Tuesday, June 08, 2010

Lies, Damned Lies, and Statistics


US Representative Lee Terry (R-NE, 2nd District) just e-mailed the results of a poll of his constituents revealing their attitudes towards certain pressing issues of the day. Included was everything from whether we should continue offshore drilling in the wake of the BP disaster to whether we should repeal the recently passed health care reform bill.

In results that should shock no one [for reasons I am about to explain--Ed.], Terry reported results reflecting that an average of over 77% of his constituents think about these issues the same way he does--which, for those of you out there [you ARE out there, aren't you?--Ed.] who read my postings, are diametrically opposed to my views on the same issues. Terry also reported that while a minority of around 20% of his constituents held views similar to mine, as many as 6% of his constituents were undecided on specific issues.

That's what gave me pause. For, you see, this poll's results were not the result of random, door-to-door canvassing of Nebraska 2nd Congressional District voters. These results came from 2nd District constituents ANSWERING an e-mail Terry himself had sent out to "ascertain" the public feeling. This "poll" was designed to elicit the answers it got, as you can be certain that at least 70% of the people who received it were Terry supporters because those voters make up the majority of his e-mail list.

Doubtless several people like me, who've contacted Terry regularly to express our opposition to his stands on particular issues, also got the e-mails, but we are in a preselected minority for several reasons which have nothing to do with the actual makeup and mindset of 2nd District voters in general. It's axiomatic that people tend to respond to and be more involved in things they support than things they oppose. So not only are those of us who oppose Terry's politics a predetermined minority of his potential polling population, we are less likely than Terry supporters to respond to Terry's e-mail poll invitation than are those who agree with him. People want validation, not rejection. That was Terry's motive in creating and distributing his "poll," and that was the motivation of most of the people who responded to it by agreeing with him.

The real problem with the numbers Terry reported, however, comes in the guise of the alleged "undecided" answers. In a "poll" such as this, where people participated by invitation, who is going to respond "undecided"? More likely, such folks just wouldn't respond at all. Terry said he got over 1,100 answers to his invitation to participate, but I do not recall seeing anywhere the total number of invitations he sent in the first place. Given that there are probably close to a million people living in the 2nd District as a whole, and that as many as half of them are not in the uber-urbanized metropolitan Omaha area, and given that many of them probably don't have computers or the time to respond to such e-mailed invitations as Terry's to begin with, it's impossible for Terry to claim that the 1,100 responses he received were any kind of valid scientific sampling.

This "poll" was much less an effort to find out what his 2nd District constituents think than it was a self-congratulatory exercise whereby he and his supporters could pat themselves on the back in mutual admiration. Mark Twain, as usual, was right. There are 3 kinds of data: lies, damned lies, and statistics. Guess which Twain held in the deepest contempt.

No comments: