¹ There are at least three studies which do not fit this generalization. First, there is a comparative political parties project headed by Allan Kornberg and Joel Smith. Four of the most relevant reports include: (1) Smith, Joel, Kornberg, Allan, and Bromley, David, “Patterns of Early Political Socialization and Adult Party Affiliation,” Canadian Review of Sociology and Anthropology V (August 1968), 123–55Google Scholar; (2) Kornberg, Allan, Smith, Joel, and Bromley, David, “Some Differences in the Political Socialization Patterns of Canadian and American Party Officials: A Preliminary Report,” this journal II, no. 1 (March 1969), 64–88Google Scholar; (3) Kornberg, Allan and Smith, Joel, “Self-concepts of American and Canadian Party Officials,” Polity VIII, no. 1 (1970), 70–99CrossRefGoogle Scholar; and (4) Allan Kornberg, Joel Smith, and Harold D. Clarke, “Semi-Careers in Political Work: The Dilemma of Party Organizations,” Sage Professional Papers in Comparative Politics, series number 01–008, Vol. 1. The second study is that done by Howard Scarrow of a Windsor federal constituency in 1962. See his “Three Dimensions of a Local Political Party”, in Papers on the 1962 Election, ed. Meisel, John (Toronto 1964), 53–67.Google Scholar Finally, there is the study by Samuel Eldersveld on political parties in the Detroit area. See his Political Parties: A Behavioural Analysis (Chicago 1964).

² Political Parties (New York 1962), see especially pages 238–270 and 333–356. This book was first published in 1911.

³ Especially chapters one and three. Eldersveld also quotes de Grazia, Alfred, Roberto Michels’ First Lectures in Political Sociology (Minneapolis 1949).Google Scholar

⁴ For a critical history of the concepts of aggregation and articulation and the problems associated with their operationalization, see Janda, Kenneth, “A Conceptual Framework for the Comparative Analysis of Political Parties,” International Comparative Political Parties Project Reports, no. 6 (Evanston, Ill. 1969), 15–19.Google Scholar

⁵ Ed. Gabriel A. Almond and James S. Coleman (Princeton, N.J. 1960), especially pp. 33–45.

⁶ Ibid., especially Gabriel A. Almond, “Introduction: A Functional Approach to Comparative Politics,” p. 33.

⁷ Almond and many others who have used the concept “aggregation” do not make clear how the concept should be systematically operationalized.

⁸ This minimum variation occurs for 61 per cent of the 44 possibilities in this study.

⁹ In a landmark article dealing with political parties in 17 countries Richard Rose and Derek Urwin use this concept but under the name of “the social cohesion of political parties.” Their analysis, however, deals only with voters and no analysis is made of party activists. See their “Social Cohesion, Political Parties and Strains in Regimes,” Comparative Political Studies II, no. 1 (April 1969), 7–67.

¹⁰ It occurs that one social category gives a political party a half or more of its support 53 per cent of the time out of the 30 possibilities in this study.

¹¹ The exact population in 1966 was 298,121. Dominion Bureau of Statistics, Population Characteristics by Census Tracts, Hamilton, Census of Canada, 1966, catalogue no. 95–610, bulletin c-10, August 1969, 4.

¹² Dominion Bureau of Statistics, Population and Housing Characteristics by Census Tracts, Hamilton, Census of Canada, 1961, catalogue no. 95–523, bulletin ct-8, June 1963, 4, 16.

¹³ The federal constituency of Hamilton-Wentworth was not sampled partly because half its polling divisions are outside the city boundaries of Hamilton.

¹⁴ For a more detailed discussion of the data-collection procedures see Winham, Gilbert and Cunningham, Robert, “Party Leader Images in the 1968 Federal Election,” this journal III, no. 1 (March 1970)Google Scholar, especially 38–9. The authors of the present article thank Winham and Cunningham for making this survey available to us.

¹⁵ This finding is identical to that of Richard Rose and Derek Urwin in that they found that religion, and not class indicators, is the most important variable in understanding the social basis of political parties.

¹⁶ For an in-depth discussion of the meaning and problems of developing natural categories for variables with nominal and ordinal measurements, see Lazarsfeld, Paul and Barton, Allan, “Qualitative Measurement in the Social Sciences: Classification, Typologies and Indices,” in The Policy Sciences, ed. Lerner, Daniel and Lasswell, Harold (Stanford, Calif. 1951), 155–92.Google Scholar

¹⁷ Unfortunately, data on the presence or absence of labour union background were not collected in the voter survey.

¹⁸ The authors recognize that some scholars may not view housewives as an occupational category. While in the past women who were not gainfully employed were viewed as not having a class position apart from their husbands, sociological thinking and research in recent years has dealt more and more with the class interests of housewives as distinct from their husbands.

¹⁹ In some cases particular unions tend to take over particular constituency organizations. Thus, the Hamilton Centre ndp has an organization dominated by the United Electrical Workers and the incumbent ndp member of the Ontario Legislative Assembly is an official of that union. The Hamilton Mountain organization and the Hamilton East executive tend to be dominated by the Steelworkers union. This situation can produce intraparty conflict over allocation of nominations and resources. Thus, in the decision to nominate a federal candidate in 1968 for Hamilton East, the Hamilton Centre Electrical Union group defeated the Steelworker official put up by the Hamilton East provincial executive. The Hamilton Mountain executive is the strongest case of a single union's power. Of the 14 ndp officials, 11 are trade unionists. Nine of the unionists are steelworkers, mainly from Local 1005, the local which represents the workers at the Steel Company of Canada, Hilton Works plant in Hamilton. If one examines the ndp executive committees in Hamilton Centre, Hamilton East, Hamilton West, Hamilton Mountain, Wentworth, and Wentworth North, one finds 53 trade unionists, many of whom are labour union officials, out of 106 executive members. Of these 53 unionists, 31 are members of the United Steelworkers of America. For an in-depth discussion of this subject see Eastham, Frank C., An Analysis of the Relationship Between the New Democratic Party and Organized Labour with Particular Reference to Hamilton, unpublished master's thesis, Department of Political Science, McMaster University, Hamilton, Ontario, 1972, especially 31–44and 60–7.Google Scholar

²⁰ “Ethnicity, Religion and Politics in Canada: A Comparative Analysis of Survey and Census Data,” Quantitative Ecological Analysis in the Social Sciences, ed. Dogan, Mattei and Rokkan, Stein (Cambridge, Mass. 1969), especially 210–14Google Scholar

²¹ The question for federal party identification was: “On the Federal level, do you think of yourself as a (1) Liberal, (2) Conservative, (3) New Democrat, or (4) uncommitted voter?” The question for provincial party identification was: “On the Provincial level, do you think of yourself as a (1) Liberal, (2) Conservative, (3) New Democrat, or (4) uncommitted voter?”

²² The purpose of these four categories is the same as the categories for the other variables, namely to group the respondents in such a way as to explain differential political responses. The “British Isles” category includes, almost entirely, those with English, Scottish, and Welsh fathers. There are no significant political differences among these three groups. Respondents with an Irish background are found in the “North America” category since Irish immigration into the Hamilton area is an overwhelmingly nineteenth-century phenomenon. Once again we did not distinguish between Eastern and Western Europeans since there are no significant political differences between these two groups.

²³ See the previous citation in footnote 20 as well as his People vs Politics (Toronto 1969).