Article contents
Non-Voting and Non-Voters: A Typology
Published online by Cambridge University Press: 07 November 2014
Extract
The studies of non-voting made in the United States and Europe have firmly established that failure to vote is associated with conflicting political preferences, with lack of interest in politics, with low position on the social stratification scale, with a feeling of political powerlessness and, to a lesser extent, with sex (women voting usually less than men) and with age (the very young and the very old voting less than the average). But while identifying the causes of non-voting these studies have been unable to determine with any precision the true level of voluntary non-voting and the relative size of the various groups of non-voters. The reason for this failure is in the non-coincidence of four types of electoral universes used and often confused during the research and the analysis: (a) the universe of the cases studied; (b) the universe of the sample (when the research is done by sampling, as is normally the case); (c) the electors registered on the rolls and thus able to vote; and (d) the potential electors. Unless there is a close coincidence between these four universes, determining the true level of electoral participation and the relative importance of the various types of non-voters becomes an open confrontation between hard data and guess work.
Cette étude est un exercice de mesure de l'abstentionnisme électoral; elle est faite à partir d'un échantillonage aléatoire de 456 noms pris sur les listes électorales d'un quartier de Vancouver avant l'élection d'avril 1963. Soit par interview personnel, soit par renseignements obtenus de voisins ou de parents nous avons établi si la personne dont le nom était apparu dans l'échantillon aurait pu voter ou non; l'accès aux listes de votes nous a permis de déterminer qui, en fait, s'était abstenu.
Le taux d'abstention officiellement enregistré dans le district électoral est de 78 pour cent. Sur les 100 personnes de notre échantillon qui n'ont pas voté, nous trouvons qu'une seule n'existait pas, que trois étaient citoyens américains qui avaient la ferme intention de ne pas voter, que trois étaient morts, trois très gravement malades et quatorze absents de la région de Vancouver pour toute la durée de la campagne. A partir d'interviews obtenues après l'élection nous estimons aussi que 1.7 pour cent des électeurs n'ont pu voter à cause d'accidents ou de maladies survenues entre notre interview pré-électorale et l'élection. Le taux réel de participation, si l'on exclue les électeurs qui ne pouvaient pas voter, passe de 78 à 83.2 pour cent. Cette différence rend sujette à caution les comparaisons de taux d'abstentions officiels, entre régions urbaines et rurales par exemple; la mobilité résidentielle étant, dans la ville que nous étudions, le facteur principal d'abstention forcée.
Les électeurs qui auraient pu mais qui n'ont pas voté nous apparaissent de quatre types principaux. (1) Le boycotteur: il refuse de voter pour des raisons personnelles, religieuses par exemple, mais non pas afin d'influer sur le système politique; il est donc distinct des boycotteurs qu'on trouve communément en Europe, en Amérique du Sud et en général dans les pays ou la légitimité du système électoral est mal établie. (2) L'électeur à la retraite : c'est un ancien électeur qui ne vote plus ; même sous ses aspects démocratiques, la politique reste un système de conflit et de violence qui n'est ni de tous les tempéraments, ni de tous les âges. (3) Le barbare: nouvel immigrant, artiste ou infirmière, la politique ne l'intéresse pas; il n'y connaît rien, n'en veut rien. (4) Le spectateur: il s'intéresse à la politique, en a une connaissance au moins moyenne, a l'intention de voter, mais pour des raisons que nous n'avons pu identifier, il s'abstient. La théorie des appels contradictoires n'est pas apparue comme une bonne explication de l'abstentionnisme parmi les spectateurs. Les boycotteurs représentaient .6 pour cent de l'échantillon; les électeurs à la retraite 2.4 pour cent; les barbares 3 pour cent; les spectateurs, 6.7 pour cent.
L'abstentionnisme volontaire n'est pas apparu plus marqué chez les femmes que chez les hommes; si l'on exclue les électeurs à la retraite, l'abstentionnisme apparaît un peu plus fort chez les hommes que chez les femmes ( 14.0 et 12.5 pour cent respectivement ).
- Type
- Articles
- Information
- Canadian Journal of Economics and Political Science/Revue canadienne de economiques et science politique , Volume 33 , Issue 1 , February 1967 , pp. 75 - 87
- Copyright
- Copyright © Canadian Political Science Association 1967
References
1 For studies of electoral participation see in particular: Merriam, C. E. and Gosnell, H. F., Non-Voting, Causes and Methods of Control (Chicago, 1924)Google Scholar; Arneson, B., “Non-Voting in a Typical Ohio Community,” American Political Science Review, 1925, 816–25Google Scholar; James, K. Pollock. Voting Behavior, A Case Study (Ann Arbor, 1939)Google Scholar; Litchfield, E. H., Voting Behavior in a Metropolitan Community (Ann Arbor, 1941)Google Scholar; Connelly, E. M. and Field, H. H., “The Non-Voter: Who He Is; What He Thinks,” Public Opinion Quarterly, 8 (Summer 1944), 175–87CrossRefGoogle Scholar; Lazarsfeld, P. F., Berelson, B., and Gaudet, H., The People's Choice (New York, 1944)Google Scholar; Miller, J. E., “Atypical Voting Behavior in Philadelphia,” Public Opinion Quarterly, 12, no. 3 (1948–1949), 489–90CrossRefGoogle Scholar; Centers, Richard, The Psychology of Social Classes (Princeton, 1949)Google Scholar; Rosenberg, M., “The Meaning of Politics in Mass Society,” Public Opinion Quarterly, 15 (Spring 1951), 5–15 CrossRefGoogle Scholar; Janowitz, M. and Miller, W. E., “Index of Political Predisposition in the 1948 Election,” Journal of Politics, 14 (1952), 710–27CrossRefGoogle Scholar; Campbell, A., Gurin, G., and Miller, W. E., The Voter Decides (Evanston, 1954)Google Scholar; Hastings, P. K., “The Non-Voter in 1952; A Study of Pittsfield, Massachusetts,” Journal of Psychology, 38 (1954), 301–12CrossRefGoogle Scholar; Berelson, B., Lazarsfeld, P., and McPhee, W., Voting (Chicago, 1954)Google Scholar; Dogan, M. and Narbonne, J., “L'Abstentionisme électoral en France,” Revue française de Science politique, 4 (1954), 5–26 and 301–25CrossRefGoogle Scholar; “La Psychologie politique des femmes,” Sondages, 1954, No. 2 and 3Google Scholar; Goguel, F., ed., Nouvelles études de Sociologie électorale (Paris, 1954)Google Scholar; Benney, Mark, Gray, A. P., and Pear, R. H., How People Vote (London, 1956)Google Scholar; Milne, R. S. and Mackenzie, H. C., Marginal Seat, 1955 (London, 1958)Google Scholar; Duverger, M., La Participation des femmes à la vie politique (Paris, 1955)Google Scholar; Campbell, A. and Stokes, D. E., “Partisan Attitudes and the Presidential Vote,” in Burdick, E. and Brodbeck, A., eds., American Voting Behaviour (Glencoe, 1959)Google Scholar; R. E. Renneker, “Some Psychodynamic Aspects of Voting Behavior,” ibid.; R. S. Milne, “Second Thoughts on ‘Straight Fight,’” ibid.; Birch, A. H. et al. Small Town Politics (London, 1959)Google Scholar; Butler, D. E. and Rose, R., The British General Election of 1959 (London, 1960)Google Scholar; Campbell, A., Converse, P., Miller, W. E., and Stokes, D., The American Voter (New York, 1960)Google Scholar; Glaser, W. A., “Fluctuations in Turnout” in McPhee, W. N. and Glaser, W. A., eds., Public Opinion and Congressional Elections (Glencoe, 1962)Google Scholar; Blondel, J., Voters, Parties and Leaders (London, 1963).Google Scholar
For a review of the literature on political participation and social class, see Erbe, William, “Social Involvement and Political Activity: Replication and Elaboration,” American Sociological Review, 29 (04 1964), 198–215.CrossRefGoogle Scholar
2 A further disadvantage in relying on the respondent's answers is that the social value attached to voting leads to misleading responses. The Lazarsfeld Elmira study had a low rate of misreporting. A comparison of aggregate statistics of official records and survey responses shows misreporting to be about 2 per cent; see Dinerman, H., “The 1948 Voter in the Making” Public Opinion Quarterly, 12, Winter 48–49, p. 598.Google Scholar But this lack of discrepancy seems limited to participants of panel studies, respondents who by the mere fact of accepting a number of reinterviews are probably more interested politically and thus less likely not to vote than the people who would have accepted, out of surprise, to be interviewed once but not twice. In their Ewing Township study, Parry & Crossley found rates of misreporting varying from 14 to 31 per cent between presidential and primary elections. See Parry, Hugh J. & Crossley, Helen M., “Validity of Responses to Survey Questions,” Public Opinion Quarterly, 14 (Spring 1950), 61–80.CrossRefGoogle Scholar In the Waukegan study, Mungo Miller, found misreporting of voting to be at 10 per cent. See Miller, Mungo, “The Waukegan Study of Voter Turnout Prediction,” Public Opinion Quarterly, 16 (1952), 381–98.CrossRefGoogle Scholar In my sample I found the rate of post-election misreporting of the vote to be 10 per cent.
3 See Miller, “The Waukegan Study.”
4 For a study of the effect of registration rules on voting participation in the United States see Report on Registration and Voting Participation, U.S. Government Printing Office, 1963, particularly Appendix III.
5 Each person on the list was assigned a number from 1 to 40,569. (456 numbers were drawn at random.) There is no easy way of checking the reliability of one's sample. The few checks available showed our sample to be good. For example, (a) if we rank the 456 sample names by polling areas (from 1 to 149) we find that the person in the median position is in the same polling division as the person at the middle point of the 40,569 names on the fists; (b) according to the official records of the rate of voting in the district was 78.40 per cent; according to our sample it was 78.01.
6 A private publication published annually, the B.C. Directory lists all residents of the Vancouver area whose names are obtained from public sources and from a door to door canvass.
7 See Report of the Chief Electoral Officer, Twenty-Sixth General Election (Ottawa, 1963).Google Scholar
8 See Scarrow, H. A., “Patterns of Voter Turnout in Canada,” Midwest Journal of Political Science, 5 (1961), 351–64.CrossRefGoogle Scholar
9 Standard error = 3.43 < E < 4.59.
10 See Scarrow, , “Patterns of Voter Turnout in Canada,’ 351 ff.Google Scholar
11 The riding is in a section of the city which excludes the poorest as well as the wealthiest districts. It is composed mostly of white collars but has a sizeable manual workers group. A large number of rest homes and rooming houses and a large hospital account for the population of the district being much older than the national average. Thirty-four per cent of the people in the sample were over 65 (13 per cent for the nation).
12 According to our maximum hypothesis, that which takes minor illness and accident into account, eight more people (100 × 1.7 = 7.6 = 8) should be added to the 24, thus leaving only 68 voluntary abstainers; but the data given by the post-election interviews are not sufficient to enable us to sort out from the whole sample the eight or so people likely to have been affected by a minor illness or accident on the day of the election. To measure the relative size of the various groups of non-voters I must thus work on the basis of the minimum hypothesis; the effect is to bias slightly upward the percentages measuring the size of the different types of non-voters, especially the “retired electors” who are more likely than any other group to be affected by illness or accident. It would be unreasonable to assume, however, that the bias in any of the categories proposed in the typology is more than a fraction of 1 per cent of the total sample.
13 After trying a battery of tests we finally chose to use only one question to separate respondents according to their knowledge of politics; the question: “Could you name the provincial and federal leaders of the following parties: Communist, Liberal, N.D.P., Progressive Conservative, Social Credit?” Two points were given for each correct answer; one point for giving a correct name but for the wrong position (federal listed instead of provincial leader or vice versa). The highest possible score was 20. Barbarians are those respondents who scored less than five; poorly informed spectators are those who scored more than 4 and less than 13, well-informed spectators those who scored more than 12. In the case of the Communist party either “Buck” or “Morris” was taken as correct answers for the national leadership. The cutting point at 12 between poorly and well-informed respondents is, of course, somewhat arbitrary. It was selected for two reasons, (a) We assumed that a well-informed person in British Columbia should have been able to identify Pearson (Liberal); Strachan and Douglas (NDP); Fulton and Diefenbaker (Conservatives) and Bennett and Thompson (Social Credit), which added up to 14. (h) The second reason was that the distribution of scores invited that we put the cutting point between 12 and 13, since 9 per cent of respondents had scored 14, only 1 per cent scored 13, and 8 per cent scored 12.
14 This definition of cross-pressure, like those of Lazarsfeld and Campbell, by equating cross-pressure with divided interests, opinions, and loyalties fails to distinguish pull in the opposite from pull in a different but non-opposite direction. While the first would act as a brake the second might only change the orientation of the vote. If an individual who likes two parties equally has to express his final choice in a single vote, the two pulls would act as a brake on each other; the result could be the standstill of abstention. But if the individual has the possibility of expressing his varied preferences by more than one vote (splitting his ticket or voting for a compromise party) the cross-pressure need not then be a brake and might even increase the tendency to vote.
15 The residual category comprises students, housewives, retired individuals, businessmen and management, shopkeepers, white collar workers, liberal professions. The occupation category was determined by the respondent's answers and by the interviewer's comments. The latter are necessary since many respondents describe their profession in very vague terms.
- 9
- Cited by