A while back Microsoft was paying hosts and registrars with large numbers of domains parked, or $30 / year type, to switch over. I don't know if that program is still active.
First, you're looking at the wrong number mathematically. The proper statistic would be "X% of black people are criminals". Remember you're trying to estimate whether or not a particular person is a criminal, not who a particular crime was committed by. because a very small number of black people or white people are criminals race is not a good indicator that a particular person is engaging in criminal behavior.
Secondly, there are behaviors that are suspicious and other evidence. In my city, for example crack dealers circle the block on bicycles in the middle of the night. If X% of crack dealers are black, approximately X% of the people circling the block in the middle of the night will be black. If you focus on the specific suspicious behavior, you'll get more accurate results AND whatever racial proportion is correct will be looked at as a side effect, without ever considering race. So mathematically, it's best to ignore race when stronger indicators are available.
Lastly, other commentors have discussed the EFFECTS of racial profiling. Harassing people on racial grounds also has negative effects.
Interestingly, your concept of "true probabilities" DOES work for solving a specific crime. If, in a certain city, the italian mob does professional hits, then when investigating a professional hit it's mathematically correct to have a look at Italian mobsters. I'll say it - if young middle eastern men are normally the ones who hijack planes, it makes sense mathematically to check which young middle eastern men are booked on the flight, AFTER you have evidence that flight is involved in a hijacking attempt
Ps. I don't think that means there can be no effective study. To find people who want to sober up, and therefore might be helped to do so, look at people CHOSE to go to AA as opposed to people who were ordered to visit.
That's what you'd do with a fertility drug - you'd study people who were trying to get pregnant.
Agreed. Further, I think you hit an important point:
"If you send alcoholics to AA, you can divide them into three groups."
It would be wise to recognize that if you're SENDING people to AA, those are people who didn't already choose to be there, so most of them will be in group #1, people who don't want to be there.
So if you count all of the people SENT to AA by a court or whatever, you're counting mostly people who shouldn't be counted. Not for the purpose of determining how well AA works.
Empirical evidence says roughly about 6% of those sent by court want to sober up. 6% want to sobering up, 4% do sober up. Since the purpose is to help people who want to stop, that's an 66% success rate or so for the first attempt. (success meaning helping people achieve their goal of sobering up - not forcing the someone's goal upon them).
Another 6% will want to quit at some future time in their lives. For many of them, their exposure to AA plants a seed that gets them somewhat sooner than they would have sobered up otherwise. When those 6% decide they want to, they know where to go.
> the app will then wait until the user is trying to install another app and will pop up an extra dialog box asking for permission
A couple of weeks ago when I described this attack, some commenters said it was impossible - an app can't wait until the user was expecting a permission prompt from a different app, then request more permissions itself, they said.
I wonder if they still think it's impossible now that it's publicly reported to be in the wild.
> What part of > > These periods of elevated and variable CO2 levels are so different from the typical measurements that is easy to > > > remove them from the final data set using a simple mathematical “filter.”
> don't you understand?
What part of 4:30 AM don't you understand? It's right there int he subject line. If you look at the data, 4:30 NOT "so different from" 4:00 AM or 5:00 AM that there's any objective reason to include one and exclude the other. As it happens, these readings that entirely subjective, like 4:30 AM, are precisely the peak readings people are getting excited about.
> And, given that it comes up with the same results as all the other measurements we have, what is your problem?
I've been asking and asking for citations to any other similar result. Dozens of people have replied to those posts, but not one has cited a single measurement, anywhere in the world, even in the middle of LA, with readings anywhere NEAR that high. I just checked the reading in downtown Houston, TX, one of the country's dirtiest cities. It's 0.2 ppm. These guys are claiming overall atmospheric C02 of 400 ppm. So according to them, Houston has air 2000 times cleaner than the atmosphere is general. I call bullshit. If you know of any other readings coming anywhere near 400, cite them. Since you don't, you just thought that's what Jon Stewart said, stop repeating it. It's bull. 2000X bull.
Oh yes, wikipedia is great. Of course _I_ am one of the authors there, so GP would probably agree it's not the most reliable source, er non-source. A lot of what's in wikipedia is cited. A lot isn't. Some came directly from my rear end. Overall, it's a great summary, and a great place to FIND citations to actual sources, which you might then cite yourself.
I just wouldn't cite it, because anyone interested in viewing the the source has to go to wikipedia, see if it's cited there, find the citation, and follow it, then look to see if the cited source is reputable. Better to just cite the source. Compare these two citations from the same actual source:
"Guns are great" [wikipedia.org] "Guns are great" [nra.org]
If NRA is the course, citing the source directly means the reader doesn't have to click ANYTHING to find out what the source is, and if they consider that source to be credible. Citing the wikipedia citation instead obscures the actual source. For the same reason, you don't cite Britannica, you cite the same source Britannica does.
In a discussion about not citing wikipedia or any encyclopedia, you cited wikipedia saying "wikipedia is reliable". I'm shocked an amazed that wikipedia isn't critical of itself.;) Seriously though, that IS a good summary and list of citations. If I had more time, I'd read more of it.
Yes of course they need to disregard some of the data from this particular station, which is the problem with this monitoring station. In the sample graph on the page, they chose to include the 4:30 AM reading, which looks like it may well be part of the volcanic breeze. Maybe it is, maybe it isn't. It's entirely subjective whether or not you want to include that.
Subjective massaging of data like that represents the person's OPINION, not an objective measurement.
Prove it? You think they were lying? They said they disregarded 15% of the days. They selected 15% as opposed to 14% or 18%, which would have given different results. The data is there if you want to calculate it for different values than the 15% they chose.
Prove it? You think they were lying? They said they disregarded 15% of the days. They selected 15% as opposed to 14% or 18%, which would have given different results. The data is there if you want to calculate it for different values than the 15% they chose.
>. Then they DELETE THOSE RECORDS. If they wanted to get evidence of increased CO2, according to your model they would leave those in.
They chose to delete 15% of the readings. They could have used to same explanation to delete 12% or 18% and ended up with any result they wanted. That is, from what they said, they chose to delete only the ones that were WAY over the top, but leave in the ones that were "only" 30% higher than average. That doesn't inspire confidence.
> Their readings are consistent with other monitoring stations.
I've seen the Mauna Lua estimate cited many times. I've never seen a consistent reading from any other station cited. Do you have a citation for that? Preferably a sensor that's not inside the smokestack of a Chinese foundry?
My point is not that AA is the best program. Certainly in absolute terms it's helped more alcoholics, but that's beside the point.
My point is that AC's attempt to judge AA by it's ability to do something contrary to it's purpose is ridiculous. AA is very, very clear that they have no interest in convincing people to stop drinking. I could fill pages with quotes from the AA book indicating that. To say that AA doesn't convince people stop drinking is like saying socks aren't effective because they don't improve your gas mileage. No shit, that's not the purpose.
If you want to say "treatment starts when someone hears about the treatment", aspirin is rarely effective. Everybody's heard of it, and most headaches go away without taking aspirin.
The fertility analogy is really more accurate, though. AC is claiming AA doesn't work because it doesn't have much effect on people who WANT to keep drinking. That's precisely the same as claiming that condoms are ineffective because their existence doesn't stop people who WANT to have a baby.
The only 400 reading publicized was here on Slashdot a few weeks ago. The measurement was from atop Mauna Lua - a volcano that emits CO2 constantly. You'll note that the defender below acknowledges that fact while saying it's okay because the person who did the reading edited the data to come up with the 400 estimate.
That's seriously the best "defense" - "it's okay, we cooked the data to say 400 because we think that's what it might have been if we weren't on a volcano."
Oh okay, I'm it's perfectly objective, then. As long as the guys trying to get evidence of increased CO2 are admittedly excluding 15% of daily data, it's a perfectly objective source.
Or you could report readings that AREN'T on top of a giant CO2 vent. Seriously, you have to REALLY want to believe them if you need to pretend that's credible. Don't you have any interest at all in knowing what's actually going on? I can sort of understand fandom, though I prefer intellectual honesty. Even hardcore sports fans WATCH THE GAME to find out what happened, though. They don't just pretend their team won.
And that "alarming" measurement was on a volcano that emits CO2. It's unfortunate that most climate science isn't science. There IS something going on, something very subtle, but it's nearly impossible to any real understanding of it because there is so much politics.
Most everything published on the topic isn't just slanted, it's so ridiculously off to one side or another that's it's useless. I mean seriously, you take your "atmospheric" CO2 measurements on a volcano? That's not even acceptable by propaganda standards. Baghdad Bob would laugh at you and suggest that somewhere in the desert or sea would be much more believable, just do it a 200 miles downwind of Los Angeles.
Are you sure about those millions? Last I heard, it was several thousand. At a total cost of around a trillion dollars, something like $275,000 per new recipient.
I guess if you count all of the people who were already getting subsidized healthcare and are being moved to Obamacare, there may be millions whose healthcare was changed from one program to another.
Agreed. The JavaScript on the comment textbox also screws up in different ways on different browsers, mostly taking the word you just typed and moving it to replace some letters in a a word several sentences prior.
The birth control numbers are people who were trying not to get pregnant, and used those methods, perfectly or imperfectly.
Do you have numbers for how often condoms prevent pregnancy in people who are trying to get pregnant? For people who never bought them? Of course not. Yet some people insist on using people who want to get drunk as the gauge for something designed to help someone who wants to NOT drink.
To directly answer your question, none of the above. I would not educate a fertility patient about birth control, and I wouldn't advise someone who wants to drink regarding a method of doing the opposite.
If you're comparing treatments and there's a difference, sure compare patients level of compliance and report that ALSO. Nobody studies non-patients with no interest in any treatment and from that concludes that the treatment "doesn't work" though.
You could order me to go to a bar. When I didn't get drunk, you wouldn't conclude "whiskey doesn't get people drunk". You'd conclude that TELLING SOMEONE to go to a bar doesn't get them drunk, when they don't care to do so.
Your article claimed that because ordering Lohan to visit a rehab doesn't work, no method for getting sober works. That's silly. The far more reasonable conclusion is that ordering someone to visit a place doesn't cause them to do anything they don't want to do.
If the judge ordered her to visit a law school and she didn't become a lawyer, you would conclude that she probably didn't want to become a lawyer. At least, not enough to put forth the effort to do law school.
If the judge orders her to visit a rehab and she doesn't do anything to become sober, the most likely conclusion is that she wasn't interested in being sober. At least, not interested enough to put forth the required effort.
In neither case does her lack of interest indicate anything at all about the effectiveness of the programs offered, making the article's primary premise garbage.
>. And you can't disregard people who drop out of the program or don't follow it.
In that case , law school is garbage. It doesn't work, since 99% of people don't become lawyers. Exercise most certainly does not work for improving health, because he doesn't work for people who don't exercise. Condoms don't work then, people who don't use them get pregnant. AccoAccording to your reasoning, virtually nothing works for any purpose.
Slashdot Mirror
A while back Microsoft was paying hosts and registrars with large numbers of domains parked, or $30 / year type, to switch over.
I don't know if that program is still active.
funny
First, you're looking at the wrong number mathematically. The proper statistic would be "X% of black people are criminals". Remember you're trying to estimate whether or not a particular person is a criminal, not who a particular crime was committed by. because a very small number of black people or white people are criminals race is not a good indicator that a particular person is engaging in criminal behavior.
Secondly, there are behaviors that are suspicious and other evidence. In my city, for example crack dealers circle the block on bicycles in the middle of the night. If X% of crack dealers are black, approximately X% of the people circling the block in the middle of the night will be black. If you focus on the specific suspicious behavior, you'll get more accurate results AND whatever racial proportion is correct will be looked at as a side effect, without ever considering race. So mathematically, it's best to ignore race when stronger indicators are available.
Lastly, other commentors have discussed the EFFECTS of racial profiling. Harassing people on racial grounds also has negative effects.
Interestingly, your concept of "true probabilities" DOES work for solving a specific crime. If, in a certain city, the italian mob does professional hits, then when investigating a professional hit it's mathematically correct to have a look at Italian mobsters. I'll say it - if young middle eastern men are normally the ones who hijack planes, it makes sense mathematically to check which young middle eastern men are booked on the flight, AFTER you have evidence that flight is involved in a hijacking attempt
Ps. I don't think that means there can be no effective study. To find people who want to sober up, and therefore might be helped to do so, look at people CHOSE to go to AA as opposed to people who were ordered to visit.
That's what you'd do with a fertility drug - you'd study people who were trying to get pregnant.
Agreed. Further, I think you hit an important point:
"If you send alcoholics to AA, you can divide them into three groups."
It would be wise to recognize that if you're SENDING people to AA, those are people who didn't already choose to be there,
so most of them will be in group #1, people who don't want to be there.
So if you count all of the people SENT to AA by a court or whatever, you're counting mostly people who shouldn't be counted. Not for the purpose of determining how well AA works.
Empirical evidence says roughly about 6% of those sent by court want to sober up.
6% want to sobering up, 4% do sober up. Since the purpose is to help people who want to stop, that's an 66% success rate or so
for the first attempt. (success meaning helping people achieve their goal of sobering up - not forcing the someone's goal upon them).
Another 6% will want to quit at some future time in their lives. For many of them, their exposure to AA plants a seed that gets them somewhat
sooner than they would have sobered up otherwise. When those 6% decide they want to, they know where to go.
If the NRA wasn't the source, then indeed you wouldn't cite NRA, you'd cite the source.
> the app will then wait until the user is trying to install another app and will pop up an extra dialog box asking for permission
A couple of weeks ago when I described this attack, some commenters said it was impossible - an app can't wait until the user was expecting a permission prompt from a different app, then request more permissions itself, they said.
I wonder if they still think it's impossible now that it's publicly reported to be in the wild.
> What part of
> > These periods of elevated and variable CO2 levels are so different from the typical measurements that is easy to >
> > remove them from the final data set using a simple mathematical “filter.”
> don't you understand?
What part of 4:30 AM don't you understand? It's right there int he subject line.
If you look at the data, 4:30 NOT "so different from" 4:00 AM or 5:00 AM that there's any objective reason to include one and exclude the other. As it happens, these readings that entirely subjective, like 4:30 AM, are precisely the peak readings people are getting excited about.
> And, given that it comes up with the same results as all the other measurements we have, what is your problem?
I've been asking and asking for citations to any other similar result. Dozens of people have replied to those posts, but not one has cited a single measurement, anywhere in the world, even in the middle of LA, with readings anywhere NEAR that high. I just checked the reading in downtown Houston, TX, one of the country's dirtiest cities. It's 0.2 ppm. These guys are claiming overall atmospheric C02 of 400 ppm.
So according to them, Houston has air 2000 times cleaner than the atmosphere is general. I call bullshit. If you know of any other readings coming anywhere near 400, cite them. Since you don't, you just thought that's what Jon Stewart said, stop repeating it. It's bull. 2000X bull.
Oh yes, wikipedia is great. Of course _I_ am one of the authors there, so GP would probably agree it's not the most reliable source, er non-source.
;) Seriously though, that IS a good summary and list of citations.
A lot of what's in wikipedia is cited. A lot isn't. Some came directly from my rear end. Overall, it's a great summary, and a great place to FIND citations
to actual sources, which you might then cite yourself.
I just wouldn't cite it, because anyone interested in viewing the the source has to go to wikipedia, see if it's cited there, find the citation, and follow it, then look to see if the cited source is reputable. Better to just cite the source. Compare these two citations from the same actual source:
"Guns are great" [wikipedia.org]
"Guns are great" [nra.org]
If NRA is the course, citing the source directly means the reader doesn't have to click ANYTHING to find out what the source is, and if they consider
that source to be credible. Citing the wikipedia citation instead obscures the actual source. For the same reason, you don't cite Britannica, you cite
the same source Britannica does.
> http://en.wikipedia.org/wiki/Reliability_of_Wikipedia [wikipedia.org]
In a discussion about not citing wikipedia or any encyclopedia, you cited wikipedia saying "wikipedia is reliable".
I'm shocked an amazed that wikipedia isn't critical of itself.
If I had more time, I'd read more of it.
Yes of course they need to disregard some of the data from this particular station, which is the problem with this monitoring station. In the sample graph on the page, they chose to include the 4:30 AM reading, which looks like it may well be part of the volcanic breeze. Maybe it is, maybe it isn't. It's entirely subjective whether or not you want to include that.
Subjective massaging of data like that represents the person's OPINION, not an objective measurement.
Prove it? You think they were lying? They said they disregarded 15% of the days. They selected 15% as opposed to 14% or 18%, which would have given different results. The data is there if you want to calculate it for different values than the 15% they chose.
Prove it? You think they were lying? They said they disregarded 15% of the days. They selected 15% as opposed to 14% or 18%, which would have given different results. The data is there if you want to calculate it for different values than the 15% they chose.
Yes, they did. They said they disregarded 15% of the days. They selected 15% as opposed to 14% or 18%, which would have given different results.
>. Then they DELETE THOSE RECORDS. If they wanted to get evidence of increased CO2, according to your model they would leave those in.
They chose to delete 15% of the readings.
They could have used to same explanation to delete 12% or 18% and ended up with any result they wanted. That is, from what they said, they chose to delete only the ones that were WAY over the top, but leave in the ones that were "only" 30% higher than average. That doesn't inspire confidence.
> Their readings are consistent with other monitoring stations.
I've seen the Mauna Lua estimate cited many times. I've never seen a consistent reading from any other station cited. Do you have a citation for that? Preferably a sensor that's not inside the smokestack of a Chinese foundry?
My point is not that AA is the best program. Certainly in absolute terms it's helped more alcoholics, but that's beside the point.
My point is that AC's attempt to judge AA by it's ability to do something contrary to it's purpose is ridiculous. AA is very, very clear that they have no interest in convincing people to stop drinking. I could fill pages with quotes from the AA book indicating that. To say that AA doesn't convince people stop drinking is like saying socks aren't effective because they don't improve your gas mileage. No shit, that's not the purpose.
If you want to say "treatment starts when someone hears about the treatment", aspirin is rarely effective. Everybody's heard of it, and most headaches go away without taking aspirin.
The fertility analogy is really more accurate, though. AC is claiming AA doesn't work because it doesn't have much effect on people who WANT to keep drinking. That's precisely the same as claiming that condoms are ineffective because their existence doesn't stop people who WANT to have a baby.
The only 400 reading publicized was here on Slashdot a few weeks ago. The measurement was from atop Mauna Lua - a volcano that emits CO2 constantly. You'll note that the defender below acknowledges that fact while saying it's okay because the person who did the reading edited the data to come up with the 400 estimate.
That's seriously the best "defense" - "it's okay, we cooked the data to say 400 because we think that's what it might have been if we weren't on a volcano."
Oh okay, I'm it's perfectly objective, then.
As long as the guys trying to get evidence of increased CO2 are admittedly excluding 15% of daily data, it's a perfectly objective source.
Or you could report readings that AREN'T on top of a giant CO2 vent. Seriously, you have to REALLY want to believe them if you need to pretend that's credible. Don't you have any interest at all in knowing what's actually going on? I can sort of understand fandom, though I prefer intellectual honesty. Even hardcore sports fans WATCH THE GAME to find out what happened, though. They don't just pretend their team won.
And that "alarming" measurement was on a volcano that emits CO2. It's unfortunate that most climate science isn't science. There IS something going on, something very subtle, but it's nearly impossible to any real understanding of it because there is so much politics.
Most everything published on the topic isn't just slanted, it's so ridiculously off to one side or another that's it's useless. I mean seriously, you take your "atmospheric" CO2 measurements on a volcano? That's not even acceptable by propaganda standards. Baghdad Bob would laugh at you and suggest that somewhere in the desert or sea would be much more believable, just do it a 200 miles downwind of Los Angeles.
Are you sure about those millions? Last I heard, it was several thousand. At a total cost of around a trillion dollars, something like $275,000 per new recipient.
I guess if you count all of the people who were already getting subsidized healthcare and are being moved to Obamacare, there may be millions whose healthcare was changed from one program to another.
Agreed. The JavaScript on the comment textbox also screws up in different ways on different browsers, mostly taking the word you just typed and moving it to replace some letters in a a word several sentences prior.
In other news, Al Gore is expected to win California in the November 2002 election.
Editors - if 2002 is submitting stories, please warn them about "we have to pass the bill to see what's in it".
The birth control numbers are people who were trying not to get pregnant, and used those methods, perfectly or imperfectly.
Do you have numbers for how often condoms prevent pregnancy in people who are trying to get pregnant? For people who never bought them? Of course not. Yet some people insist on using people who want to get drunk as the gauge for something designed to help someone who wants to NOT drink.
To directly answer your question, none of the above. I would not educate a fertility patient about birth control, and I wouldn't advise someone who wants to drink regarding a method of doing the opposite.
If you're comparing treatments and there's a difference, sure compare patients level of compliance and report that ALSO. Nobody studies non-patients with no interest in any treatment and from that concludes that the treatment "doesn't work" though.
You could order me to go to a bar. When I didn't get drunk, you wouldn't conclude "whiskey doesn't get people drunk". You'd conclude that TELLING SOMEONE to go to a bar doesn't get them drunk, when they don't care to do so.
Your article claimed that because ordering Lohan to visit a rehab doesn't work, no method for getting sober works. That's silly. The far more reasonable conclusion is that ordering someone to visit a place doesn't cause them to do anything they don't want to do.
If the judge ordered her to visit a law school and she didn't become a lawyer, you would conclude that she probably didn't want to become a lawyer. At least, not enough to put forth the effort to do law school.
If the judge orders her to visit a rehab and she doesn't do anything to become sober, the most likely conclusion is that she wasn't interested in being sober. At least, not interested enough to put forth the required effort.
In neither case does her lack of interest indicate anything at all about the effectiveness of the programs offered, making the article's primary premise garbage.
>. And you can't disregard people who drop out of the program or don't follow it.
In that case , law school is garbage. It doesn't work, since 99% of people don't become lawyers. Exercise most certainly does not work for improving health, because he doesn't work for people who don't exercise. Condoms don't work then, people who don't use them get pregnant.
AccoAccording to your reasoning, virtually nothing works for any purpose.