Oct 20, 2012

Statistics and manipulation of the ignoramuses (contd.)

[Previously in the series]

There is a commonly employed fallacy known as Hasty Generalization. It is often used to support a position that a person aligns with and as evidence cites rather a viscerally satisfying personal anecdote or two. The proposition that the person argues for may well be true, but not because of the use of this fallacy. In these days of news being delivered by entertainers, it has become commonplace to force-feed conclusions based on such evidence. For example, you might hear that there is an epidemic of people dying through lead poisoning in Orlando. The evidence?  The testimony of Sheila before the city council. Sheila's aunt drank water from a lead pipe and died, her dog fell sick, and her two kids are sick. Sheila is "known" to be honest, and has gotten several awards for her exemplary character.
Can we conclude that Orlando is likely experiencing an epidemic?  If we have no other data that show fatalities, it is impossible to do so.  For one thing, Orlando has about 243,000 residents, and we would know from a lot more if there were such an epidemic.

One might think that such is an uncommon misuse in society today, but I would venture to guess, perhaps using hasty generalization that it is not the case. Levity aside, I would assert that one will find many fall prey to this, though. Certain news agencies purposefully use this to propagate myths and draw spurious conclusions.

For example, Jimmy Kimmel, a late night talk show host and comedian often has skits that show how people lie and make asses of themselves. On camera. One such segment he ran recently involved asking people on the street how they viewed the Presidential debate conducted the previous day. Of course, the catch being that since there was no such debate, people who talk about it are lying and generally clueless.

Here, watch:

Funny, isn't it?

Yes, but that's about all its value. Yet, you will find people drawing conclusions from this. For instance, here is a news contributor Regina Conley who covered this and another skit by Kimmel ominously reaching a conclusion [emphasis mine]:
While Kimmel’s sketches are all in good fun, they point out a serious problem with the national electorate in their level of ignorance about this Presidential election
It makes you wonder, just how many people are going to head into that voting booth in November without having watched a single Presidential debate, read any election articles or informed themselves about any of the issues at stake?
Fair enough, but it also makes you wonder whether Regina considered some of these possibilities about  Kimmel's video:
  • How many people did Kimmel's team interview before they found this set?
  • Were these people playing along, knowing fully well that there was no debate?
  • Were these people paid by Kimmel or his staff to play along?
  • Does the location represent the entire city, let alone the national electorate as Conley seems to indict?
  • Does the "overwhelming number" citing Obama point to a plethora of ignorant people in Obama's corner?
  • Wait, isn't this a ridiculously small-sized sample, and a lopsided survey to even draw any conclusions?
Now, can it still be true that the conclusion is correct?  Yes, of course. But not based on the hasty generalization.

In the weeks preceding the 2008 election, shock-jock Howard Stern conducted a similar interview with some voters in Harlem. Hilarity aside, many rightwingnuts took the video viral citing it as ostensible proof of... well, you can guess what.

Buoyed by its success, he did it again this time. Predictably the rightwingnuts took it viral again. There is a reason polls cost $60,000 to $100,000 to conduct. It requires careful selection of the sample size, the sample set, phrasing of the proper questions, times to conduct the poll, how to eliminate bias, etc. There is a good reason why sites often publish disclaimers such as this is not a scientific poll.

But then such polls may not yield the result we desire, it may not confirm our biases. Therefore, we should let Howard Stern and Kimmel do them. Then insist that while it is not a scientific poll, it is nevertheless true because, er, common sense. Or because science has not solved all unsolved problems.

[Next in the series]

No comments: