Tuesday, February 03, 2004

I was walking by the post office today and noticed the headline of The USA Today: "Kerry leading Bush in new poll." It reminded me of why I object to the media's overreliance on polls.

Media members like to solemly proclaim, "We don't make news. We just report it." The media makes news every time it makes out an issue or a broadcast. I had chicken wings for dinner yesterday. It was an important part of my day. It was news to me. The local paper didn't deem it worthy of publication. Each news outlet decides what it considers "newsworthy" and what it doesn't. It has to. The amount of news is infinite while the amount of space in a publication or time in a broadcast is limited. It don't begrudge it that, but it should be ingenuous enough to acknowledge this fact.

That's how the media decides what it considers news. But the use of polls is a different kettle of fish. The use of polls not only replaces exploration of real issues with discussion of a horse race. But it's a case of the media INVENTING a story out of nothing. A report on a poll which claims 52% of Americans think the president is not doing a good job doesn't explain why or how he should do things differently.

Then there's the whole question of the numbers. Take the afforementioned USA Today article on their poll. Even the phrasing of this sentence demonstrates something. They had no story before so they commissioned a poll and then reported on it. Isn't this inventing a story?

The text of the article read, "Kerry defeated Bush 53% to 46%, a lead outside the poll's margin of error. North Carolina Sen. John Edwards edged Bush at 49%-48%, a statistical tie. Bush bested former Vermont governor Howard Dean by 7 points and retired Army general Wesley Clark by 3."

Yet in the small print at the bottom of the actual poll numbers, "Results are based on telephone interviews with 1,001 National Adults, aged 18+, conducted January 29-February 1, 2004."

So what people assume is 53% of all American voters would support John Kerry and only 46% George W. Bush. What the poll results actually demonstrate is that 530 Americans would support Sen. Kerry today in race vs Pres. Bush and 460 Americans would support the incumbent.

But "530 individuals support Kerry" doesn't make good copy.

All that this poll proves is that Sen. Kerry is more popular among the 1001 specific Americans chosen for this particular poll. As someone with a math degree and a concentration in statistics, I know all about scientific sampling and all that. But it doesn't change the fundamental fact that people extrapolate too much out of polls.

And even if the poll numbers are accurate in relation to the population as a whole, so what? Knowing that Sen. Kerry is more popular at this very moment tells me nothing about his plans for the country. It tells me nothing about his foreign policy ideas or his character. It tells me nothing about the president's plans for the country either. It tells me nothing about if I should or shouldn't vote for him. The common thread: it tells me nothing. Why should I care about this "fact"? Such numbers are great for oddsmakers but useless to voters. Yet polls are consistently given lots of air time and column inches in lieu of stuff that might inform me.

This is the fundamental flaw of how coverage of politics now mirrors coverage of sports. The focus on personality and entertainment value and fabricated drama. Both sports and politics have their round tables of hyperactive "experts" shouting at each other. The sports' page has standings; the politics' page has its "who's hot and who's not" table, which is an attempt at the same thing. At least we generally give athletes' spouses a break.

This is great for junkies who are more intrigued with the process but of little relevance to anyone else. Besides, at least sports' commentators take pains to emphasize that just because you're ahead in the 1st quarter doesn't mean you're guaranteed to win the game.

Furthermore, as one acquiantance noted "polls force answers on issues people may not care about and they assume the person they ask knows what they're talking about." This is an excellent point.

Let's say 69% of people surveyed think that universal health care is a good thing. It doesn't tell you which of those people vaguely think it's a good idea and which think it's an absolute imperative for the immediate future. Would all of those 69% support universal health care if it meant raising taxes by 5%? By 25%? Would more support universal health care if the actual care remained private and the insurance was universalized? Do all 69% (or even the other 31%) even share the same conception of what universal health care is? If not, do the numbers mean anything at all?

Polls are like Big Macs. They spice things up from time to time but unhealthy if consumed in large quantities.

No comments: