¹ Key, V. O., The Responsible Electorate (Cambridge, Mass.: Belknap Press, 1966)CrossRefGoogle Scholar; Boyd, Richard W., “Popular Control of Public Policy: ANormal Vote Analysis of the 1968 Election,” American Political Science Review, 66 (June 1972), 429–49CrossRefGoogle Scholar; Fishbein, Martin and Coombs, Fred S., “Basis for Decision: An Attitudinal Approach Toward an Understanding of Voting Behavior” (paper delivered at Annual Meeting of the American Political Science Association, Chicago, 1971)Google Scholar; Pomper, Gerald M., “From Confusion to Clarity: Issues and American Voters, 1956–1968,” American Political Science Review, 66 (June 1972), 415–28CrossRefGoogle Scholar; Schulman, Mark A. and Pomper, Gerald M., “Variability in Electoral Behavior: Longitudinal Perspectives from Causal Modeling,” American Journal of Political Science, 19 (February 1975), 1–18CrossRefGoogle Scholar; Jackson, John E., “Issues, Party Choices, and Presidential Votes,” American Journal of Political Science, 19 (May 1975), 161–85CrossRefGoogle Scholar; Miller, Arthur H. et al., “Majority Party in Disarray: Policy Polarization in the 1972 Election,” American Political Science Review, 70 (September 1976), 753–78CrossRefGoogle Scholar.

² Lazarsfeld, Paul F., Berelson, Bernard R., and Gaudet, Hazel, The People's Choice, 3rd ed. (New York: Columbia University Press, 1968)Google Scholar; Berelson, Bernard R., Lazarsfeld, Paul F., and McPhee, William N., Voting (Chicago: University of Chicago Press, 1954)Google Scholar; Campbell, Angus, Gurin, Gerald, and Miller, Warren E., The Voter Decides (Evanston, Illinois: Row, Peterson, 1954)Google Scholar; Campbell, Anguset al., The American Voter(New York: John Wiley, 1960)Google Scholar; Converse, Philip E., “The Nature of Belief Systems in Mass Publics,” in Ideology and Discontent, ed. Apter, David E. (Glencoe, I11.: Free Press, 1964), pp. 206–61Google Scholar; and Campbell, Anguset al., Elections and the Political Order (New York: John Wiley, 1966)Google Scholar. A recent critique of the issue-voting findings is Margolis, Michael, “From Confusion to Confusion: Issues and the American Voter (1956–1972),” American Political Science Review, 71 (March 1977), 31–43Google Scholar.

³ The latest statement on the responsible parties controversy, with copious references to previous writings on the subject, is Kirkpatrick, Evron M., “‘Toward a More Responsible Two-Party System’: Political Science, Policy Science, or Pseudo-Science?” American Political Science Review, 65 (December 1971), pp. 965–90CrossRefGoogle Scholar.

⁴ See Benney, Mark, Gray, A. P., and Pear, R. H., How People Vote (London: Roudedge and Kegan Paul, 1956), pp. 140–46Google Scholar; Milne, R. S. and Mackenzie, H. C., Marginal Seat, 1955 (London: Hansard Society, 1958), pp. 117–21, 126Google Scholar; Abrams, Mark and Rose, Richard, Must Labour Lose? (Harmondsworth, Middlesex: Penguin Books, 1960), pp. 19–20Google Scholar; Blondel, Jean, Voters, Parties, and Leaders, rev. ed. (Baltimore: Penguin Books, 1966), pp. 77–87Google Scholar; Pulzer, Peter G. J., Political Representation and Elections (New York: Praeger, 1967), pp. 114–15Google Scholar; and Benewick, R. J.et al., “The Floating Voter and the Liberal View of Representation,” Political Studies, 27 (June 1969), 191–92, 194Google Scholar.

⁵ Butler, David and Stokes, Donald,Political Change in Britain, 2nd ed. (New York: St. Martin's Press, 1974), pp. 296–302; 382–83,414–16CrossRefGoogle Scholar.

⁶ Foot, Paul, Immigration and Race in British Politics (Harmondsworth, Middlesex: Penguin Books, 1965)Google Scholar; Deakin, Nicholas, ed., Colour and the British Electorate, 1964 (London: Pall Mall Press, 1965)Google Scholar; Butler, David and King, Anthony, The British General Election of 1964 (New York: St. Martin's Press, 1965), pp. 224–25, 296CrossRefGoogle Scholar; and Butler, and Stokes, , Political Change in Britain, pp. 304–05Google Scholar.

⁷ Deakin, N. D.et al., “Colour and the 1966 General Election,” Race, 8 (July 1966), 17–42CrossRefGoogle Scholar; Butler, David and King, Anthony, The British General Election of 1966 (London: Macmillan, 1966), pp. 250–53, 268–69, 289CrossRefGoogle Scholar; Nicholas Deakin, “British Voters and the Immigration Issue,” Appendix VII in Patterson, Sheila, Immigration and Race Relations in Britain, 1960–67 (London: Oxford University Press, 1969), pp. 417–22Google Scholar.

⁸ Butler, and Stokes, , Political Change in Britain, pp. 303–05Google Scholar.

⁹ Deakin, Nicholas and Bourne, Jenny, “Powell, the Minorities, and the 1970 Election,” Political Quarterly, 41 (October-December 1970), 399–415CrossRefGoogle Scholar; Butler, David and Pinto-Duschinsky, Michael, The British General Election of 1970 (London: Macmillan, 1970), pp. 341, 405–07Google Scholar; Hoinville, Gerald and Jowell, Roger, “What Happened in the Election?” New Society, 16 (2 July 1970), 14Google Scholar; Crewe, Ivor and Payne, Clive, “Another Game with Nature: An Ecological Regression Model of the British Two-Party Vote Ratio in 1970,” British Journal of Political Science, 6 (January 1976), 68–73CrossRefGoogle Scholar.

¹⁰ See Crossman, Richard, “Understanding the Profusion of Shrinking Violets,” The Times, 6 September 1972Google Scholar; Wood, John, ed., Powell and the 1970 Election (Kingswood, Surrey: Elliot Right Way Books,1970)Google Scholar; and Butler, and Stokes, , Political Change in Britain, pp. 305–08, 415Google Scholar. The post-electionanalyses offered by Louis Harris Research and National Opinion Polls in The Polls and the 1970 Election, ed. Rose, Richard (Glasgow: University of Strathclyde Survey Research Centre Occasional Paper No. 7, 1970), pp. 24–25, 39–40Google Scholar, indicate that the combination of the immigration issue and Powell's personality significantly influenced voting behavior. A similar view is expressed in Johnson, R. W. and Schoen, Douglas, “The ‘Powell Effect’: or How One Man Can Win,” New Society, 37 (22 July 1976), 168–72Google Scholar.

¹¹ Butler, and Stokes, , Political Change in Britain, pp. 306–08Google Scholar.

¹² Preliminary analyses of the two British general elections in 1974 indicate that immigration was not an important factor in the outcome. See Fox, A. D., “Attitudes to Immigration: A Comparison of Data from the 1970 and 1974 General Election Surveys,” New Community, 4 (Summer 1975), 167Google Scholar; Butler, David and Kavanagh, Dennis, The British General Election of February, 1974 (New York: St. Martin's Press, 1975), pp. 24, 332CrossRefGoogle Scholar; Community Relations Commission, Reference and Technical Services Division, Participation of Ethnic Minorities in the October General Election, 1974 (London: Community Relations Commission, 1975), pp. 33–43Google Scholar.

¹³ Because of the large number of variables employed and the fact that a few questions were asked of only half-samples, the working N for much of the analysis is considerably smaller than this. A check of the major variables of interest in this analysis did not indicate any significant bias in the results because of this.

¹⁴ Details on the measurement and selection of all background variables–individual socioeconomic, social context, and political context–are available in Studlar, Donley T., “The Impact of the Colored Immigration Issue on British Electoral Politics, 1964–1970” (Ph.D. dissertation, Indiana University, 1975)Google Scholar.

¹⁵ Butler, David and Stokes, Donald, Political Change in Britain, (New York: St. Martin's Press, 1969), pp. 351–53Google Scholar; Deakin, , “British Voters and the Immigration Issue,” pp. 418–19Google Scholar; Bagley, Christopher, Social Structure and Prejudice in Five English Boroughs (London: Institute of Race Relations, 1970), p. xGoogle Scholar; Deakin, , Colour and the British Electorate, 1964, pp. 163–66Google Scholar.

¹⁶ Britain, Great, General Register Office, Census 1966: United Kingdom General and Parliamentary Constituency Tables (London: Her Majesty's Stationery Office, 1969), pp. 79–88Google Scholar.

¹⁷ Attempting to analyze votes for other parties, even the Liberals, presents a problem because of the low N in those categories and the multiple controls employed in this analysis.

¹⁸ Bivariate correlation tests for the effect of opinion were run with the abstainers and Labour voters together in one category of the dependent variable and the Conservatives in the other. The results did not differ appreciably from the same analysis with the abstainers removed. When one compares the two-party distribution of the vote fa the Butler-Stokes surveys with the actual recorded distributions for the two-party vote in England for the 1964, 1966, and 1970 election, he finds that the respondents in the Butler-Stokes surveys tend toward Labour in reported voting behavior slightly more than the population at large. This discrepancy is not large, however, and except for 1966 is within the normal limits of sampling error. The gap between the surveyed and actual vote was 2.7 percent in 1964, 4.1 percent in 1966, and 0.9 percent in 1970. Rose, Richard, “Britain: Simple Abstractions and Complex Realities” in Electoral Behavior: A Comparative Handbook, ed. Rose, Richard (New York: Free Press, 1974), p. 495Google Scholar, points out that the overall Butler-Stokes errors in estimating the gap for Britain (exclusive of Northern Ireland) in 1964 and 1966 were 3.0 percent and 4.8 percent respectively. “While the degree of error involved in a sample survey may be crucial in forecasting which party wins a highly competitive British election, it is of very limited concern in an analysis of major social characteristics of voters.” Similar reasoning would seem to apply when a major attitudinal characteristic of voters is added to the analysis. The results for the actual population of England include the nonwhite voters, which is thought to shift the results in the direction of Labour at all three elections by a small amount. Nonwhites are not included in the survey results. Since the N for the regression runs was reduced considerably, especially in 1970, by pairwise deletion of cases, it is noteworthy that the surveyed vote for the multivariate analysis is as follows:

¹⁹ On the use of dummy variables, see Cohen, Jacob, “Multiple Regression as a General Data-Analytic System,” Psychological Bulletin, 70 (1968), 426–43CrossRefGoogle Scholar; Suits, Daniel B., “Use of Dummy Variables in Regression Equations,” Journal of the American Statistical Association, 52 (December 1957), 548–51CrossRefGoogle Scholar; and Franklin, Mark N., Daedal: A Data Archiving, Editing, Describing and Analyzing Language (Edinburgh: University of Edinburgh, 1973), pp. 89–90Google Scholar. Treating ordinal-level variables as if they were interval is discussed in Labovitz, Sanford, “Some Observations on Measurement and Statistics,” Social Forces, 46 (December 1967), 151–60CrossRefGoogle Scholar; Labovitz, Sanford, “The Assignment of Numbers to Rank Order Categories,” American Sociological Review, 35 (June 1970), 515–24CrossRefGoogle Scholar; Reynolds, H. T., “Ordinal Partial Correlation and Causal Inferences,” in Measurement in the Social Sciences: Theories and Strategies, ed. Blalock, Hubert M. Jr., (Chicago: Aldine, 1974), pp. 399–423CrossRefGoogle Scholar; Reynolds, H. T., “On Testing Causal Models with Ordinal Data” (University of Delaware, mimeographed, 1972)Google Scholar.

²⁰ Labovitz, , “Some Observations on Measurement and Statistics,” pp. 156–58Google Scholar; Reynolds, , “Ordinal Partial Correlation and Causal Inferences,” p. 417Google Scholar.

²¹ Other research employing this procedure includes Stokes, Donald E., “Some Dynamic Elements of Contests for the Presidency,” American Political Science Review, 60 (March 1966), esp. 27–28CrossRefGoogle Scholar; Kim, Jaeon, Petrocik, John R., and Enokson, Stephen N., “Voter Turnout Among the American States: Systemic and Individual Components,” American Political Science Review, 69 (March 1975), 107–23CrossRefGoogle Scholar. This method is also the equivalent of the dichotomization of the dependent variables employed in Multiple Classification Analysis (MCA). See Andrews, Frank M., Morgan, James N., and Sonquist, John A., Multiple Classification Analysis: A Report on a Computer Program for Multiple Regression Using Categorical Predictors (Ann Arbor, Mich.: Institute for Social Research, University of Michigan, 1967)Google Scholar. Fred Kort argues that discriminant function analysis is preferable to multiple regression for a dichotomous dependent variable. See “Regression Analysis and Discriminant Analysis: An Application of R. A. Fisher's Theorem to Data in Political Science,” American Political Science Review, 67 (June 1973), 555–59CrossRefGoogle Scholar. His example, however, belies his argument. See also the exchange between Lemieux, Peter H. and Kort, in the “Communications” section of the American Political Science Review, 68 (March 1974), 202–05CrossRefGoogle Scholar.

²² Blalock, Hubert M. Jr.,, Causal Inferences in Nonexperimental Research (Chapel Hill: University of North Carolina Press, 1964), p. 52Google Scholar.

²³ The b weights and the actual averages of the percentage differences by crosstabulation were calculated for the bivariate relationship of opinion and vote, controlling for each of the party perception categories in each year. The comparative results were as follows (b weight presented first):

Party More

Likely to Keep

Immigrants Out

This procedure, unlike that using dummy variables, does not indicate which categories make the most difference in the scores of the dependent variable.

²⁴ This is the method used by Stokes to calculate attitudinal components of electoral decision in “Some Dynamic Elements of Contests for the Presidency.” See esp. 27–28.

²⁵ In a multiple regression equation, the b weights and beta weights of all the independent variables can be compared to see which ones make more of an impact on the dependent variable net of the others. A small b, of course, would indicate small percentage differences among the categories and a conclusion of no effect for that particular variable. Contrariwise, a large b can be interpreted as signifying large percentage differences in Conservative vote among the categories and a conclusion of substantial impact of that variable on voting behavior. The beta weights of all the independent variables can be used to calculate the net effects of each variable in a path analysis. If the b coefficient is 1.5 times its standard error of estimate, a statistically significant path exists between two variables. This is a liberal standard for determining the significance of a path. Since this is an exploratory study incorporating more variables than previously examined in studies of this kind, it is more important to include any possibly important variables in the equations than to exclude variables in the interest of a rigorously specified theory. In this case, no rigorously specified theory would justify more stringent standards for including variables. Using only those variables found to be significantly related, one then calculates a second set of equations. The values from this second set provide estimates of the path coefficients. This procedure enables one to assess both the direct and indirect contributions of a particular variable to variation in the dependent variable. Any masking of the effects of a variable by its relation to another variable will thus be revealed. In other words, the net effect coefficients can be considered partial correlations, with multiple controls for the effects of the other independent variables. See Knoke, David, “A Causal Model for the Political Party Preferences of American Men,” American Sociological Review, 35 (December 1972), 679–89CrossRefGoogle Scholar; Sani, Giacomo, “Determinants of Party Preference in Italy: Toward the Integration of Complementary Models,” American Journal of Political Science, 18 (May 1974), 315–29CrossRefGoogle Scholar.

²⁶ Butler, and Stokes, , Political Change in Britain, 2nd ed., pp. 287–95Google Scholar. The first development of these conditions was in Angus Campbell et al., The American Voter, pp. 168–87.

²⁷ Brody, Richard A. and Page, Benjamin I., “Comment: The Assessment of Policy Voting,” American Political Science Review, 66 (June 1972), 454–55CrossRefGoogle Scholar.

²⁸ Two exceptions to this generalization are Fishbein and Coombs, “Basis for Decision: An Attitudinal Approach Toward an Understanding of Voting Behavior,” and Knoke, David, “A Causal Synthesis of Sociological and Psychological Models of American Voting Behavior,” Social Forces, 53 (September 1974), 92–101CrossRefGoogle Scholar.

²⁹ This model is a variation of the one employed by Matthews, Donald R. and Prothro, James W., Negroes and the New Southern Politics (New York: Harcourt, Brace and World, 1966), pp. 27, 323Google Scholar. See also Sani, , “Determinants of Party Preference in Italy: Toward the Integration of Complementary Models”; Butler, and Stokes, , Political Change in Britain, 2nd ed., pp. 287–95Google Scholar.

³⁰ On developmental models, see Blalock, , Causal Inferences in Nonexperimental Research, pp. 85–87Google Scholar; Cnudde, Charles F. and McCrone, Donald J., “Party Competition and Welfare Policies in the American States,” American Political Science Review, 63 (September 1969), 860–62CrossRefGoogle Scholar. The causal sequence of the variables reflects a temporal order, that is, a person's social context and individual attributes are temporally antecedent to his political context, which in turn is assumed to be antecedent to his opinion on immigration and his vote. The individual attribute variables are also modeled according to an assumed temporal order in the life of an individual:

The categories of the individual attribute variables are as follows:

l. Age cohort (1 = less than 36 years, 2 = 36–50 years, 3 = 51–65 years, 4 = 66 years+)

2. Religion 1 (0 = unbelievers, 1 = Anglicans).

3. Religion 2 (0 = unbelievers, 1 = other Protestants)

4. Religion 3 (0 = unbelievers, 1 = Roman Catholics)

5. Church attendance (1 = never–unbelievers, 2 = never–believers, 3 = less than once a year, 4 = several times a year, 5 = several times a month)

6. Kind of school (1 = elementary, secondary modern, 2 = grammar, 3 = public school)

7. Education (0 = left school at minimum age, 1 = stayed in school beyond minimum age)

8. Social class (1 = D, 2 = C2, 3 = C1B, 4 = CIA, 5 = B, 6 = A)

9. Income (1 = under £350, 2 = £350–549, 3 = £550–749,4 = £750–1199, 5 = £1200 +)

10. Subjective social class ( 0 = working class, 1 = middle class).

³¹ The two questions were combined. Those who felt very strongly that there were too many immigrants in the country were assigned to the “strongly anti-immigrant” category. Those who felt only fairly strongly that there were too many immigrants in the country were assigned to the “moderately anti-immigrant” category. Those who thought that there were too many immigrants in the country but did not feel strongly about it were assigned to the “weakly anti-immigrant” category. Finally, those who did not think that there were too many immigrants in the country were assigned to the “favorable” category. There is a discrepancy in the N of question 2 in Table 1 and the N in Table 2 because in 1964 and 1966 people who did not think that there were too many immigrants in the country were not asked how strongly they felt about this issue. Since by themselves these people constitute the “favorable” category in Table 2, the N is larger for that table.

³² The stability of findings in electoral situations is similar to that in nonelectoral situations found by Studlar, Donley T., “British Public Opinion, Colour Issues, and Enoch Powell: A Longitudinal Analysis,” British Journal of Political Science, 4 (July 1974), 371–81CrossRefGoogle Scholar. Since electoral campaigns present the ordinary elector with stimuli that may cause him to behave differently than he would in electoral situations, the persistence of attitudes on immigration is worthy of emphasis. On the differential stimuli of electoral campaigns, see Berelson, , Lazarsfeld, , and McPhee, , Voting, pp. 291–96Google Scholar; and Miller, W. L. and Mackie, M., “The Electoral Cycle and the Asymmetry of Government and Opposition Popularity: An Alternative Model of the Relationship Between Economic Conditions and Political Popularity,” Political Studies, 21 (September 1973), 263–79Google Scholar.

³³ Since the surveys under examination are samples of the English electorate rather than panels, these figures do not mean that the direct movement of individual respondents is necessarily indicated by the overall figures.

³⁴ Accounts of the 1970 campaign and of the relative attention given in it to the immigration issue are contained in Butler and Pinto-Duschinsky, The British General Election of 1970, ch. 7, and Gale, George, “The 1970 Election Campaign” in Powell and the 1970 Election, ed. Wood, , pp. 50–82Google Scholar.

³⁵ The 1970 manifestoes of the two major parties read as follows:

Conservative Party: Good race relations are of immense importance. We are determined that all citizens shall continue to be treated as equal before the law, and without discrimination…. Local authority services are under great strain in many of the towns and cities where large numbers of immigrants have settled. We believe that additional funds should be made available to these local authorities in order that they can deal with these problems effectively without placing heavy burdens on their ratepayers.

We will establish a new single system of control over all immigration from overseas…. We believe it right to allow an existing Commonwealth immigrant who is already here to bring his wife and young children to join him in this country. But for the future, work permits will not carry the right of permanent settlement for the holder or his dependants….

These policies mean that future immigration will be allowed only in strictly defined special cases. There will be no further large-scale permanent immigration.

We will give assistance to Commonwealth immigrants who wish to return to their countries of origin, but we will not tolerate any attempt to harass or compel them to go against their will.

Labour Party: With the rate of immigration under firm cdhtrol and much lower than in past years, we shall be able still more to concentrate our resources in the major task of securing good race relations. The urban programme includes help to areas of high immigrant population, where special social needs exist. The Race Relations Act has outlawed incitement to racial hatred and discrimination in housing, employment and credit facilities…. We now propose to review the law relating to citizenship and to give the Race Relations Board powers of discretion in taking up complaints.

³⁶ Examples of mass detachment from political events are given in Butler, and Stokes, , Political Change in Britain, 2nd ed., pp. 22–23Google Scholar; and Abrams, and Rose, , Must Labour Lose? pp. 73–74Google Scholar. The argument is familiar, of course, to those who have read the SRC studies of American elections. See especially Stokes, Donald E. and Miller, Warren E., “Party Government and the Saliency of Congress,” in Campbell, et al., Elections and the Political Order, pp. 194–211Google Scholar.

³⁷ See the discussion of the conditions for issue voting in Butler, and Stokes, , Political Change in Britain, 2nd ed., pp. 287–95Google Scholar.

³⁸ Evidence supporting this contention for elections in the United States is presented in Berelson, , Lazarsfeld, , and McPhee, , Voting, pp. 215–33Google Scholar, and Campbell, et al., The American Voter, pp. 179–87Google Scholar. British findings on the same point are available in Benney, , Gray, , and Pear, , How People Vote, pp. 140–46Google Scholar; Milne, and Mackenzie, , Marginal Seat, 1955, pp. 117–21Google Scholar; and Pulzer, , Political Representation and Elections, pp. 114–15Google Scholar.

³⁹ See Stokes, “Some Dynamic Elements of Contests for the Presidency.”

⁴⁰ Butler, and Stokes, , Political Change in Britain, 2nd ed., pp. 303–05Google Scholar.

⁴¹ Ibid., pp. 307–08.

⁴² Full tabulations are available on request. Part of the data on party differences for 1964 and 1966 is simulated. This was necessitated by the original Butler-Stokes questionnaire for those years, which directed that only those who thought that there were too many immigrants in the country were to be asked which party was more restrictive. Those respondents favorable toward immigrants were not asked for their perceptions of party differences. A continuation of this procedure here would result in undesirable complications since perception of party differences is a necessary link between opinion on immigration and voting behavior. It was assumed that the closest available approximation to the actual distribution of “favorable” respondents over the perception categories in 1964 and 1966 would be the 1970 distribution. Accordingly, respondents favorable to immigrants in the 1964 and 1966 surveys were randomly assigned to one of the perception categories on the basis of the distribution of favorable respondents over the perception categories in the 1970 data.

⁴³ Butler, and Stokes, , Political Change in Britain, 2nd ed., pp. 307–08Google Scholar.

⁴⁴ Butler, and Pinto-Duschinsky, , The British General Election if 1970, pp. 159–63Google Scholar; Gale, , “The 1970 Election Campaign,” pp. 66–82Google Scholar.

⁴⁵ Fox finds that by 1974 the electorate was once more confused about which party was more restrictive. This is not surprising, since the Conservative government admission of refugee Asians from Uganda in the fall of 1972 served to alter the restrictive image Powell had given the party. See Fox, , “Attitudes to Immigration: A Comparison of Data from the 1970 and 1974 General Election Surveys,” 176–78Google Scholar. On the Uganda Asians episode, see Humphry, Derek and Ward, Michael, Passports and Politics (Harmondsworth, Middlesex: Penguin Books, 1974)Google Scholar.

⁴⁶ Page and Brody suggest that this test is sufficient to establish that no issue voting is taking place; however they do note that one should examine perceptions before drawing a final conclusion. They do not, however, consider that other variables should also be controlled in a multivariate design. See Page, Benjamin I. and Brody, Richard A., “Policy Voting and the Electoral Process: The Vietnam War Issue”, American Political Science Review, 61 (September 1972), 982Google Scholar.

⁴⁷ The full correlation matrices are available on request from the author.

⁴⁸ This average net effect for the entire sample in each year was calculated in the following manner. The b weight for each perception category was multiplied by the N in that category. Since the “no difference” b weights were not significant, 0 was assigned as the weight for this group of respondents. The results of these three multiplications were added together. This means that the figure from the Labour perceivers, with a negative b weight, was subtracted from the figure from the Conservative perceivers, with a positive b weight. This number was then divided by the total N in the sample for that year, resulting in the average net effect for the entire sample. The average net effect was standardized according to the formula provided by Stokes, , “Some Dynamic Elements of Contests for the Presidency,” 27–28Google Scholar, to yield the total net effect.

⁴⁹ The overall R² in all three regressions shows that this particular set of variables accounts very well for the variance in voting behavior at the 1964 British general election when compared with previous efforts, even though it is not the purpose of the study to explain the vote, but only to determine whether knowing a person's opinion on immigration adds to the explanation. See Rose, , “Britain: Simple Abstractions and Complex Realities,” pp. 525–30Google Scholar; Tate, C. Neal, “The Utility of Contextural Variables in Predicting Left-Right Voting Choice and Social Welfare Policy Preferences: Great Britain, Norway, the Netherlands, Germany” (paper delivered at Annual Meeting of the Southern Political Science Association, Atlanta, 1972)Google Scholar.

⁵⁰ The overall explained variance again compares favorably with previous attempts at accounting for voting behavior in Britain, especially in the “Labour best” category.

⁵¹ The total variance explained is much lower in 1970 than in the previous years. In the causal analysis, only variables directly or indirectly related to opinion on immigration were retained for the assessment of the impact of immigration opinion on voting behavior. Thus several of the variables contributing significantly to voting behavior, such as social class, are not included in the 1970 analysis. Since they did not affect immigration opinion, they could not account for any of the variation in the vote attributed to immigration opinion. The addition of social class to the variables in the full model in 1970 resulted in an R² for the three categories of perception of .35 (Labour perceivers), .23 (those seeing no difference), and .19 (Conservative perceivers).

⁵² Studlar, , “British Public Opinion, Colour Issues, and Enoch Powell: A Longitudinal Analysis”; Hiro, Dilip, Black British, White British (London: Eyre and Spottiswoode, 1971)Google Scholar; Bacharach, Elinor B., “‘Powellism’ and Political Culture: The Impact of Race Problems on British Politics” (M.A. thesis, University of Chicago, 1969)Google Scholar; Daniel, W. W., Racial Discrimination in England (Harmondsworth, Middlesex: Penguin Books, 1968)Google Scholar.

⁵³ An excellent illustration of the pitfalls of attributing large changes in election results to single factors is found in Wildavsky, Aaron B., “The Intelligent Citizen's Guide to the Abuses of Statistics: The Kennedy Document and the Catholic Vote,” in The Quantitative Analysis of Social Problems, ed. Tufte, Edward R. (Reading, Mass.: Addison-Wesley, 1970), pp. 37–67Google Scholar.

⁵⁴ The phrase is that of R. B. McCallum as quoted in Rose, Richard, Politics in England (Boston: Little, Brown, 1964), p. 3Google Scholar.

⁵⁵ Deakin, and Bourne, , “Powell, the Minorities, and the 1970 Election,” 402–15Google Scholar. In general, ethnic minority voting patterns are a neglected research question, largely because of the difficulty of ascertaining trends in so small a portion of the population, even with large random samples. Three limited inquiries into this problem are Lawrence, Daniel, Black Migrants: White Natives (Cambridge: Cambridge University Press, 1974), pp. 132–50Google Scholar; Community Relations Commission, Reference and Technical Services Division, Participation of Ethnic Minorities in the October General Election, 1974; and Anwar, Muhammed, “Asian Participation in the 1974 Autumn Election,” New Community, 4 (Autumn 1975), 376–83Google Scholar.

⁵⁶ Schaefer, Richard T., “Party Affiliation and Prejudice in Britain,” New Community, 2 (Summer 1973), 196–99Google Scholar; Schaefer, Richard T., “Correlates of Racial Prejudice,” in Sociological Theory and Survey Research, ed. Legatt, Timothy (Beverly Hills, Calif.: Sage Publications, 1974), pp. 237–64Google Scholar; Rose, E. J. B.et al., Colour and Citizenship (London: Oxford University Press, 1969), pp. 557–59Google Scholar.

⁵⁷ Butler, and Stokes, , Political Change in Britain, 2nd ed., pp. 42–47Google Scholar.

⁵⁸ Rose, , “Britain: Simple Abstractions and Complex Realities,” pp. 496–97Google Scholar.

⁵⁹ Page, and Brody, give extended consideration to this phenomenon in “Comment: The Assessment of Policy Voting,” 452–58Google Scholar.

⁶⁰ Butler, and Stokes, , Political Change in Britain, 2nd ed., p. 307Google Scholar.

⁶¹ Fox found considerable projection of the voters' immigration views onto their chosen party in the February, 1974 election. As has been argued previously, the Uganda Asians episode blurred distinctions between party policies. In such an ambiguous situation, it would be easier for voters to project their own immigration views onto their favored party than in 1970. Furthermore, in 1974 the Conservatives increased their vote over 1970 among those favorable toward immigrants and lost votes among those unfavorable; the Labour party situation was the reverse. See Fox, , “Attitudes to Immigrants: A Comparison of Data from the 1970 and 1974 General Election Surveys,” 175–78Google Scholar.