Tuesday, October 19, 2010

How to lie with statistics, once again

David Brooks in his column in the New York Times today:
"The vast majority of campaign spending is done by candidates and political parties. Over the past year, the Democrats, most of whom are incumbents, have been raising and spending far more than the Republicans.

And from the Center for Responsive Politics overview of raising and spending (a source that Brooks cites):
Democratic House candidates have raised $444 million, while Republican candidates have raised $473 million (rounded to the nearest million). Democratic Senate candidates have raised
$229 million, while Republican candidates have raised $238 million.

I don't know in what world $444 million is "far more" than $473 million, but apparently that's where Brooks lives.

David Brooks again:
"According to the Wesleyan Media Project, between Sept. 1 and Oct. 7, Democrats running for the House and the Senate spent $1.50 on advertising for every $1 spent by Republicans. Despite this financial advantage, Democrats have been sinking in the polls."

And from the Wesleyan Media Project's press release that is the source of this data:
"Candidates in federal races have spent roughly $130M and Democrats have a 1.5:1 advantage in that spending. However, between 9/1 and 10/7, almost $65M has been spent by interest groups in key federal and gubernatorial races. “Breaking down the air war by party reveals big advantages to Republicans in both party and interest group investment in federal races,” said Franz. “Combining party and coordinated totals, Republicans are outspending Democrats by almost 3 to 2. Among interest group spenders, Republican-leaning organizations are outspending Democrats by a margin of almost 9:1 in House and Senate contests.”

Brooks argues in the remainder of the column that outside spending doesn't matter (although the best he can offer is some anecdotes of questionable relevance), but even so his highly selective use of data from the Wesleyan Media Project strikes me as disingenuous, to say the least.

More generally, it may well be true that Democratic party candidates are doing better at raising funds in some key races than the general public believes at the moment (and Brooks cites some data that suggests this is true). But if money is as unimportant to election outcomes (or to the behaviour of those elected afterwards) as Brooks suggests, then why does he feel it necessary to distort the evidence as he does?

Update: some of the other "evidence" in Brooks' column is also highly misleading (or worse), it turns out, as Glenn Greenwald discusses at Salon.com.

No comments:

Post a Comment