Lies, Damn Lies and Statistics


problsky1“A pollster’s job is to put voters in a box. Make them choose a side – see how they would respond to a neighbor or co-worker’s words like ‘we need to give someone else a chance’ or we need more checks and balances or ‘we need to back the president.”

When we read this steaming pile of horsepucky from pollster Adam Probolsky, the guy who does the polling for Capitol Weekly, Calbuzz decided we couldn’t put much stock in their new survey announcing that Tom Campbell is leading the GOP race for governor.

Since at least 50% of the principals at Calbuzz have been professional pollsters, we figure we can offer at least a half-assed comment, which would be one-half less asinine than some of the questions in the Capitol Weekly poll, to wit:

“Governor Schwarzenegger has said he will seek to close the state budget gap by borrowing or taking money from California’s local governments like cities, counties and water districts. However, the state is unable to perform any such cash grab from funds that are generated from local taxes and fees. Would you be more likely or less likely to support a new tax or fee, tax increase or fee increase by your city, county or water district knowing that these funds would be shielded from the state and stay local?”


“Thinking about the job that U. S. Senator Barbara Boxer is doing, would you say she deserves to be re-elected or do you think it’s time to give someone else a chance?”’

WTF?? “Taking money from California’s local governments,” “cash grab,” “shielded from the state” – what the hell kind of survey uses this kind of loaded language? And the Boxer question confuses job approval with re-elect and then casts her re-election as fundamentally unfair because it wouldn’t “give someone else a chance.” C’mon.

“The questions are clearly biased questions asked in an unbiased way. This is a critical point. We want people to react, to respond. Not prompt them how to respond, but to respond,” Probolsky said in John Howard’s Capitol Weekly story on their own poll.

This is utter bushwah. Not only were the questions loaded, but there’s no evidence in the report of how the sample was drawn or how it was weighted except that we could see the final results have too few voters from Los Angeles and too many from the Central Valley.

The ballot designations for candidates seemed to have been drawn from a hat – every elected official is given a title except Insurance Commissioner Steve Poizner who is called entrepreneur/educator. The question order is bizarre, with Nancy Pelosi’s approval and Barbara Boxer’s re-elect at the top – both federal questions – followed by questions about the governor’s race.

Give us a break.

If you want to study the “poll” for yourself you can find it here.

subscribe to comments RSS

There is one comment for this post

  1. avatar Adam D. Probolsky says:

    Our research is based on real world politics and policy, not an exercise in academics.

    For instance, when local government is looking to pass a revenue measure, the campaign for the tax in fact uses language like “cash grab” and “shield from the state” as part of their message. Anyone who practically applies opinion research on a regular basis would recognize this detail.

    The Boxer question is designed to flush out the base support for her re-election, clearly she has a majority of Democrats, but huge GOP opposition. There is nothing wrong with this question.

    The methodology for the poll is clearly stated in the full report for all to see.

    Given that the GOP candidate field has settled out, we asked each of the campaigns to provide a realistic ballot title. Poizner’s campaign was the only one that provided one. Since “Former eBay CEO” and “Former Congressman” are not allowed on the ballot, we chose titles that would be.

    Since the Democratic field is still in flux, we continue to allow words like former in the titles.

    It should not be hard to believe that Campbell is leading in the polls. Especially since older voters are more likely to choose him, they remember him on their statewide ballot several times before. For all the insider buzz and recent media blitz’ Whitman and Poizer just don’t have the same name ID as Campbell.

    If I was such a biased right-winger why would I publish a poll showing moderate Campbell leading. And why would I show that Democrats lead in the generic Congressional ballot test.

    Truth is, my firm’s research practice is note exclusively partisan. We conduct research for government, corporate, union and political clients. We poll for Republican and Democratic (in non-partisan races) candidates.

    Our partnership with Capitol Weekly allows us to ask questions that other statewide polls do not ask, publish results that speak directly to the issues facing voters and policy makers.

Please, feel free to post your own comment

You must be logged in to post a comment.