Wednesday, November 4, 2009

Misinformation technology

I was just having a conversation with a friend about vaccination, and marvelling at how this mass movement of spectacularly ill-informed people has arisen to rail against, of all things, giving people injections to prevent them from getting crippling or deadly diseases.


Autism is bad. Polio is worse. Mercury poisoning is bad. Diptheria, tetanus and hepatitis are worse. Doctors are people who, for all their flaws as people, spend years studying how the human body works. Jim Carrey is a rubber-faced comedian who makes fart noises.

One would think this would represent a decisive end to the debate.

I think the story of information technology has not even been a teensy bit told yet. It's still so new that people are marvelling at what it can do and what it appears to have done. But it's like standing there after a giant meteor has struck and saying, "Gee, there sure is a lot of dust in the air... I guess the main effect of that giant meteor has been to make a lot of dust." But you're not even close to seeing the whole story. Wait for some of the dust to settle, or to block out the sun for a few seasons.

But it's been difficult so far to articulate a position of suspicion or caution because it's hard to get the nuance across. ("What, you don't want people to have access to information? You think it should be controlled by someone?") It's hard to be anti-information and that's how suspicion of information technology is liable to be taken.

The availability of blogging is great to the extent that it allows smart people to get their views out there without being censored by e.g. corporate media concerns. It's bad to the extent that it allows dumb, ignorant or misinformed people to get their views out there without being censored by e.g. a basic regard for science and facts.

How many are there of the latter, compared to the former? What proportion of each group blogs? How much influence to bloggers have? I don't know. Maybe I'm worrying about nothing here.

It makes it all the worse when the "quality" information--peer-reviewed scientific journal articles, which, whatever their flaws, have at least been evaluated by people qualified to do so before being published--is restricted to people actually in universities or willing to pay thousands of dollars for a subscription to Elsevier.

But information technology is even more dangerous in other applications. I was having a conversation in class today, where we're learning to use Excel to do moderately complex techno-economic modelling. The prof warned us to be careful, because sometimes you can construct a model that makes your conclusions appear extremely sensitive to a change in one variable--much more so than they are in real life. (Anything where you use a quotient i.e. rate of return as a percentage of an investment is particularly vulnerable to this.) Any model is limited in its ability to reflect reality, and sometimes numbers do lie--or, rather, they conceal and misdirect.

My dad always says, "He who lives by the spreadsheet dies by the spreadsheet." More generally, I got to thinking that once upon a time, if you wanted to work in a job that involves a lot of math, you had to be smart. Smart enough to have learned the math. That doesn't mean wise, necessarily, or good, but at least of a level of intelligence (whatever that means) to be able to learn challenging and not-immediately-exciting stuff. It demanded, arguably, a certain strength of character in this regard. Homework sucks. If you're going to learn math, you have to do a lot of it.

But now that we have spreadsheets, you don't have to learn how to do math. You can create all kinds of complex models with only a bare understanding of what these concepts represent.

Excel has a function called NPV (Net Present Value) that allows you to reduce the value of the future with a couple of keystrokes. Frankly, I have a lot of problems with NPV and discounting, but at least, if you have to understand the math yourself, there's an opportunity for it to become apparent what you're really assuming here.

But with a pre-programmed spreadsheet function, you add a level of abstraction to the process. You can do it without understanding--indeed, without ever being presented with the argument--how crazy the conclusions are when applied to real life. And you can put those conclusions in an executive summary and no one has time to reverse-engineer the logic that got them there, and so bit by bit all kinds of mathematically-sound but completely crazy conclusions get adopted as reality.

1 comment:

  1. i saw a book in my uni bookstore today, "The Limits of Mathematics: A Course On Information Theory And The Limits Of Formal Reasoning"

    now, it was a math book, so i have no idea what its about, other than it seem to be demonstrating that math is only "guasi-empirical". but i did wonder if it was pondering the question of maps not being territories, whether spreadsheets, computer models or game theory models.