How to make a political scientist cry -- Scott Rasmussen's "Hillary Meter"

Here's the New York Post on pollster Scott Rasmussen's effort to put the pseudo-science in political science:

One of the nation's top pollsters has created a new "Hillary Meter" to measure Sen. Clinton's move to the political center for a 2008 White House run - it shows she's made progress but has a long way to go. "Right now, she's too far away from the center to win unless Republicans nominate someone who is even further away," said independent pollster Scott Rasmussen.

...Clinton is now 52 points to the left of the political center -- a bit closer to center than in January, when she was 59 points to the left. Rasmussen said he created the meter because of his "fascination" with Clinton as "the only defining actor" for 2008.

So I went and looked up how the "Hillary meter" is created. Get ready for this nonsense:

Today's Hillary Meter shows that perceptions of the New York Senator have moved four points to the left over the past two weeks.

Forty-seven percent (47%) of Americans now believe New York Senator Hillary Clinton is politically liberal. That's up from 43% two weeks ago, but down from 51% in late January.

Thirty-two percent (32%) of Americans view the former First Lady as a moderate while 8% believe she is politically conservative.

Collectively, these numbers place Senator Clinton a net 56 points to the left of the nation's political center. Two weeks ago, she was 52 points left of center.

The political center is calculated by subtracting the number of liberals from the number of conservatives among the general public (35% conservative, 18% liberal for a net +17). For the Senator, 8% conservative minus 47% liberal equals a net -39. The minus 39 reading for Senator Clinton is 56 points away from the plus 17 reading for the general public.

This is the dumbest methodology I've ever heard of. For one thing, it's well known that many Americans describe themselves as conservatives but are "operationally liberal" when it comes to their specific policy stances. So the net liberal/conservative differential doesn't tell us very much about "the center" or where a presidential candidate should be located on a left-right scale. And it's necessarily a lagging indicator -- people aren't paying close attention to what she's doing right now. Plus he's interpreting minor fluctuations in the polls as relevant that could just be survey error.

Here's a better measure of how liberal Hillary is: her location on a left-right scale compared to other senators in the 108th Congress. This takes into account all the votes she's cast:

It turns out she was the 12th most liberal senator -- almost a standard deviation to the left of the Democratic mean. If she doesn't reposition herself on policy, she'll probably lose in 2008 (though a bad economy could put almost any Democrat over the top). But the first step isn't winning over the American people in Rasmussen's poll; it's winning over elites in the media. Her husband convinced them that he was a different kind of Democrat, and then was able to make a similar case to the voters in 1992, in large part because the media reinforced his message. Can she do it? I doubt it. Perceptions of her are so entrenched that it would take the continued slaughter of liberal sacred cows from now until Election Day.

Comments

[Jami]

Thank you so much! I knew there was something fishy about Rasmussen, but I didn't think I could withstand looking into it myself.

Even the voting record, though, could be misleading. given the crapola the Republicans are presenting congress these days, 12th-most-liberal could be interpreted as 12th-least-crazy, 12th-sanest, or 88th-least-insane.