In recent weeks, polls kept showing solid support for a public insurance option, seeming to breathe new life into its viability as a provision of the health care legislation under way in Congress. In fact, advocates of a public option, from left-leaning groups to pundits to lawmakers, seized on each new number and trumpeted the news across the 24/7 news spectrum of Twitter, TV ads, blogs and headlines.

And while those polls may have bolstered Senator Harry Reid’s decision to include the public option in the merged Senate bill this week, a closer examination shows once again that public opinion on this issue shifts and shimmies depending on how you phrase the question and what you strip away from (or add to) a compound sentence.

In nearly all recent surveys, a majority of Americans simply approve of providing coverage for the uninsured, suggesting that on an altruistic level at least, they believe people deserve health care.

But differences emerge in the details. For example, support for a public health insurance depends on the order of questions, the language and the arguments posed in favor or in opposition.

For example, in a poll that NBC News and The Wall Street Journal released on Tuesday, half the respondents were asked one question about the public option, and half were asked a different one.

Just under 50 percent favored a health care plan administered by the federal government to compete with private insurance companies, while 4 in 10 opposed. But, almost three-fourths said it was important to have a choice between a public plan and a private plan.

Comment: People aren't being stupid when they answer questions in different ways depending on the wording. But we as a society are stupid to place as much stock in polls as we do. Walter Lippmann reminded us a generation ago that polls are just a snapshot of the electorate's opinion. Why should that snapshot be given reverential treatment? That was a good question when he posed it. It's still a good question.