« Questioning the popularity of fact-checking | Main | The Obama oil spill/approval narrative spreads »

June 24, 2010


Minimum wage workers who work more than 20 hours a week on average might make $217.50/week while minimum wage workers who work less than 20 hours a week on average might make only $72.50/week. Excluding benefits I don't really see a discontinuity in weekly wages vs hours worked but this is the same reasoning given by Mellman.

Nice job, Brendan.

The thesis that 50% is "magic number" is inherantly implausible. Unlike Mellman's example of water temperature, Presidential popularity is not a well-defined number. Mellman himself admits that although the average is currently 46%, Gallup was at 49%. Presumably other polls are in the low 40's. And, each of these has an error margin of several percentage points. Which of these results is supposed to be "magic"?

Furthermore, it seems unlikely to me that Presidential popularity is the best predictor of Congressional elections. I'd think Congressional popularity, general Dem vs. Rep comparisons, and race-by-race comparisons would all be better predictors. They're more directly relevant.

I suspect this "50% rule" is mostly a lazy writer's idea for a column.

Perhaps I'm wrong:

Gallup has found that presidential approval ratings seem to be more predictive of midterm election outcomes than congressional approval is, when the two are at odds. That was the case in 1990, for example, when 26% approved of Congress but 58% approved of President George H.W. Bush, and only eight seats changed party hands


The odd case is 1958, when Ike had 57% or so approval and the Republicans lost 48 seats in the House and 13 in the Senate. Historians blame the recession and Sputnik; I believe they're underestimating the detrimental influence of the Edsel.

The comments to this entry are closed.