¹ See Benewick, R. J., Birch, A. H., Blumler, J. G. and Ewbank, A., ‘The Floating Voter and the Liberal View of Representation’, Political Studies, XVII (1969), 177–95.CrossRefGoogle Scholar

² See Trenaman, J. and McQuail, D., Television and the Political Image (London: Methuen, 1961).Google Scholar

³ See Parry, H. J. and Crossely, H. M., ‘Validity of Responses to Survey Questions’, Public Opinion Quarterly, XIV (1950), 61–80.CrossRefGoogle Scholar

⁴ See Calahan, D., ‘Correlates of Respondent Accuracy in the Denver Validity Survey’, Public Opinion Quarterly, XXXII (1969), 607–21.Google Scholar

⁵ See Haggard, E. A., Brekstad, A. and Skard, A. G., ‘On the Reliability of the Anamnestic Interview’, Journal of Abnormal and Social Psychology, LXI (1960), 311–18CrossRefGoogle Scholar; Livson, N. and McNeill, D., ‘The Accuracy of the Recalled Age of Menarche’, Human Biology, XXXIV (1962), 218–21Google Scholar; Pyles, M. K., Stolz, H. R. and McFarlane, J. W., ‘The Accuracy of Mothers' Reports on Birth and Developmental Data’, Child Development, VI (1935), 165–76.Google Scholar

⁶ See Baddeley, A. D., The Psychology of Memory (New York: Harper and Row, 1976)Google Scholar; Murdock, B. B., Human Memory: Theory and Data (Washington, D.C.: Erlbaum, 1974).Google Scholar

⁷ All measures of ‘actual’ vote in these and other surveys, e.g. those of Benewick, et al. and of Butler, David and Stokes, Donald (see Political Change in Britain: The Evolution of Electoral Choice, 2nd edn. (London: Macmillan, 1974))Google Scholar are, of necessity, ‘recalled’ votes in the sense that they were obtained just after the event. There is no way of ‘observing’ the behaviour as it occurs.

⁸ See Himmelweit, H. T. and Swift, B., Social and Personality Factors in the Development of Adult Attitudes Towards Self and Society: An Eleven-year Follow-up Study of Adolescent Boys (Report to the Social Science Research Council, London, 1971).Google Scholar

⁹ Of the 450 young men interviewed in 1962, 40 per cent stayed the course over the full twelve-year period. Since we were only interested in complete voting histories, the 1970 questionnaire was sent only to those who until 1966 had returned all the questionnaires (N = 365). Again, in 1974 we contacted only those who had returned the 1970 questionnaire (N = 246). The loss of respondents on each occasion (i.e. generally after an interval of four years) was between 20 and 25 per cent. Although more of those with a grammar school education and those holding middle-class jobs remained in the sample, there remained sufficient diversity of present status, background and vote to make the analysis meaningful. Comparison of those who left with those who stayed showed fewer differences than might be expected by chance and none that related, within social status groups, either to interest in politics or to vote cast.

¹⁰ See Himmelweit, H. T. and Bond, R., Social and Political Attitudes: Voting Stability and Change: A Developmental Study from Adolescence to Age Thirty-three (Report to the Social Science Research Council, London, 1974).Google Scholar

¹¹ We should like to thank David Butler and Donald Stokes, and the SSRC Survey Archive at the University of Essex, for making these data available to us.

¹² The response ‘can't remember’ was treated as an inaccurate recollection. The actual numbers of such responses were very small and were related neither to the vote to be recalled nor to the vote cast at the time of recall.

¹³ See Warrington, E. K. and Silberstein, M., ‘A Questionnaire Technique for Investigating Very Long Term Memory’, Quarterly Journal of Experimental Psychology, XXII (1970), 508–12.CrossRefGoogle Scholar

¹⁴ See Benewick, et al. , ‘The Floating Voter’.Google Scholar

¹⁵ In inspecting this table, it is worth remembering that for any one individual an incorrect recollection consistent with present vote would occur in one instance out of three by chance. There are four possibilities in the recall of a vote (given three parties and the alternative of abstention). Since one possibility is that the individual was correct, there are three ways left of being incorrect.

¹⁶ See Haggard, , Brekstad, and Skard, , ‘On the Reliability of the Anamnestic Interview’Google Scholar, and Pyles, , Stolz, and McFarlane, , ‘The Accuracy of Mothers' Reports on Births and Developmental Data’.Google Scholar

¹⁷ See Hagburg, E. C., ‘Validity of Questionnaire Data: Reported and Observed Attendance in an Adult Education Program’, Public Opinion Quarterly, XXXII (1968), 453–6.CrossRefGoogle Scholar

¹⁸ See Himmelweit, H. T., Katz, M. and Humphreys, P., Shopping for Value: The Voter as Consumer (mimeographed report, 1976).Google Scholar