian posted beyond the obvious conceptual errors in that this does not necessarily mean what he appears to think it means, at least in relation to my views.

It doesn't even show that in the narrow circumstances - death of an unarmed man at the hands of the police - that this is not disproportionate.

I would say that baseline competency in quantitative estimates by the general population means that the question as posed doesn't even tell you whether the public thinks it is more proportionate or less proportionate.

Even the most conservative, over half thought it was at least an order of magnitude higher than it was, at which point it would be larger than the total number of armed people killed by the policy each year of any race.

So does the difference come because the Liberals think the police kill a greater proportion of black people than white people, or because liberals think police kill more people than conservatives do?

Similarly the percentage question is off too because people will often confuse percentage with liklihood. I.e. if you ask someone what percentage of people killed by a policeman are black, and think a policeman is 3 times more likely to kill a black man as a white man, they will tend to not account for the fact there are far fewer black men and give you an answer like 60%.

Such questions can give an indication of bias, but comparing quantitative estimates from the general public to baseline stats/reality isn't terribly helpful because people are so terrible at doing this even on issues where they are not biased.

This is not surprising - it is well known that asking the general population to estimate stats and probability produce hilarious results for those with some training.

But this puts the cart before the horse - you yourself have published articles on peoples biases and how they tend to be correct in direction but quantitatively off.

All in all, it doesn't really prove anything useful - it's quantitative data that lacks a qualitative context and understanding of what the number represent - and this is a common problem with Rugian and Sam and others who preach "evidence base" for their biases, but are incapable of engaging with the evidence in an objective fashion because of their biases.

The question really is which perception of reality is more right: do police tend to disproportionately target black people and disproportionately more likely to use violence in their engagement or not, and if so, why and what can be done about it?
habebe
Member
Wed Feb 24 09:46:35
Nim, You had me at anal.
habebe
Member
Wed Feb 24 09:49:08
Rev "I would say that baseline competency in quantitative estimates by the general population means that the question as posed doesn't even tell you whether the public thinks it is more proportionate or less proportionate."

That seems to imply that the public doesn't realize that blacks are about 13% of US population.
habebe
Member
Wed Feb 24 09:50:50
Rev was supposed to be Seb.
Seb
Member
Wed Feb 24 10:11:37
Habebe:

http://new...lack-hispanic-populations.aspx

Slightly less than 1 in 10 get that stat correct.

"On average, Americans say that 33% of the U.S. population is black. In fact, a majority of Americans (56%) estimate that the percentage of blacks in this country stands at 30% or higher. As many as 17% of Americans say the percentage of blacks is 50% or greater. Only 7% accurately state that the percentage of blacks falls between 10% and 14% of the entire population."


And I would bet good money, counterintuitively, that Conservatives would consistently over-estimate compared to the rest of the populations.

People are astonishingly bad at quantitative reasoning, particularly accurately guessing proportions and probabilities.

And even if someone knows that the black population is 13% Black, they would still likely not factor that into a survey response in number of black people killed by police.
habebe
Member
Wed Feb 24 10:34:17
Well, for starters you poll is 20 years old.

However, if that were to hold the numbers in the OP might be better understood.

"People are astonishingly bad at quantitative reasoning, particularly accurately guessing proportions and probabilities."

As far as the general population goes, I agree 112%!

Actually there are several Freakonomics podcasts that display just that.

"And even if someone knows that the black population is 13% Black, they would still likely not factor that into a survey response in number of black people killed by police."

Again, your likely right , I be more specific anyway its black males. I am far more likely to be shot by police than a black female with my Golden blonde hair and blue eyes.

However, this really just goes to show that people in general have no clue what's really going on, even worse that the people pushing for legislation dont either.

Which is why I think expanding the vote usually is a bad idea.
Nimatzo
iChihuaha
Wed Feb 24 10:41:27
"The question really is which perception of reality is more right: do police tend to disproportionately target black people and disproportionately more likely to use violence in their engagement or not, and if so, why and what can be done about it?"

That is fair, but that isn't how the question was posed when people were energized about racism last summer, including yourself. People ran into that issue with emotions (as did I, just very doubtful as to how far this can be extrapolated and what it actually says about American society. I remember that anyone that tried to provide a more nuanced picture with facts and logic, was at best ignored.

You can provide reasoned and factual argument to an emotional outpouring, but it will rarely work as intended and often blow up in your face. The people energized into action by emotions, will look down on your "facts", as you clearly do not "get it", and thus are "part of the problem". Next thing you know you are fired from your job.

Whatever problems there are here, we seem to all be part of it to varying degrees.
Seb
Member
Wed Feb 24 10:44:35
habebe:

Has there been a big public education campaign since 2001 that would make people more informed about the demographics of the country?

If you ran the poll today, I think you would get the same results!

Seb
Member
Wed Feb 24 10:48:03
Nim:

"but that isn't how the question was posed when people were energized about racism last summer,"

Because this question has been asked and answered again and again and again - there is a reason that these single events blow up.

It's not like these were the first instances and this was a hitherto undiscovered issue.
Seb
Member
Wed Feb 24 10:59:48
Bottom line, when you have police caught on camera multiple times blatantly behaving in unacceptable ways, and getting off Scott free, saying "hmm, this is a terribly bad thing and it's frustrating he didn't face consequences, but statistically, doess this phenomenon really exist?" is a terribly bad response.

It's like a govt responding to 9/11 by saying "yes, but statistically terror attacks are terribly rare and not particularly proportionate to go to war over it".

Arguably, yes, rationally speaking, the US response was terrible prioritisation of resources and priorities.

But there is a difference between arguing how prevalent a phenomenon is, whether it exists at all, and then how much priority should be given to it.

That there is racism and impunity in the US police forces is obvious. The degree is kinda irrelevant, and the priority it gets is not really amenable to statistical analysis.

habebe
Member
Wed Feb 24 11:22:15
Of course your mind only goes to a "public education campaign"

You may be right, the numbers may be about the same.But I'm not so certain, with the highlighted emphasis on race and such in the news and online it could change.

However, its rarely a good idea to compare stats from 20 years apart like that.

Nimatzo
iChihuaha
Wed Feb 24 12:05:13
"Because this question has been asked and answered again and again and again - there is a reason that these single events blow up."

It hasn't been answered or throughly dealt with, otherwise, what is the problem? I said as much back then and I actually didn't presuppose which way the "dealing with" would go, I am assuming I will be as shocked as you. At any rate, it seems wise to keep asking the question "where are we", since the answer is expected to change across time and place and not be a constant. Otherwise, everything you and I or anyone else says or does to combat racism is utterly futile.

Anyway what you just responded with makes no sense at all, in light of your first formulation of the question, which didn't presupose the answer, that you now say, has been asked and answered many times. Either there is something to figure out, or this has been done to death already. Which is it?

Clearly you and I (and a bunch of people) disagree about the state of the world on some very important issues, issues with empirical answers. Some kind of third party should look into it and provide facts for shared sensemaking. We either reconcile with reality or there will be a reckoning.
werewolf dictator
Member
Wed Feb 24 17:31:28
60.40% <--- "very liberal" average guess on what percentage of people killed by police were black

51.16% <--- "liberal" guess

45.88% <--- moderate guess

37.80% <--- conservative guess

44.50% <--- very conservative guess

23.86% <--- actual black percentage of those killed by police in wapo database

_____

38.6% <--- homicide offenders that were black males in 2017 [fbi]

23.1% <--- black male percentage of those killed by police in wapo database

4.6% <--- homicide offenders that were black females in 2017 [fbi]

0.8% <--- black female percentage of those killed by police in wapo database

werewolf dictator
Member
Wed Feb 24 17:35:59
"And I would bet good money, counterintuitively, that Conservatives would consistently over-estimate compared to the rest of the populations."

if you read your own link you don't need to guess

avg [mean] guessing % black
democrats.. 35%
republicans.. 32%
liberals.. 32%
conservatives.. 33%

so conservatives overestimated by 1% compared to liberals [which will be meaningless after margin of error]
werewolf dictator
Member
Wed Feb 24 17:56:55
people should have to guess percentage of top political donors who are jewish

http://www...top-15-jewish-political-donors

i'd say anyone very far off is clueless on american politics and its true oligarch masters
werewolf dictator
Member
Wed Feb 24 18:07:41
"ferguson effect" is about as close to solid social science as you find

so no surprise that sebs make homicide rate increase by record year-to-year change with their george floyd and blm garbage

http://the...0-murder-homicide-rate-causes/

it's sebs who are stupidly killing blacks
Seb
Member
Thu Feb 25 02:07:31
Nim:

*Sigh*. I'm not interested in archeology right now.

But seriously, if your response to decades of examples of police shooting black men under dubious circumstances and facing no consequences, as the cherry on top of numerous other specific examples of structural discrimination, is to dig out a slide rule and table to see if you can use fairly patchy, incomplete and biased data sets (how trustworthy is the data on unarmed black men when there are numerous cases where guns, drugs, knives etc. have been planted after the fact to justify an outcome that shouldn't have occurred?) to determine if the well documented phenomenon exists in the stats, I think you are missing the point.

Failure to meet a large section of the publics demands on issues that they think critical results in civil unrest, because legitimacy and consent to be policed/governed has been lost. This cannot be rectified by a graph, nor can it be dismissed with a graph.

Seb
Member
Thu Feb 25 02:09:56
Werewolf:

I didn't say compared to liberals etc. I said compared to the general population.
show deleted posts

Your Name:
Your Password:
Your Message:
Bookmark and Share