10-09-08, 10:34 am

In his book, The Opinion Makers, David W. Moore devotes his entire 4th chapter, title Inscrutable Elections, to showing how the 'polls often present a highly misleading if not outright false picture of how the candidates are faring and what voters are thinking.' A good example is his report on a media poll conducted in New Hampshire in September 2007 to try and determine how the primary might turn out. There were two versions – one widely reported in the media, the other basically ignored.

In the first poll, the widely reported one, the standard 'if the election was held today' who would you vote for forced choice question yielded an 11 percent undecided total and 43 percent for Clinton to 20 percent for Obama (26 percent for others). But when the poll was given again with the FIRST question asking if people were still undecided about whom to vote for the undecided was 55 percent, Clinton 24 percent, Obama 10 percent and 11 percent other. Even though Clinton eventually won the huge 43 percent early lead was an illusion.

Besides the forced choice question, which leads to unreliable results, there is another tactic pollsters use that leads to misleading results. If you remember, at the beginning of the 2008 race for president, we were told that the polls revealed two 'front runners' – i.e., Clinton and Giuliani. They were never really 'front runners,' at that time, at all. How did the pollsters determine that they were?

Moore says, 'The problem with the apparent front-runner status of both Giuliani and Clinton is that it was based on national polls of Republicans and Democrats, respectively. These are people who don't vote – at least not all in the same primary election, and maybe not at all.'

To get an accurate picture the pollsters would have to run separate polls on a state by state basis in the states with the first primaries. 'Polling a national sample of Democrats and Republicans,' Moore writes, 'reveals nothing about the dynamics of the state-by-state nomination contest.' There are three reasons why this isn't done – it costs too much, people are too undecided, and the results of one state can effect the results in another state. But a nice juicy 'front runner' story is better then nothing from the media's point of view.

The conclusion to be drawn from this chapter is that polls won't tell you how the voters are really thinking (or feeling) about the candidates during most of the election cycle. Within about two weeks of the actual election most people have finally made up their minds so at this time, and only at this time, the 'forced choice' question method will give a 'fairly accurate estimate' of what is going on. 'But,' Moore warns, 'in the weeks and months and years before the election, these polls give a misleading, even false, picture about the electorate.'

What is true about the electorate is also true about polls on public policy such as peace and war, abortion, the environment, etc.

--Thomas Riggins is associate editor of Political Affairs.