Sunday, January 3, 2010

Rasmussen polls are bogus until election

It's clear pollster Scot Rasmussen tries to use his polls and the results to advance a conservative Republican agenda. Politico has done a story about how Democrats are increasingly discrediting and discounting his work. Yet, Rasmussen rightfully points out that neutral observers have noted that Rasmussen polls have been among the most accurate polls over the past few election cycles.

So that proves that Democrats are just blaming the messenger for the bad results right?

I think these daily Rasmussen polls are bogus. Not all the time, but whenever they need to be.
Rasmussen has a financial incentive to get results that are less favorable to the Democrats than other polls. In today's political environment there isn't any benefit of being an "honest" anything. You really need to pick left or right, Democrat or Republican. That is why CNN has lost out in cable news. They have a liberal bias but they work hard to be "objective" which in the end turns off nearly everybody but some of the "moderates". Most people gravitate toward excellence and projections of power. It's hard to do that with "on this hand, but on the other hand" presentations. People want an emotional energy burst.

So here is Rasmussen who asks questions like, "Who do you think can best decide how to spend money, the American people or the federal government?" and then can give interviews in Conservative outlets and soberly explain the results as an "independent" pollster. It drives visitors to his site, speaking opportunities, etc.

Rasmussen also has a habit of promoting good results for Republicans like "Republicans gain on the generic ballot"! when I didn't remember when Democrats were winning in the first place. That's because he chooses not to release and/or highlight more liberal leaning findings until later on when there is bad news for Democrats.

So could it be fraud, clever marketing, or selective science?

All of the above.

By Rasmussen being the first pollster to publish data pretty much everyday he gets viewers of all political stripes because we just want anything political. Even if we doubt the authenticity. Lately, Gallup has gotten in that business too, so now it is natural to just compare Gallup and Rasmussen. But the difference is in general, Gallup has polls that swing in an unpredictable direction, sort of like public opinion.

I think as the election season comes to a close, Rasmussen adjusts his computer programs to do the job accurately. So that he can be among the best pollsters. Remember a poll today is really only news to the hardcore political types and only for a day or so. If in 2008, Rasmussen had a poll that displayed McCain leading Obama for instance, there's no indepedent verification of this. If his is the only poll to show a particular result that is even better, especially if he can point to the accurate election results from the previous election day.

So if he mixes in his Republican push polls and maybe outright fraud in between elections, with a state of the art polling at election time, Rasmussen can be the talk of Republican circles.
It doesn't need to be completely made up, it can be a shift in the sample size, the demographics, the time of day polled, etc.
It clearly is profitable, and I think possible.

Unless every poll is 100% transparent and independently verified, Rasmussen can do whatever he wants. I judge pollster by whether they produce polls that equally benefit and/or frustate the two parties.

No comments: