Bias befalls inaugural polls

A trio of pre-inaugural polls from Annenberg, Pew and AP-Ipsos are discouraging.

They suggest that, rather than accepting the outcome of the 2004 election, these pollsters are ready for another round of Bush-bashing. They are like the defeated Confederates of old shouting, “Forget? Hell!”A trio of pre-inaugural polls from Annenberg, Pew and AP-Ipsos are discouraging.

They suggest that, rather than accepting the outcome of the 2004 election, these pollsters are ready for another round of Bush-bashing. They are like the defeated Confederates of old shouting, “Forget? Hell!”

I could critique these polls from a partisan perspective, but, more important, these surveys simply don’t meet the criteria we would associate with good research design and methodology.

Let’s look first at AP-Ipsos. The author of this survey had a flair for the dramatic, launching each interview with the blockbuster question: “Do any of these words describe your own feelings as you think ahead to President Bush’s second term in office?” The four words tested were: hopeful, angry, excited and worried.

I suppose the researcher felt there was some balance to the question by having two positive words — hopeful and excited — balanced against two negative words. The problem is that the supposed negative terms — angry and worried — may not necessarily be a negative reflection on the president, as Ipsos seems to think. I may be “angry,” not because Bush was reelected but because I think the Democrats are going to try to thwart his second-term initiatives. I may be “worried” that Democratic-leaning pollsters are going to try and undermine the president’s agenda.

So if the pollsters thought that the 21 percent who described themselves as “angry” or the 47 percent who described themselves as “worried” are all critics of the president, then they should go back to school. Questions must have a universally understood meaning to be valid.

The Pew Research Center poll’s faults revolve around language. For example, one question asked voters about their “priorities for President Bush and Congress this year.” Twenty-three potential priorities were described. The lowest rated one, chosen as a “top priority” by only 27 percent of the sample, was “passing a constitutional amendment that would prohibit gay marriages.” This wording likely elicited a lower rating than would have been recorded for a more general priority such as “protecting traditional marriage.” Similarly, the poll recorded a low priority for “making the recent federal tax cuts permanent.” This formulation likely would have been beaten by “letting taxpayers keep more of what they earn.”

From the other perspective, Democratic priorities are given their best possible send-off. For example, 60 percent assigned a top priority to “providing health insurance to the uninsured.” Do you think 60 percent would have been so enthusiastic about “raising taxes to pay for health insurance for the uninsured?” In short, liberal and Democratic priorities were described in more flattering terms while conservative
and Republican priorities were given less appealing descriptions.

The ugliest portion of the Pew poll was a section designed to identify winners and losers in a second Bush term. For example, the poll asked whether “the military” will “gain influence, lose influence or not be affected by Bush being in office for another term.” Rather than asking coldly of “the military,” Pew could just as well have asked about “soldiers and their families” or “members of the armed services.”

Most of the groups were left-leaning in order to identify and lift up “losers”: blacks, Hispanics, older people, union leaders and poor people. The “winners” were all stereotypically despicable: conservative Christians, Washington lobbyists, corporations and the military. Here are some winners that Pew forgot to ask about: traditional families, small-business owners, taxpayers, faith-based charities and schoolchildren and their families.

The Annenberg survey focused on Social Security but biased the key question on private savings accounts by tying privatization to reductions in benefits and reduced taxes. This was followed by a one-sided question that threatened to “borrow as much as $2 trillion” to finance private accounts.

I’m amazed that we didn’t see a question about when Bush will stop beating his wife.

Hill is director of Hill Research Consultants, a Texas-based firm that has polled for GOP candidates and causes since 1988.