Been reading the latest issue of The Weekly Standard, and in particular an editorial smugly commenting on the collapse of the Left (they have a point, but sadly write about the Left in a way that could just as easily be turned around on the Right – as they seemingly always do). Aside from my political biases, I have a major problem with the editorial relating to the “evidence” the writer provides to support his claim that Obama’s administration is being ‘refudiated’ (yuck, don’t get me started on that…) by the voters:
“In a recent poll of likely voters, 54 percent strongly or somewhat supported the Tea Party movement, with 38 percent strongly or somewhat opposed. In another poll, of all adults, the movement did have a slightly negative image (30 percent positive, 34 negative)— but it was considerably better than the image of either the Democratic or Republican party.”
In a recent poll conducted by who? Other than “likely voters”, we know nothing. This is a common practice in journalism the Western World over (I can’t speak for the rest of the Globe) – the use of unattributed “polls”. It’s useless, and it still irritates me that publications can get away with it. A poll of ‘likely voters’ at a Head Shop would likely suggest overwhelming support for the legalisation of marijuana, but the same poll at an AA or temperance meeting would have the opposite effect. By not stating where the poll was conducted, or by whom, however, you now have support for both sides of the debate, and can give yourself a smug pat on the back. This is another example of the real world having nothing to do with academia – when I did my MA in International Journalism, we were specifically told that this sort of statement was useless and should be avoided at all costs.
There is no reason one shouldn’t suspect that the 54% of ‘likely voters’ in the poll above comprised the colleagues and family members of the author. The 38% could well have been ‘conducted’ by watching a couple of news programs on liberal news channels. Voila! A poll conducted, indicating support for my beliefs! This likely isn’t what happened, but by not properly attributing the poll, or giving us more information, they have presented ‘evidence’ for nothing save for their ability to pick and choose their data/poll sources.