Perhaps all polls on public questions will be looked upon with distrust by some people until they believe firmly that these surveys are above suspicion and that they are made as scientifically as up-to-date knowledge will permit. Should people be assured too that the polling organizations are not linked up with special interest groups? Would it be a service to the public and to the polls if regular checkups of the results were made by some unbiased and expert group? Could the polls be expected to state that their methods have been passed upon by unbiased investigators?

Careful examination of results are already made of figures gathered by some major polls. These “audits” are somewhat like those made of bookkeeping records or accounts of business concerns and government offices. Similarly, audits could be made of all polls on public questions, perhaps under the guidance of expert statisticians who are not connected with the polls. And the techniques of sampling, of getting closer and closer to a cross section of national thought should be constantly perfected.

Actually any poll which turns out to be wrong or misleading may lose its standing with the public. The very existence of a poll depends on its record of accuracy. The Literary Digest folded after its grossly erroneous prediction of the 1936 elections. Would a careful and unbiased checkup of results be a forward step in the development of public opinion polling?

From EM 4: Are Opinion Polls Useful? (1946)