Sunday, 21 February 2016

People are quite stupid. A look at reasoning and statistics.

And road managers and people advocating for people's safety have to keep you from killing yourself with your own incompetence. Not saying that you will preform poorly on a math test or anything like that, but humans have a lot of biases, they gauge the wrong risks, and tend to think that they are more capable than everyone else. This isn't an insult towards my audience. This is a statement about humans in general.

I learned a lot about bias and risk from Vsauce on Youtube. The survivorship bias might point us towards a correlation, perhaps we say that based on the fact that most people who come home after riding with a helmet, then we should all wear bike helmets, but probably the most important thing in science, which you should all know from school by now unless you are maybe in the third grade, is that correlation is NOT causation. And phrases that appear closely related is often doing something wrong in the equation P therefore Q, Wonderwhy on Youtube made a video on that.

An example of how correlation is not causation. The number of deaths while skiing (or possibly because of skiing. There's a difference) and the number of drivers killed by train crash is actually quite a close correlation, though anyone who actually knows anything about math knows that should be used to imply cause. Or, you could say that all ravens are black, everything that's not black is not a raven and this green apple is not a raven. Clearly the apple and raven shouldn't have anything to do with each other, yet all of these phrases are true.

Of course there must be logical explanations for things that do happen. For example we find that roundabouts in the Netherlands with separate bicycle paths have much better safety records than those roundabouts that do assign cycle priority on an annular bicycle path. It could be that the priority roundabouts were unlucky, that they have higher traffic, that they handle less trained drivers or younger children, but the simplest explanation that works for roundabouts is that drivers simply have too much to do in order to let cyclists proceed safely, and especially not with the sightlines of ring shaped bicycle paths. The simplest explanation is usually the right answer, though there are of course exceptions.

Some correlations do lead us to useful results. If you do see a correlation, often it leads to finding reasons for that correlation that use the deductive reasoning method. And sometimes invalid reasoning can be perfectly true. For example: The Sun will rise tomorrow. How can I prove that? Just because it's done that every day that we know of before? Yet it is one of the most obvious things that really happens.

Things that are not obvious must not be overlooked. Nor things that are counter intuitive. The Dutch found that even though you are free to buy up to 5 grams of marijuana if you are 18 or over in Amsterdam, the usage rate is actually lower than Canada and especially the US. That shouldn't make sense at first glance. But for various reasons, it works. Bike helmets too. If you make people wear helmets, that should be a medical benefit. But apparently, the rate of cycling declines so much that the health benefits that outweigh helmets are lost, and possibly more reckless behavior. I mean if you are in a tank like a Sherman or a T34, vs a car, which one would you more dangerous things in willingly?

Of course there must be a logical reason for the Dutch cycling so much more than other industrialized nations. Wealth? No, The Dutch are among the richest in the world. Training? No. Even an idiot can ride a bike. All you need is balance and the ability to push your feet up and down and control a handlebar. Car driving incidentally needs very little training to get fairly good at too. I was driving on a 90 km/h expressway on my second lesson (true story). The weather? Australia has pretty warm weather and that doesn't seem to make cycling more attractive enough that 30% of people use it on a daily basis for their practical journeys. It's weather is very similar to Britain, especially Kent, and yet Kent has much less cycling.

No. The reason is the existence of cycle paths and low volume low speed roads that make the final link between origin and destination, plus the efficiency offered by the bicycle paths and the bicycles they ride. These are the core reasons. Various things could make it a bit more attractive, but these are the things that make or break the cycling society.

When you look at data, and this goes for everything not just cycling and walking, you must examine the data vociferously. And you must look at the context, error margins, how it was reported, why it's being collected, how was it collected, over what time frame, what other things could be influencing . the results. Look at all the data you can, and don't ignore conflicting data without at least seeing if the data is right or wrong yet. Use the scientific method, question, hypothesis, observation and confirmation (simplified) to look at the data you have. And make sure that your sources are reliable and could be repeated as much as possible. For example if you decide to grow two plants with different fertilizers but everything else was controlled, then the growth rate and end height of each plant should be the same if you try the experiment in April of this year or the next year for example. Control the variables as much as possible, ideally all but one (if you controlled all the variables then what are you testing for?). You must be accurate with your data if you are to be a reliable source yourself and not draw yourself into fallacies. You must make sure that you aren't prioritizing or denying something just because it's new or hyped about.

Your beliefs are your beliefs, but what you actually do with other people should be based on facts that can be proven. If I was designing a road, shouldn't I use the most accurate information and not just make people wear a helmet and say "off you go" in the 60 km/h divided 4 lane arterial 20000 vehicle per day traffic while riding a bicycle? Would anyone accept that with their children or grandparents, husband, wife, friend, other special relationship?

People can be stupid, only when we not ignore our biases, and look at actual data can we avoid the stupidity.

No comments:

Post a Comment

Thanks for commenting