--The most obvious: It's more than three years before Republicans start voting for any candidate for president. The notion that 71 percent of the Republicans can predict today whom they might support in 2102 is simply ludicrous.

--As an example, the USA Today/Gallup poll, along with all the other major polls, told us in 2007 that Rudy Giuliani was the dominant front-runner for the Republican presidential nomination. Even as late as the fall of 2007, shortly before the caucus and primary voting period began (with the Iowa Caucuses), polls still reported that Giuliani was the candidate to beat. In the end, he won no primaries and no caucuses, and earned no delegates to the 2008 Republican Convention. The polls misled the country for more than a year with their bogus measurements.

--The current poll on Palin was conducted on one night, Monday, July 6, rather than over a longer three-day period (which means if you weren't home that night, there was no chance Gallup could interview you on a different night).

--Palin's announcement was made just before the July 4^th holidays, which meant that large numbers of people probably had not heard about it by the time of the poll. On July 6, when they conducted their poll, Gallup and USA Today refused to measure how many people had actually known of Palin's resignation. Instead, their interviewers informed respondents of the resignation, and then asked for their immediate opinion. By informing respondents of the resignation, the sample of respondents no longer represented the larger American population.

--Contrary to Rich's interpretation, the poll did not exactly find that a large majority of Republicans would vote for Palin. According to Gallup, the poll found that "if Sarah Palin were to run for president in 2012," 35 percent of currently registered Republicans would be "very likely" to vote for her, 37 percent would be "somewhat likely," and 25 percent would be "not too" or "not at all" likely to vote for her. That does – as Rich inferred – show more than 70 percent who say they would be "very" or "somewhat" likely to support her, but 1) it does not say in what context they would support her, and 2) it includes more than half of that group who at best have only weak support ("somewhat likely" to vote for her), hardly the ringing endorsement Rich implied.

--The context is important. It could be that a majority of Republicans would support Palin only if she were the actual Republican nominee. If so, it would not be surprising that a majority of voters would support their own party's nominee for president. But that hardly means Palin is the first choice of the party's voters. The poll makes no attempt to measure whether the support is for the nomination, or for the general election assuming she gets the nomination.

By conducting a quickie one-night poll, with at best ambiguous findings, and at worst findings that have absolutely no relevance to what will happen in 2012, USA Today and Gallup provided pundits and journalists with fodder for their incessant demands for "news." But that doesn't mean the poll provided the public with any real information. CNN's Wolf Blitzer found the poll useful, because it helped spur a discussion about the possible consequences of Palin's resignation. The Times' Frank Rich found the poll useful for his larger critique of the Republican Party. For the average citizen, however, the poll is just another example of how irrelevant real public opinion is to the polling industry.