The Gallup Poll and the 1950 Election

Abstract

BEYOND reaffirming the danger with which all prediction in the social sciences is fraught, the 1948 presidential election spurred public opinion researchers to both a reexamination of existing techniques and redoubled effort to develop less fallible methods of prediction. In this article, Dr. Gallup analyzes the lessons which were learned in 1948 and various refinements of method which enabled the American Institute of Public Opinion to predict more accurately the outcome of the Congressional elections of 1950.