Examines the Record in respect of Elec-
tion Forecasts and ventures Opinions on
why those of 1948 were so Sadly Awry.
BEFORE THE VOTERS deposited their ballots on the first Tuesday after the first Monday in November 1948, the pollsters claimed, and outsiders did not deny, that they had had a good record in predicting the results of elections. By stating their errors in averages, they had made this record appear to be much better than it really was. Moreover, they had camouflaged the size of each such error by stating it in terms of what they called "percentage points."
In the case of the Italian general election of April 1948, for example, the polling organization affiliated with the American Institute of Public Opinion claimed that it had an "average error for all parties of 2.6 per cent." But one of the predictions had been that the National Bloc would get 9 per cent of all the votes cast. It actually got 3.8 per cent of the votes. Dr. Gallup, describes this error as amounting to 5.2 "percentage points," but to me, who am not a statistician, the real prediction was that the National Bloc would poll two and one third more votes than it did. Similarly, in 1948 Dr. Gallup prophesied that Henry Wallace would receive 4 per cent of the total vote. Mr. Wallace got 2.4