3 ms·

Much of this criticism seems to be "Economist says X causes Y in situation A, assuming B. Politician uses that to force change in X. How dare the economist igno

Much of this criticism seems to be "Economist says X causes Y in situation A, assuming B. Politician uses that to force change in X. How dare the economist ignore Z?!" Like the draft example. An economist estimated the effect on labor, politicians use the finding to abolish the draft, then the economist gets shit on for not considering the effect on civil-military relations. It's much the same, but more obviously ridiculous if you rephrase it with a physical scientist: Chemist estimates some pesticide is lingering in unsafe levels on one particular kind of fruit in certain limited circumstances, then politicians use the finding to enact a sweeping farming regulation, then the chemist gets shit on for not considering the effect on the price of food. As a data scientist, I find this happening to me sometimes. I tell a PM that we can say X causes Y, but can't really say what the relationship between X and Z is, and people get really frustrated by that. Then they really want an estimate. If you tell them, "fine it's 3 +/- 15" all they hear is "3" and now it's your fault if they make decisions as if we knew for a fact that 3 is the answer. Damned if you do, damned if you don't. I'm sorry that the scientific method isn't an oracle, but that's life. We still have to make decisions.

bertil 7y ago
Don’t say “3 +/- 15”, say “between -12 and 18”. Avoid showing the number you don’t want them to see.
- 6gvONxR4sf7o 7y ago
  I certainly agree with this and follow this, but even then I've found communications downstream in the game of telephone using the midpoint of what I tell them, dropping any mention of uncertainty. In fact, many times it's the very first thing they do. :facepalm:
  - bertil 7y ago
    What I’ve found work quite well is to remove the mid-point entirely, event graphically, represent the edges with more than three significant numbers (which I’d happily argue against, except): it makes the mental math to find the average harder, and less likely to happen. Finally, explain that out of 10 or 20 results, one has their actual impact outside of that zone. Show simulation.
xkcd-sucks 7y ago
> Chemist invents a fuel additive which improves fuel economy and extends engine life, then gas+car companies integrate it into their product, then everyone gets shit on for not considering the effect of tetraethyl lead on human health... after the damage is already done
- 6gvONxR4sf7o 7y ago
  The way this would usually pan out is the chemist publishes a paper on "this one chemical seems to have high energy density after undergoing a particular reaction." Then car companies rush to put it in front of the public. You're really going to include the chemist in the blame there? Tetraethyl is an easy case, because it was car companies who realized its use in fuel in the first place, so we don't even need to figure out who to blame, unless you want to blame its original inventor who did it almost a century earlier, before the gas engine was even invented.
jayd16 7y ago
Your chemist analogy doesn't fit. No where is it in a chemist's job description that they should consider the price of food where as an economist's job is to consider the economy and consequences of change.
- 6gvONxR4sf7o 7y ago
  No that's exactly the point. There are a ton of kinds of economists who each study specific things. It's as limited as any other science. Of course you get people like Neil deGrasse Tyson in any field, who feel its their job to talk overconfidently about all things, but that's another issue.

bertil 7y ago

Don’t say “3 +/- 15”, say “between -12 and 18”. Avoid showing the number you don’t want them to see.

6gvONxR4sf7o 7y ago
I certainly agree with this and follow this, but even then I've found communications downstream in the game of telephone using the midpoint of what I tell them, dropping any mention of uncertainty. In fact, many times it's the very first thing they do. :facepalm:
- bertil 7y ago
  What I’ve found work quite well is to remove the mid-point entirely, event graphically, represent the edges with more than three significant numbers (which I’d happily argue against, except): it makes the mental math to find the average harder, and less likely to happen. Finally, explain that out of 10 or 20 results, one has their actual impact outside of that zone. Show simulation.

xkcd-sucks 7y ago

> Chemist invents a fuel additive which improves fuel economy and extends engine life, then gas+car companies integrate it into their product, then everyone gets shit on for not considering the effect of tetraethyl lead on human health... after the damage is already done

6gvONxR4sf7o 7y ago
The way this would usually pan out is the chemist publishes a paper on "this one chemical seems to have high energy density after undergoing a particular reaction." Then car companies rush to put it in front of the public. You're really going to include the chemist in the blame there? Tetraethyl is an easy case, because it was car companies who realized its use in fuel in the first place, so we don't even need to figure out who to blame, unless you want to blame its original inventor who did it almost a century earlier, before the gas engine was even invented.

jayd16 7y ago

Your chemist analogy doesn't fit. No where is it in a chemist's job description that they should consider the price of food where as an economist's job is to consider the economy and consequences of change.

6gvONxR4sf7o 7y ago
No that's exactly the point. There are a ton of kinds of economists who each study specific things. It's as limited as any other science. Of course you get people like Neil deGrasse Tyson in any field, who feel its their job to talk overconfidently about all things, but that's another issue.