Wednesday, May 06, 2009

Will Hiatt Sea Ice part MMMCMDCCCLXXXIV

update: While my post debates a month old column, it turns out that Will has made an absolutely false claim on a matter of fact accounting
May 3rd. As noted by Steve Benen
George Will was unimpressed.

"I assume the president is talking about the Prius. It's affordable because Toyota sells it at a loss, and it can afford to sell it at a loss because it is selling twice as many gas-guzzling pickup trucks of the sort our president detests. So as an auto executive, he's off to a rocky start."

Actually, the only thing "rocky" here is the quality of Will's analysis.

In reality, Toyota used to sell hybrids at a loss -- in 1997. The industry and consumer trends have changed considerably over the last 12 years, and Toyota started making a profit on each Prius sold way back in 2001.

Indeed, reader R.H. directed me to this item, which noted, "[T]he Nikkei newspaper in Japan estimated just last week that both Honda and Toyota make over $3,000 of profit on each hybrid sold."


OK back to sea Ice.
Can't let it go. Here I object to something written by Eugene Robinson and quoted here.

ROBINSON: What George Will did was cherry-pick a sentence in a report, you know, be very persnickety in the way he parsed his sentences, and end up making it sound as if the report had said the exact opposite of what it actually said. He was persnickety enough that his editors, who also happen to be my editors, felt he didn’t quite cross the line. I thought he did.

Will was *not* "persnickety enough" in two ways. The sentence to which he pointed when his claim was challenged did not correspond to his claim (even ignoring all context) for two reasons.

He said global sea ice coverage was the same, the sentence said similar or slighly below (or something like that). OK so the context is that further two numbers implying it was lower by 500,000 sq km appeared in the document. If the defence of Will is that what he wrote was technically true, although highly misleading, then it can't also be that 500,000 is about equal to zero. One has not made a false claim on a matter of fact if one either picks an unrepresentative data point or replaces "similar" with "the same." However, a claim which is misleading and technically false is just an error and should have been corrected.

The second reason is that Will made his claim in the present tense, but it didn't accurately describe the most recently published data. In this case it doesn't depend on what the definition of "is" is. The word can be interpreted in various ways, but no matter how it is interpreted, Will's claim was false.

If can be interpreted narrowly so that it refers to the most recent observation. In that case Will's claim was false (as immediately noted by his cited source). Or it can be interpreted broadly as referring to an interval of time which is certainly recent, but not the most recent available. In that case his statement is false. Will's claim was true of a brief interval of time which had ended when he made the claim. When the document which he cited was written, that interval was the most recent available, so the claim in the present tense was true. As the document went on to argue, the particular (they were quoting a global warming skeptic denier) was unrepresentative. It was true for a brief interval of time.

To get really geeky, the interval was during the Northern Winter. The facts are that the area covered by sea ice in the Northern hemisphere has decreased and the area covered by sea ice in the Southern hemisphere has increased. In the Northern winter, the area of *sea* covered by ice is partly limited by Canada, Norway and Russia (the expanding ice runs out of sea to cover as there is land instead). This doesn't happen in the Southern hemisphere. This means that the reduction in global sea ice coverage is lower in the Northern winter than in the Northern summer. Picking an observation from the Northern winter is, therefore, a trick. Will fell for the trick. He assumed that "is" was meant broadly, when for the claim to be true, it had to be interpreted narrowly, and, by the time he wrote his column, as referring narrowly to an interval in the past not including the most recent past.

The problem isn't just that Will picked cheries -- it is that he tried to pick cheries but was clumsy and picked leaves instead. His assertions were both misleading and technically false. The Washington Post should have printed a correction.

1 comment:

Anonymous said...

You have an unclosed strike tag at the beginning.