Strengthening Constitutional Self-Government

No Left Turns

The Tyranny of Statistics

Most people who have studied the matter note that Body Mass Index is a deeply flawed statistic. What sense does it make to claim that Kobe Bryant is overweight?

Why is the number so popular? I suspect it’s because it is easy to calculate. Our bureaucrats and number crunchers need statistics with which to do their magic. Hence it is important to them that there be reliable statistics. BMI is easy to calculate, and, therefore, is very useful. What use will be made of this and other like statistics as health care gets even more regulated than it already is, one can only imagine. The more unified our health care system is, the harder it will be for a doctor to be skeptical of the standards created in Washington. The trouble is those standards are often created for convenience of calculation, rather than because of their accuracy.

The abuse of statistics is an old story, of course. Why did "disparate impact" become the standard in race discrimination cases? Because it’s much easier to compare the proportion of each race in a given population with the percentage of each race in a given profession or workplace than it is actually to study each population, workplace, etc. closely. The more things we ask Washington to do, the more likely it is that imperfect statistical models will become the standard for operation.

Discussions - 11 Comments

Lies, damn lies and statistics.

Yes, those pesky government statistics and definitions can be such tyranny, especially when utilized in deeply flawed environmental studies, health care studies, and the science of natural selection. It is a good thing that the government statistics on corporate bailouts, military applications of land-mines, and gun ownership, using the same actuarial science, are so accurate. The CBO that the Bush administration lambasted as 'not to be trusted' is now the beacon of truth in projecting Obama's healthcare costs. The 'tyranny of statistics' sign flashes on and off in syncopation with republican schemes.

I sometimes wonder if ren is actually trying to parody himself.

It seems to escape out goofy little friends notice that his own party is the party of "corporate bailouts", and that the Evil Republicans opposed them. But mere reality is always less satisfying to the left than their preferred narrative.

Why did "disparate impact" become the standard in race discrimination cases? Because it’s much easier to compare the proportion of each race in a given population with the percentage of each race in a given profession or workplace than it is actually to study each population, workplace, etc. closely.

You're partly right - but you miss the context within which this test arose, namely, litigation (often by private individuals who had limited resources) after overt discrimination had become harder to get away with.


At any rate, here, as in many of your other comments, I agree with much of what you say about bureaucracies, but think that you need to apply it to private bureaucracies as well as public ones. The last time I was categorized according to BMI was at my private health provider, not by the government.

No one wants to admit he or she is overweight, so someone came up with a general way to measure overweightness. The measure is imperfect, so those who are overweight now feel better that the "objective" measure was not so scientific after all.

But those who are overweight remain overweight.

If McCain had opposed bailouts, or the sitting president, we would not have had them and McCain would have won the election.

Frank Warner, I will admit I am overweight and neither know nor care what my BMI is. However, as the article points out, there are many people who are not carrying excess fat whose BMI says that they are. This is a stupid standard.

The BMI ignores waist size which is the most obvious and clear-cut indicator of obesity level and makes no allowance for the relative proportions of bone, muscle and fat in the body of different individuals.

As to the other matter, how the heck is anyone going to find an accurate statistical model for discerning racism or race bias in hiring? As to "knowing it when you see it", you can go stand with Henry Gates and tell us all about it.

But do not talk to me, I can't even spell my own name.

If McCain had opposed bailouts

McCain was never going to oppose bailouts. The left won the election the instant that man won the Republican nominaton. At that point they were in a no-lose situation.

The problem with statistics is that the general population is completely ignorant of them, and so can be easily fooled. Consider the headline, "100% Increase In Drug-resistant TB Cases". Sound scary? Sure, unless you realize that it simply doubled. Still scared? Suppose you found out that out of 100,000 TB cases last year there was only one drug-resistant, and this year there were two. Still scared? Of course not, since that's normal variation. But the headline is technically true, and can be used by politicians (and the media) to buffalo the public, to its detriment. Happens all the time.

Leave a Comment

* denotes a required field
 

No TrackBacks
TrackBack URL: http://nlt.ashbrook.org/movabletype/mt-tb.cgi/14213