Tuesday, May 29, 2007

When big numbers mean nothing.

One annoying aspect of modern life is the use of statistics to drive public opinion. Please, someone help me see why this example is not a particularly bad one:

100,000 Chinese die annually from passive smoking.

A study says that 100,000 people a year die in China from passive smoking. I don't know how they came up with that figure, but I am pretty sure it is meaningless.

China has about 1.3 billion people. According to the CIA World Factbook, China's death rate is about 6.9 per 1000. That means that about 9 million people die each year. What is the level of confidence of this figure? I don't know, but I'd be surprised if the figure is more precise than + or - 100,000 deaths.

So the 100,000 figure probably falls within the range of error. If so, it can't be used to make any meaningful statement.

Yet it is front page news on Yahoo!

Another sobering thought: it is estimated that more than 99% of households who have suffered a death this past year possessed a refrigerator*.

No question that all that smoking in China is unhealthy, but using a such a level of precision to support a policy change is plain misrepresentation. And it happens all the time. The global warming hysteria is another late example. Government inflation figures are just as misleading. Is it no wonder that people distrust the authorities?

Yes, there are too many dangers to comprehend. The biggest one, though, is almost statistically certain**: 90% of statistics used by policy makers are lies.

* Source: personal observation, every household I've been in in the past 30 years owned a refrigerator. (In other words, I made this up).

** This is my personal estimate based upon a tightly focused study(in other words, I personally took a sample of 1 statistic and extrapolated from there. . . .)


Lauren said...

We expect poor statistical applications in advertising, social engineering directives on the part of government, and the promulgation of false theories too often known as "science." What has specifically annoyed me the most, however, is misleading--or really just ignorant or bad--uses of statistics by so many evangelists.

Do they think their big numbers will give an aura of intelligence to their cause? Are they afraid that otherwise people will think that God is stupid, or that they are? Are they really just afraid to be fools for Christ?

Ruben said...

I love your methods. And I have begun to cast sidelong glances at my refridgerator.