STABILITY AND CHANGE IN PATTERNS OF FRENCH PARTISANSHIP

CO ONVERSE AND DUPEUX, in their well-known study of politicization in the French and American electorates, suspect that the "flash party" phenomenon in France (the Gaullist R.P.F. in 1947-1951, the Poujadist U.F.F. in 1956, and the Gaullist U.N.R. in 1958) occurs because a considerable portion of the electorate is not anchored to self-reinforcing long-term partisan commitments.' While 75 percent of the American sample identified with a political party, less than half of the 1958 French sample identified with a party or a tendance such as "left" or "right." Noting that the difference in proportions able to classify the father's partisanship revealed an even greater cross-national difference-86 percent of the American sample but only 26 percent of the French sampleConverse and Dupeux compared levels of partisanship in the two samples after controlling for knowledge of father's partisanship. The controlling operation removed the cross-national difference in partisanship to the marginals, i.e. to the difference in the proportion of respondents in each sample who were aware of their father's partisanship. This led the authors to conclude that the difference in