1 Leacock, Stephen, “The Great Election in Missinaba Country,” in his Sunshine Sketches of a Little Town (London: John Lane, 1912), 213Google Scholar.

2 We know of only one other attempt at econometric analysis of the 1911 election: Allan, John, “Reciprocity and the Canadian General Election of 1911: A Re-examination of Economic Self-interest in Voting” (unpublished M.A. thesis, Queen's University, 1971)Google Scholar. Allan found that potential gainers and losers from Reciprocity did vote according to economic self-interest. The present study differs from Allan's in the range and definition of variables, in the composition of the sample, and in estimation procedures.

3 See especially MacQuarrie, Heath, “Robert Borden and the Election of 1911,” Canadian Journal of Economics and Political Science 25 (1959), 171–286CrossRefGoogle Scholar.

4 The figure in the text is the net 1908–1911 Conservative gain. The Liberal loss was even smaller, 2.8 points. The asymmetry between Conservative and Liberal swings was made possible by flux in the third-party vote, such as it was.

5 Brandon, Portage la Prairie, Pro vencher, and Winnipeg. Note that all four were on the Canadian Pacific main line.

6 Prince Albert and Saskatoon.

7 Masters, D. C., Reciprocity, 1846–1911, Booklet No. 12 (Ottawa: Canadian Historical Association, 1969), 19Google Scholar.

8 The text of the agreement may be found in Hopkins, J. Castell, The Canadian AnnualReview of Public Affairs 1911 (Toronto: The Annual Review Publishing Company, 1912), 28–30Google Scholar.

9 Skelton, O. D., The Life and Letters of Sir Wilfrid Laurier, Vol. 2 (Toronto: S. B. Gundy, 1921), 370Google Scholar; see also Cuff, Robert D., “The Toronto Eighteen and the Election of 1911,” Ontario History 57 (1965), 169–80Google Scholar.

10 Careless, J. M. S., Canada: A Story of Challenge (Toronto: Macmillan, 1963), 323Google Scholar.

11 Creighton, D. G., Canada's First Century (Toronto: Macmillan, 1970), 124Google Scholar.

12 The classical statements that the National Policy's greatest effect was on property values rather than incomes and that western values were discounted in advance of settlement are Mackintosh, W. A., The Economic Background of Dominion-Provincial Relations (Toronto: McClelland and Stewart, 1964)Google Scholar, and FowkeV, C. V, C., The National Policy and the Wheat Economy (Toronto: University of Toronto Press, 1957Google Scholar). Fora recent formal statement with estimates of magnitude of effect, see Norrie, K. H., “Agricultural Implement Tariffs, the National Policy, and Income Distribution in the Wheat Economy,” Canadian Journal of Economics 7 (1974), 449–62CrossRefGoogle Scholar.

13 The effect of Reciprocity on primary manufacturers, industries primarily engaged in the processing of staple exports and involving little value-added, would have been especially ambiguous. The effect would be contingent on the behaviour of both input and output prices in the sector. Estimation of effects would require a general equilibrium analysis. For such an analysis, see Johnston, R. G. C., Norrie, K. H., and Percy, M. B., “The Impact on Canadian Incomes of the Proposed Reciprocity Agreement of 1911: A General Equilibrium Analysis” (manuscript, University of Alberta, 1980Google Scholar).

14 Fowke, V. C., Canadian Agricultural Policy: The Historical Pattern (Toronto: University of Toronto Press, 1946), 265–69Google Scholar; Ellis, L. E., Reciprocity, 1911 (New Haven: Yale University Press, 1939)Google Scholar, 16ff.

15 Skelton, Laurier, 375. Other commentators emphasizing the sociological factors include Baker, W. M., “A Case Study of Anti-Americanism in English-Speaking Canada: The Election Campaign of 1911,” Canadian Historical Review 51 (1970), 426–49CrossRefGoogle Scholar, and Creighton, D. G., Dominion of the North (Toronto: Macmillan, 1957), 434Google Scholar (compare with Creighton, Canada's First Century, quoted in footnote 11).

16 (Toronto: Canadian National League, 1911).

17 For a summary of recent evidence, see Clarke, Harold D., Jensen, Jane, Due, Lawrence Le and Pammett, Jon H., Political Choice in Canada (Toronto: McGraw-Hill Ryerson, 1979Google Scholar), chap. 4. See also Irvine, William P., “Explaining the Religious Basis of Canadian Partisan Identity: Success on the Third Try,” this JOURNAL 7 (1974), 560–63Google Scholar.

18 Skelton, Laurier, 377. In private correspondence, prominent Liberals commonly emphasized the sectarian campaign. See the letter from Laurier to a supporter in Prince Edward Island, in Skelton, Laurier, 382, and Humphries, Charles W., “Mac Kenzie [sic] King Looks at Two Elections,” Ontario History 56 (1964), 203–06Google Scholar.

19 Ideally, both ethnic and religious variables would appear in a multivariate test. Unfortunately, the inclusion of both would introduce multicollinearity into the regression. In the national data, the correlation between the percentage of British ancestry and the percentage Roman Catholic is −0.86. A correlation of this magnitude between two independent variables would undermine the estimation of the regression coefficient on each variable. In an earlier version of this article, we sought to unpack the collinearity of ethnicity and religion with ridge regression. Although the ridge specification did make a difference in some of the tests, we ultimately decided the effort was not warranted. The reduction in variance of coefficients came at some risk of bias in the estimates of the same coefficients, although by Wallace's noncentral F test that risk did not appear too great. More compelling was our concern for the interpretability of the results. Further, we have no particular stake in distinguishing the religious and ethnic variables from each other. We see the social and ideological processes generating each effect as essentially similar. Our real concern is to pit the best specification of a sociological effect against the economic variables. With the percentage Roman Catholic as the sole sociological variable in the regression, the most robust correlation between independent variables, between the percentage in agriculture and the percentage in manufacturing, is −0.22, no threat to the estimated effect of either variable.

20 Dafoe, J. W., Laurier: A Study in Canadian Politics, Carleton Library No. 3 (Toronto: McClelland and Stewart, 1963), 44CrossRefGoogle Scholar.

21 A good capsule account of the charges against the government can be found in MacQuarrie, “Robert Borden,” 283.

22 Creighton, , Canada's First Century 125; John English, The Decline of Politics: The Conservatives and the Party System 1901–20 (Toronto: University of Toronto Press, 1977), 59–60Google Scholar.

23 Stevens, Paul, “Laurier, Aylesworth, and the Decline of the Liberal Party in Ontario,” Historical Papers, Canadian Historical Association, 1968, 94–113Google Scholar.

24 Cuff, Robert “The Conservative Party Machine and the Election of 1911 in Ontario,” Ontario History 57 (1965), 149–56Google Scholar; English, The Decline of Politics, 61–64.

25 Cuff, “The Conservative Party Machine,” 156.

26 We could as easily have examined the Liberal vote. Indeed, we have performed such an analysis. The findings of that analysis simply mirror the findings here and so we have omitted them for reasons of space. A copy of the Liberal analysis is available from the authors on request.

The vote and religious variables are from Donald E. Blake, Canadian Census and Election Data 1908–1968 (Vancouver: University of British Columbia, Data Library, n.d.), File 1: 1908 and 1911 Elections, 1911 Census. For a detailed description of Blake's data sources, see his “Introduction and Data Description.” Neither Professor Blake nor the UBC Data Library bear any responsibility for analyses in this paper.

Note that our analysis includes only 200 of the 217 constituencies. We excluded from analysis the 15 constituencies decided by acclamation in either 1908 or 1911. Acclamations commonly reflect the determinants of competition in the constituency, but are atone remove from actual voting. It is the latter which principally interests us here. We were especially concerned to purge acclaimed seats from the ordinary least squares regressions (OLS) below, as OLS is very sensitive to outliers. Further, inclusion of acclaimed seats would have introduced a significant nonlinearity into the estimation. The 15 acclamations were mainly in Ontario and Quebec. The other two exclusions were St. John City and St. John City and County, in New Brunswick. The former is nested within the latter and so the two observations are not independent. The route of least resistance seemed to be to delete both.

Some will object that our analysis will embody the ecological fallacy, in that we shall test propositions about individuals with data from aggregates. The fallacy does not concern us much. By adopting the constituency as the unit of analysis, we in effect group individual voters by geographical proximity. Such a grouping scheme does not bias estimates of b, however much it may inflate estimates of R². Aggregation by proximity probably reduces the efficiency of the estimates of b, although not differentially between variables. See Blalock, Hubert M., Causal Inferences in Nonexperimental Research (Chapel Hill: University of North Carolina Press, 1964), 97–114Google Scholar, and Hannan, Michael T., Aggregation and Disaggregation in Sociology (Lexington: Heath, 1971), 38–54Google Scholar.

27 Data for the calculation of the Nurseries and Orchards variable and the number of occupiers in agriculture are taken from Census of Canada, 1911, Vol. 4, “Agriculture” (Ottawa: King's Printer, 1914)Google Scholar, Table II. The numberof employees in manufacturing is taken from Census of Canada, 1911, Vol. 3, “Manufactures” (Ottawa: King's Printer, 1913Google Scholar), Table X. The numberof occupiers in agriculture and of employees in manufacturing were standardized by the electoral district population given in the Blake file.

28 For which see Johnston, Norrie, and Percy,“The Impact on Canadian Incomes of the Proposed Reciprocity Treaty.”

29 But see Cuff, “The Conservative Party Machine.”

30 For illuminating recent discussions of this matter see Sears, David O., Hensler, Carl P., AND Speer, Leslie K., “Whites' Opposition to ‘;Busing’: Self Interest or Symbolic Politics?” American Political Science Review 73 (1979), 369–84CrossRefGoogle Scholar; and Weatherford, M. Stephen, “Economic Conditions and Electoral Outcomes: Class Differences in Political Response to Recession,” American Journal of Political Science 22 (1978), 917–38CrossRefGoogle Scholar.