Comment by DontchaKnowit | Hacker News Reader

Comment by DontchaKnowit | Hacker News Reader

Parent Full thread

What drives me nuts about the usage of statistics in common discourse is that it is used prescriptively instead of descriptively.

E.g. lets say 50% of gun owners accidentally shoot themselves. Now lets say Im talking to someone about how I want to buy a gun for home defense. They tell me "well 50% of gun owners accidentally shoot themselves, so you're safer if you dont have a gun"

This is a gross misunderstanding if statistics. There is no way whatsoever to take the observed general trend, and apply it to a future possibioity in a single instance. What this interpretation misses is that there are variables well within your control. A gun owner who follows all safety rules will virtually never shoot themselves accidentally. Turns out 50% of gun owners are stupid as hell, and you are in control of yiur own outcomes.

I see this problem CONSTANTLY in debates online, on the news, etc. It is very frustrating and contributes greatly to propogating dumb ideas.

There's a risk of considering oneself better than average and that aggregate statistics don't apply to you. Like most drivers consider them better drivers than average. I'd bet the same with gun owners or any other activity. You don't think you can repeat the stupid mistakes, but we all can, you get stressed, sick, drunk,... life is long and full opportunities to make a mistake and get you back to the average case.

By buying a gun and keeping it in your house you move yourself from a population basin of “I will never shoot myself” to one where “If I follow these dozen rules perfectly I will probably never shoot myself.”

You go from a stable basin to an astable mountaintop basin where you are responsible for the maintenance of the banks.

I'm sure you can appreciate that both imperfect causal inference from statistics and the belief that one is somehow "special" (or exempt from accident and disaster if one simply "puts in the effort") are pervasive forms of human folly that cause a lot of human misery.

Lots of humans are special though and don't have the same problems or weaknesses as average people. For example, some poor people manage to not get fat in USA, if you are one of those then likely many other things that applies to other poor people don't apply to you.

"Prescriptive" statistics in the sense you're describing are probabilities.

And they're both used and useful much of the time. The field of risk management and industry of insurance are based on it.

Your frustrations are not uncommon. They are, however, rather poorly-founded.

They're an excellent illustration failing to get a handle on this safety thing (where "safety" is the inverse of "risk", as one of yesterday's submissions addresses:

<https://news.ycombinator.com/item?id=34036978>

There's also the philosophical question around the "sea battle" question, in which the truth of a statement about a future possibility ("there will be a sea battle tomorrow") is assessed.

<https://en.wikipedia.org/wiki/Problem_of_future_contingents#...>

People ask each other where they're from, even though it gives you precisely zero amount of concrete information about a person, and then proceed to paint a picture going forward of that person based on assumptions about other people from there. They also do it to show interest, but it doesn't stop the painting.

So you say where you grow up has zero influence on who you are?

No, it does. I'm saying asking someone where they're from yeilds no facts about that person if all you get back is a name of the country. It doesn't mean they definitely grew up there, or that they hold citizenship. But it's easy to just take the name of the country and assume a bunch of things about the person based on things you heard about people from there. There may be statistical probabilities there but no certainty. People have done it since forever and it's essentially the same mechanism as OP described.

Mostly I do it to make conversation. I am almost never asking the question of random strangers, and probably in most cases the social context tells me more about the person than their growing up in Winnetka or Brookline would.

It is true that fifty years ago many people in Denver (and no doubt not only Denver) regarded New York as the embodiment of The Other and were very quick to pick up on a New York accent. But this was silly.

Reply on news.ycombinator.com