Pollster's X

There had been, said the House Campaign Expenditures Committee, "numerous reports that the Gallup poll might have been used to try to influence the outcome of the elections." Last week in Washington, appearing before the committee at his own request, big, grave Dr. George H. Gallup set out to answer them. He admitted that early in the campaign he had adjusted his prediction of the Presidential returns in a way that seemed to favor Tom Dewey, deducting 2% from the Roosevelt plurality actually indicated by the polls. His explanation: he had expected a light vote, which would presumably have benefited the G.O.P. candidate.

Scolded Committee Chairman Clinton Anderson: "These appear to be dangerous adjustments . . . when you have secret polls and checks which come directly from the voters."

Replied Dr. Gallup: "This is an empirical science. We must learn by experience."

Actually, a random sampling of opinion "directly from the voters" is highly untrustworthy. Accurate results can be obtained only by applying an elaborate series of checks and balances concerning types of voters, voting habits, etc. His polls, explained Dr. Gallup, depend on "at least twelve important adjustments for their accuracy." In this complex mathematical calculation, involving such unknown quantities as election-day weather, the only sin of which Pollster Gallup appeared guilty was that he may have used his Xs to bring down Roosevelt's vote and failed to recognize other X factors which would bring it up again. The astonishing accuracy of this revolutionary new method of measuring public opinion was proven once more by the fact that no scientific poll, including Gallup's, erred by more than 1.1% in forecasting the civilian vote for President.

This file is automatically generated by a robot program, so reader's discretion is required.