Poll Data--Again

We have warned our readers in past blogs (Poll Data) to be wary of figures reported to be the results of a public poll until you know how the questions were asked. In USA Today the paper reports the results of a recent poll with the following "table."

"Who wins? What group of Americans will get the most benefits from the tax-cut bill, if one is signed into law?
The poor: 6%
The Middle Class: 17%
The Wealthy: 64% "

It would be helpful to know what dollar value was used (or if any was used) to define the meaning of "the poor" or the other two categories. The paper also reports three "volunteered" categories of Everybody, Nobody, and Undecided/Refused, but do not tell readers how the "volunteered" responses were obtained (were they responses from those who did not answer in terms of the three established categories, or were they in addition to established category responses). This kind of table and this kind of "data" reporting would never make it through an introductory research methods course in sociology, but USA Today thinks it is fine for their readers.