Posts by BenWilson
Last ←Newer Page 1 2 3 4 5 Older→ First
-
Access: Autism, ABA and being a bad mother, in reply to Mark Easterbrook,
My son is my son, a unique individual who at 2 years old was labelled as being on the autism spectrum. Can I choose to not use that label in the future? Sure. Can I rewrite the past and stop a doctor giving it to him? No I can’t. So apologies if the use of it has caused confusion or bemusement. I’m 3 short years into a long journey.
Wow, that sums up my situation in a nutshell. So well put. My boy might be autistic. Or maybe he's just different because of his other disabilities, stemming from brain damage. But he has autistic features and that's enough to put him on the spectrum. In the end, does it matter?
Yes, and no. To me, how I think about him daily, no, it doesn't matter much at all. But when it comes to leveraging support, explaining his situation, and giving people hints about how to deal with him, then it's a useful idea. It talks to a difference in communication style. Sometimes, it does more harm than good, since autism is not the beginning and end of his problems. It might even not be the biggest problem - I don't think it is. Much more severe is his limited eyesight and gross motor development. But the toolkit for dealing with autism is really good. It's good stuff even for non-autistic children - all people communicate through non-verbals, and it's always a good idea to try to judge to what level you're communicating, and what a child can even hear, let alone understand, in an stream of adult words. Sometimes it's a toolbox that isn't very helpful. Then I put it down for a bit.
-
Hard News: A law gone awry, in reply to Steve Barnes,
is still bullshite in my humble opinion.
I think it's bullshite too, but for a different reason. Which is that it could be true but missing the point. Something could be a contributing factor in more fatal road crashes, whilst not contributing anywhere near as much to the road crashes. Not all contributing factors are equal. Minor contributing factors can be found in big data sets, especially if they pertain to frequently occurring and clearly measurable things, like the presence of some traces of marijuana consumption - you'll have a lot of data about that across thousands of fatalities.
Just been doing a bunch of this in a statistical data analysis course. It's been an eye opener about what a black art modeling data really is. The well known issue of over-fitting is there, because of course you can fit any data set with a sufficiently high order polynomial - but you get a useless model. So you reduce the number of factors you look at, modeling it around the terms with the greatest statistical significance (NOT the greatest effect). You put terms in until they start dropping below your chosen significance level. The more data you have, the tighter your confidence intervals get, and the better chance you have of finding things that have statistical significance. They do not have to be big effects for this to occur, if your data is big.
-
Hard News: A law gone awry, in reply to Ian Dalziel,
Yes, they're a danger. To be consistent, I should be saying that there is a counterbalancing utility in being able to use those things while driving. That utility isn't worth nothing at all. I don't know if it adds up to enough to justify death on the road for some people. But there is also the question of what could be done about it. Banning it seems pretty ineffectual at stopping it. But it might make people who do it a lot more wary of being caught, which means that they will at least be concentrating on the road somewhat, looking out for police and keeping a low profile during it. And as an imposition, it's pretty minor. Except on the motorway, it's usually possible to pull over to use the phone. So it should probably not be allowed.
On the other hand, we already have laws about causing accidents, which is actually what's really important here. It's not the using of the phone which is the harm, it's the use of the phone leading to harm. It will usually be pretty clear who is at fault when an accident happens in which distraction was the cause, and careless driving carries penalties already. So it's kind of not allowed anyway - at least it's not an acceptable excuse for a lapse in concentration when involved in an accident.
Seems like a good policy to ban it, and leave it to police discretion. They'll probably only notice the cellphone user who is a completely oblivious fool, driving carelessly, and such people should certainly get a rap on the knuckles. The rest will either abide by the law scrupulously, or they will disabide it, but carefully (which is a good effect). I wouldn't care for police time to be wasted picking on people talking on their cellphone whilst stopped at the lights.
-
On of the more problematic parts of this is to establish a baseline of how stoned the average population is the rest of the time, when they’re not in a car. If 75% of accidents involve someone under the influence, but we find that 75% of the population is under the influence all the time, then we haven’t proved any kind of connection.
My point here isn’t to say these statistics are proving anything. It’s to say that we need to be careful when challenging them that we haven’t actually missed the point. Because they might come back later and prove the thing we denied, and by even engaging in the discourse we implicitly approved the underlying argument that all they have to do is prove some harm. I say that’s not enough, and to challenge the juggernaut that is Big Data is a mug’s game. Our part as humans is not to hope the numbers come out in our favour, it’s to seize control of the decisions that we make using the numbers so that they reflect our real values and aims.
I say this as someone who has experienced many times, firsthand, the danger of letting blind mechanistic optimizers make decisions, when their “objective function” (what it is that they are trying to minimize or maximize) is wrong. They will very systematically fuck things up to a much greater degree than any person could. And I think we have our social objective function all wrong, if harm minimization is our aim. It should be (harm – good) minimization.
Otherwise we unleash the power of science to ruin us.
-
Hard News: A law gone awry, in reply to Steve Barnes,
I would posit the proposition that the majority of people who have just smoked cannabis to the level that could influence their driving can’t be fucked to even get in a car, let alone drive it.
I really don't know if that is true. This is the land of the automobile. When you want to do stuff, a car is pretty common. If you don't think it impairs you, then you will often drive, if it's convenient for you to do so.
I think that there's some truth in what you say. I certainly don't think stoned driving is a big contributor to accident statistics. But that doesn't mean it's not a contributor at all. And with sufficient data, what small level of contribution it does make can be ascertained with a high level of confidence. Even if, on the whole, stoners don't drive when really blazed (if that is true), and they're only a little bit impaired, and the way they're impaired tends to make them more careful drivers, there are still going to be accidents involving them, in which the impairment was a factor - if you have enough data.
To me that seems highly likely. Even if only 1% of the time someone who is very stoned decides to drive and 99.999% of those trips don't end in an accident, there are millions of people getting wasted, so there's going to be a ton of data on people who are in accidents, and stoned as. And if it was a contributing factor, that will probably show up in those data.
Hence my whole point. It's not about whether it is a factor, it's about how much it's a factor, and whether or not, even if it is a factor, prohibition actually helps change anything (we could still, after all, have a law against stoner driving, without prohibition. We do already have laws against driving under the influence, and also for alcohol, which is not prohibited generally), and even after all that, whether or not the contribution of all of that is enough to override the basic right to consume cannabis.
In the case of alcohol, the prohibition against driving under the influence is well founded, in my opinion. With cannabis, there is likely to be some issue, but the question of whether it's significant is important. There are a great many things that have influence over our ability to drive, and cannabis may be well down the list of danger behaviours. It might be that having a working car radio is more dangerous. Or a working cellphone. It's well known that being very tired is extremely dangerous, and yet we have no laws against that, except in the case of professional drivers who have to keep logs of their time behind the wheel. We don't legislate against those things because it is felt that the right to have a radio or a cellphone is important enough to preserve, and we can't even measure how tired people are, except perhaps subjectively (and of course they will lie if they could get an expensive ticket).
But, at the end of the day, even if it's clearly a factor that we'd actually care about, that still doesn't imply that prohibition is the way to prevent it. It just means that prohibition of driving while stoned is probably a good idea. I even think it actually is a good idea, so long as a decent link between impairment level and some measurable chemical signals can be developed. We're not there yet.
-
For something that should be prohibited from general usage, because of the harm, the obvious one is guns. Their whole purpose is harm. I do see the right-to-a-gun argument, but practically, I don't think the huge number of gun related murders justifies the ownership of something that has little other purpose than to produce extreme harm. There are obvious exceptions. Police should be able to get them, if a criminal is armed. The army should obviously have them. Farmers should probably have limited access to them for pest control. And in a highly controlled sporting environment they should be available, with many, many caveats.
-
A friend gave me a very good analogy the other day on this. He pointed out that gay rights didn't come about because of a harm minimization argument. They didn't have to show that the harm from being gay was non-existent. They saw quite clearly that the path to freedom was not to allow of the idea that society had any right to interfere, even if there is harm in being gay. They simply have the right. It really is that simple.
I think rights are not sovereign - in many cases they should be overridden when they cause serious harm. But the threshold should be pretty high, and the flipside always has to be considered.
To see the flipside, you need only consider that there is a very strong correlation between vehicle-related deaths, and driving being legal. If all driving were banned, then there could be no vehicular accidents. People might still make sneaky cars and drive them, without any control, but car usage would generally be basically eradicated. But would that be good? I don't think so, I think cars have tremendous utility, that more than justifies their usage, despite them being THE major cause of death by accident in this country and most others. From a pure harm point of view, there's an extremely strong case to ban them. So much the worse for the pure harm point of view.
-
Also, considering that the default is that cannabis is illegal, even if they never show any harm from it whatsoever, there is still the possibility that they might find some in future, to justify continuing to ban it. This is a tenable argument if all all focus on this harm becomes the only justification. This is precisely the argument that we just used to ban all psychoactive substances, even ones that haven't even been invented yet. It's not only an argument that flies, it's an argument that wins.
I think I may have hammered this point enough. But then again....
-
Hard News: A law gone awry, in reply to Russell Brown,
But they can’t really determine impairment or even whether the driver had smoked pot in the past 24 hours. It’s a very, very perilous measure to stake big claims on.
It is. But the war to undermine correlations found between cannabis use and increased accident rates, to show that cannabis should be legalized, is one you can only lose. You might win many battles, show many stats are spurious. But in the end, with enough data, it will probably be shown that the chance of an accident is increased by using cannabis, with very high levels of certainty. They aggregate it into a binary factor (used or not-used) precisely because the levels of impairment are not clear. But if a correlation exists, there is also a pretty clear mechanism (impaired judgment) that gives us reason to suspect that the correlation might also be on account of some causation.
What I'm saying here builds on what I've said before about harm minimization. What is dodgy about this reasoning isn't the level of confidence in harm found. That might be statistically significant to the 99.9999% level. What is dodgy is that only harm is taken into account, and that any harm is bad. The amount of harm could still be incredibly small, even if we have extremely high levels of confidence that it exists. It could be the Higgs Boson of harms, but we'll still be pretty sure it's there, deep in our Big Data analysis.
To me, the question will always be "how does that harm balance against the good". Also "what other harms are caused by the alternative course? (in this case continued prohibition)". If one extra person is killed in accidents every million years because of cannabis use, and we're very, very sure about that, does that justify:
-That every single cannabis user should have to be a criminal for a million years?
-That every use of it should be accompanied by secrecy, paranoia?
-That a whiff of it can invoke a fishing expedition by police?
-That people get criminal records for possession?
-That the production has to be done by gangsters with no controls at all?
-That tens of thousands of citizens are guilty not just of possession, but of the more serious charge, supply?
-That billions of law abiding citizens will never ever even try it, thus never finding out if they actually like it, that trillions of enjoyable highs might never be had?I think it does not. Indeed, considering that alcohol clearly kills a lot more people than that on roads, we can at least, as a society, see that it's not as simple as a clear cause of some harm in some situation. If this really is all the police have as their ongoing reason for opposing decriminalization, they've got basically nothing.
Which is of course NOT their only reason. Cannabis being illegal is extremely convenient for the police. It's one of the few drugs that has a strong smell, and the plant is recognizable at a glance. They can invoke the Misuse of Drugs act with much greater confidence. Prohibition gives them power, and not just over cannabis.
-
Hard News: A law gone awry, in reply to Steve Barnes,
It seems likely that there is a correlation, though. And that does make it rather tempting to ascribe that to causation by virtue of fact that we know our judgment is affected by cannabis.
What is missing is saying how much of a contributing factor it is. It is very possible to find a statistically significant factor that only contributes a very, very small amount to the overall explanatory power of the model. Maybe if you model crashes using only alcohol and excessive speed you can explain 90% of the variation in the data. Adding in cannabis you get 90.1%. It's still statistically significant, but possibly not one to focus on compared to other possible indicators. But the temptation is to use it, because it's data that they can certainly collect. They can measure the drug levels of people in the crashes. They can't always know the speed of the crash, and definitely can't reliably know if the person was on a cellphone, or just had an angry fight with their boss and was in a state of blind rage.
Of course police will be against decriminalization. When has it ever been different? Fortunately, the police are only a "contributing factor" to how our laws are decided.
Last ←Newer Page 1 … 252 253 254 255 256 … 1066 Older→ First