British Journal of Political Science, Vol. 19, No. 1. (Jan., 1989), pp

Candidates for Parliament: Transforming the Agenda? Pippa Norris; Joni Lovenduski British Journal of Political Science, Vol. 19, No. 1. (Jan., 1989), pp. 106-115. Stable URL: http://links.jstor.org/sici?sici=0007-1234%28198901%2919%3a1%3c106%3awcfptt%3e2.0.co%3b2-0 British Journal of Political Science is currently published by Cambridge University Press. Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at http://www.jstor.org/about/terms.html. JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use. Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at http://www.jstor.org/journals/cup.html. Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed page of such transmission. The JSTOR Archive is a trusted digital repository providing for long-term preservation and access to leading academic journals and scholarly literature from around the world. The Archive is supported by libraries, scholarly societies, publishers, and foundations. It is an initiative of JSTOR, a not-for-profit organization with a mission to help the scholarly community take advantage of advances in technology. For more information regarding JSTOR, please contact support@jstor.org. http://www.jstor.org Sun Feb 10 11:27:45 2008

TABLE 2 Threshold Model Predicting Monthly Success Rate SUCCESS, = 140.1 5 + 6.375 HISKILL,-,* + 0.649 APPROVAL,-,* (1.66) (1.71) - 14.085 MISERY,- I* + 5.34 PRES (- 2.53) (0.329) R2 = 0.383 Durbin-Watson Statistic = 1.91 * Significant at 0.05 level, one-tailed. variable for incumbent president is still in the pro-reagan direction, but is much further from significance than in the initial model. The biggest news, of course, is that our new operationalization of skill fits the data much better than does the original full index. The percentage of variance explained improves slightly over the first model, and the coefficient for the attenuated skill measure is significant in the expected direction. Taking the interaction into account, a president who crosses the threshold of high skill is at least 20 per cent more successful per month (on average) than one who remains below the threshold. The implication is that skill does not exert consistent effects on success over the entire range of skill; only when skill surpasses a certain level does it overcome the contextual influences on legislative success. Our analysis suggests a new way of looking at the controversy between the historical and the quantitative schools of presidency research. The usual way to reconcile the disparity between the two schools' findings on skill is to say that skill operates only in special contexts; only when conditions are 'right', as in Roosevelt's Hundred Days and Reagan's extended 1981 honeymoon, is it possible for skill to be efficacious. We have presented evidence that extraordinary displays of skill can prevail over contextual factors. Our analysis supports the contention that presidents, when they are perceived to be highly skilful, can influence their own legislative success. Candidates for Parliament: Transforming the Agenda? PIPPA NORRIS AND JON1 LOVENDUSKI Why should more women be elected to positions of power in Britain? What difference would it make? This Note aims to examine these questions using data from a survey of almost six hundred men and women candidates to the British Parliament in the 1987 general election. There are two major arguments for increasing women's representation. Department of Politics, Edinburgh University; Department of European Studies. Loughborough University. We should like to thank officers at party headquarters for supporting this research and providing assistance with lists of candidate addresses. particularly the help of Joyce Gould, Anne Wilkinson, Peter Knowlson and Wendy Buckley. The authors are grateful to all the candidates who completed the questionnaire while they were fully occupied by the campaign, and to Denise Ralph for the administration of the survey. Assistance towards the costs of this project was kindly made available by Professor Stewart Kirk of the Department of Human Sciences, Loughborough University. We should also like to thank Susan Welch for helpful comments concerning an earlier draft.

Notes and Comments 107 One argument is simply that there should be more women in politics on the grounds of symbolic equity. This symbolic argument is based on a theory of representation in which Parliament can only function democratically if it acts as a public forum for all points of view. reflecting the major divisions in society. Reform is justified by the self-evident disparity between the proportions of women in the electorate and in public office. In this view Parliamentary democracy requires increased female representation as an end in itself, irrespective of whether women make a difference. This view has an immediate straightforward appeal but it is not clear why women should be represented qua women unless they have a distinct set of concerns and interests. The substantitie argument is stronger. It claims that more women in Parliament would make a substantive difference. From this perspective more women should be elected because they represent a 'woman's point of view', with distinctive values, attitudes and concerns which may have an impact on legislative behaviour and the content of public policy. Alternative studies have claimed that women are more 'liberal', 'conservative'. 'pacific', 'moralistic', 'humane' or 'feminist' in their ideological beliefs. These gender differences in attitudes can be attributed to women's experience or to innate differences. However, there is little agreement on whether there are significant gender differences in politicians, and, if so, what their nature is. The evidence, particularly concerning Britain, can only be described as inconclusive. PREVIOUS STUDIES In the United States Stanwick and Kleeman found that elected and appointed women officeholders are likely to be more sympathetic to liberal policies such as nuclear disarmament and welfare programmes.' In separate studies Leader, Frankovic, Welch and Norris have confirmed that in Congress women are more liberal than their male colleagues in their attitudes and in their legislative behaviour, although the differences were not great and may be dimini~hing.~ Similar attitudinal gender differences have been found among US party delegates and state legislator^.^ Studies in the Nordic countries suggest that women have played a crucial role in raising the profile of women's issues on the legislative agenda.4 ' K. Stanwick and K. Kleeman, Make a DifSrrence (New Brunswick, NJ: Center for the American Woman and Politics, Rutgers University, 1983): see also S. Carroll, ' Candidates and Support for Feminist Concerns: The Closet Feminist Syndrome'. Western Political Quarterl!. 37 ( 1984) 307-23. S. G. Leader, 'The Policy Impact of Elected Officials'. in L. Maisel and J. Cooper. eds. The Impact of the Electoral Process (Beverly Hills. Calif.: Sage, 1977): K. Frankovic. 'Sex and Voting in the US House of Representatives, 1961-1975', American Politics Quarterly, 5 (1977). 315-30: F. L. Gehlen, ' Members of Congress: A Distinctive Role', in M. Githens and J. Prestage. eds. A Portrait of,mar(]inality (New York: McKay. 1977): P. Norris, ' in Congress: A Policy Difference?'. Politics, 6 ( 1 986). 3440: S. Welch, 'Are More Liberal than in the US Congress?'. Legi.slati~.e Studies Quarterly, 10 (1 985). 125-34. S. Baxter and M. Lansing, and Politics (Ann Arbor: University of Michigan Press. 1983). p. 133; J. W. Soule and W. E. McGrath, 'A Comparative Study of Male-Female Political Attitudes at Citizen and Elite Levels', in Githens and Prestage. A Portrait of Marqinalitj,: I. Diamond, Seu Roles in the State House (New Haven. Conn.: Yale University Press, 1977). D. Dahlerup and E. Haavio-Mannila. in E. Haavio-Mannila et a/., eds, Unfinished Democracy: in Nordic Politics (London: Pergamon Press. 1985): M. Eduards, 'Sweden'. in J. Lovenduski and J. Hills, eds. The Politics ofthe Second Electorate (London: Routledge and Kegan Paul. 1981).

108 Notes and Comments Yet some observers remain sceptical, arguing that women who enter political elites successfully are largely similar in their political attitudes to their male colleagues, due to the process of selective recruitment, party pressure and political socialization within legislature~.~ The minority status of women in public office keeps women within the mainstream of politics. Vallance found that in the British Parliament most women felt that their loyalty lay with their party and constituents, rather than with women per ~ e In. the ~ European Parliament Vallance and Davies found that women MEPs were pre-occupied with specialist interests or constituency interests, rather than 'women's issues', although women were strongly involved in sex equality legislation.' Diamond suggested that the majority of female state representatives in the Unites States do not campaign on women's issues nor, once elected, do they actively pursue explicitly feminist goals.' Holsti and Rosenau found considerable similarities in the foreign and defence policy attitudes of women and men in the American political elite.g Sapiro has suggested that perceptions about gender differences in politicians may reflect common stereotypes about women as 'nurturers and carers', rather than adequate evidence.'' Whatever the explanation, there is strong evidence that, once elected, women politicians behave similarly to their male counterparts, which implies that the case for more women in office must be argued on the grounds of symbolic equity rather than substantive effect. This long-standing debate has been given added impetus by recent party efforts to increase the number of women candidates. The 1987 election was notable for the entry of a record number of women to the House of Commons. The proportion of women candidates for the major parties increased from 194 (10.2 per cent) to 250 (12.8 per cent) and there was a sharp rise in the number who were elected. Between 1983 and 1987 the number of women in Parliament doubled from twenty-three to forty-one (see Table 1). Yet in 1987 women MPs still constituted only 6.3 per cent of the House of Commons, one of the lowest proportion of women in European legislatures." Immediately before dissolution there had been twenty-eight women MPs due to a series of by-election victories since 1983. The change was therefore less dramatic than it appeared at first sight. DATA AND METHODS But will more women in Parliament make a difference to public policy? In Britain we lack systematic evidence to assess this question. The overriding problem is that with so few women MPs we have to rely upon anecdotal and impressionistic evidence rather than a more systematic approach. It is difficult to generalize about such diverse politicians as V. Randall, and Politics (London: Macmillan, 1987). E. Vallance, in the House (London: Athlone, 1976); see also M. Currell, Political (London: Croom Helm, 1974). ' E. Vallance and E. Davies, of Europe (Cambridge: Cambridge University Press, 1986). ' Diamond, Sex Roles in the Stare House; T. Bers, 'Local Political Elites', Western Political Quarterly, 31 (1978), 381-91; S. Mezey, 'Support for 's Rights' Policy', American Politics Quarterly, 6 (1978), 485-97; Gehlen, ' Members of Congress'; Frankovic, 'Sex and Voting in the US Representatives'; C. Mueller, 'Feminism and the New in Public Office', and Politics, 2 (1982), 7-21; Carroll, ' Candidates and Support for Feminist Concerns'. Holsti and Rosenau, in E. Boneparth, ed.,, Power and Policy (New York: Pergamon, 1982). lo V. Sapiro, The Political Integration of (Chicago: University of Illinois Press, 1983). 'I E. Vallance, 'Two Cheers for Democracy: Candidates in the 1987 General Election', Parliamentary Affairs, 41 (1988), 86-91.

TABLE 1 Female Candidates and MPs, 1983 and 1987 Notes and Comments 109 1983 1983 1987 1987 Candidates M Ps Candidates M Ps Conservative 40 6.3 13 3.3 46 7.3 17 4.5 Labour 78 12.3 10 4.8 92 14.5 21 9.2 LiberalISDP 76 12.0 0 0.0 105 16.6 2 9.1 PC/SN P 16 14.5 0 0.0 7 6.6 1 16.6 Total above 210 10.2 23 3.5 250 12.8 41 6.4 Note: The percentages are based on the total number of candidates and MPs in the major parties Source: The Times Guide to the House ofcommons. 1983, 1987. Edwina Currie, Diane Abbott and Margaret Thatcher. In addition, most legislative behaviour in Parliament is severly constrained by the nature of party discipline. It is possible to analyse the record on relevant free votes but these involve only a restricted range of issues. We can compare gender differences in behaviour through the membership of Parliamentary committees, affiliation to party groups, or Parliamentary Questions, but again we have few cases for comparison. Yet evidence is available which enables us to assess the logically prior question: are women candidates for Parliament more liberal or feminist in their attitudes and priorities? Gender differences in attitudes can be seen as a necessary but not sufficient condition for gender differences in behaviour. A wider sample drawn from candidates for the major parties allows for more reliable generalizations than a study restricted to the Parliamentary elite. Accordingly this Note analyses data from a survey of men and women candidates from the major British parties in the 1987 general election (British Candidate Survey, 1987).12In late April/early May 1987 a postal questionnaire designed to gather information about the social background, political experience and attitudes of candidates was distributed to 2,004 candidates, including 250 women. Minor parties with no Parliamentary representation were excluded from the British Candidate Survey, including the Green and Communist parties, both with a relatively high proportion of women candidates. In total the British Candidate Survey included 590 completed questionnaires, replies from a third of all candidates (including 106 women and 484 men). The sample included 139 British MPs which represents 22 per cent of the current Parliament.13 The low response rate was not unexpected as most candidates were fully occupied by the general election campaign in May/June. Nevertheless, we were concerned to establish whether there was a bias in the sample, for example, whether the questionnaire was more likely to be completed by losing candidates than by winning ones. Accordingly we checked the characteristics of the respondents against a profile of all candidates derived from The l 2 For other results from the survey see J. Lovenduski and P. Norris, 'Political Recruitment and the Masculine Candidate Model in the British General Election of 1987: Obstacles to the Feminisation of Political Elites', European Journal ofpolitica1 Research (1989). l3 The survey included responses from 139 MPs, who represented 24 per cent of the sample, compared with 633 elected British MPs out of 2,004 candidates, which represents 31.5 per cent of all candidates for the major parties.

1 10 Notes and Comments Times Guide to the House ofcommons, 1987.14Using t-tests we found no statistically significant differences (p >0.05) between the sample and all candidates in terms of gender, age and education, and the slightly lower response rate amongst Conservative candidates was adjusted by weighting for party.15 On this basis we concluded that although the response rate was not ideal, the sample was representative of all candidates. Certain previous studies claim that the gender gap in attitudes extends to different priorities about the major issues in politics. Some argue that women candidates, given their experience of family responsibilities, or their occupational background, may have different concerns about public policy issues such as childcare, education or social services while male politicians may give a higher priority to issues like defence, trade unions or unemployment. Is this is the case them more women in politics might change the agenda, as well as the direction, of the Parliamentary debate? TABLE 2 Mean Issue Salience All Conservative Liberal/SDP Labour M F M F M F Unemployment 7.2 6.0 6.9 7.3 7.3 7.9 7.7 NHS 5.7 4.3 5.1 5.9 6.1 6.8 7.1 Education 5.2 4.3 5.5 5.7 5.8 5.5 5.9 Defence 4.1 5.4 4.5 4.2 3.9 2.9 4.1 Law and order 3.6 4.7 5.3 3.3 2.5 3.7 3.9 Social services 2.8 1.8 2.3 2.3 2.9 3.9 3.5 Inflation 2.4 4.4 4.4 2.2 2.0 1.6 0.6 Nuclear energy 1.6 0.8 1.1 1.3 1.6 1.7 2.0 Equal opportunities 1.8 0.6 0.3 1.5 1.9 2.8 2.3 Local government 1.8 1.4 2.2 1.9 1.6 1.7 1.3 Unions 1.1 1.9 1.1 0.7 1.1 0.9 1.3 Childcare 0.9 0.5 0.1 0.4 0.7 1.8 1.9 Agriculture 0.5 0.6 1.1 0.6 0.5 0.3 0.3 Note: Respondents were asked to rank these issues from most important (10) to least important (1). Source: British Candidate Survey, 1987. l 4 Alan H. Wood, ed., The Times Guide of the House of Commons, 1987 (London: The Times, 1987). As a result of the first mailing, 542 completed questionnaires were received, along with some refusals (N = 28) and non-contacts at the given address (N = 10). Given limited resources we were unable to send a second mailing to all non-respondents, but reminders were sent out after the election to female non-respondents because we were concerned about the number in the sub-sample. Following the second reminder we received 628 replies in total, including 590 completed questionnaires (29.5 per cent of all candidates). The response rate from Conservative candidates was slightly lower (26.2 per cent) than from Labour (33.5 per cent) or the Alliance (31.6 per cent). Accordingly, the Conservative response was adjusted by applying a positive weighting (1.26) to the results so that the sample remained representative of all candidates. It should be noted that given the small number of cases in some sub-samples (e.g. female Conservative candidates) considerable caution is needed when interpreting the results in cross-tabulations.

Notes and Comments 11 1 To assess this hypothesis, we asked candidates to rank the most important issues in the general election, ordering ten from a list of items. The mean results were then ranked from most important (10) to least important (1). The issues given a high priority by the candidates were similar to those nominated by the public in most public opinion polls during the campaign, namely unemployment, the National Health Service, education, defence and law and order (see Table 2). There was quite a high level of consensus among all candidates about the relative rank order of the issues, although the social services were seen as more important by Labour while inflation was given greater priority by the Conservatives. There was also agreement that certain issues were not considered important, including those such as childcare and equal opportunities, which are part of the feminist agenda. Within each party, there were certain gender differences in priorities which proved to be statistically significant. Across all parties women put a slightly higher priority than men on the ~lurturing issues of the NHS and education, as well as on nuclear energy; only in the Labour party do women give higher priority than men to 'defence' (which probably means 'peace'). There was no strong tendency for women to give higher priority than men to equal opportunities. Overall there was a large measure of agreement about the relative rank orderings. On this evidence, therefore, there are gender differences in candidate issue priorities but the differences are minor. To analyse gender differences in political attitudes we asked candidates to respond to a series of six standard items, used since 1979 in the (Essex/Oxford) British Election Survey, which can be employed to measure liberal or conservative attitudes towards social change.i6 These items ask respondents whether they feel that certain aspects of social change, such as welfare benefits, the availability of abortion, equal opportunities for women, nuclear power, racial equality or defence spending, have gone too far, not far enough, or are about right (see Table 3). These items were designed to examine enduring and fundamental aspects of liberal and conservative attitudes, not the specific policy proposals or issues which featured in the 1987 campaign debate. Responses were coded to range from liberal (+ 1) to conservative (-1) and the mean scores were calculated for each item. There was a strong correlation between these items, which suggested that we could use the series as the basis for a summary CON-LIB Index ranging from liberal (+6) to most conservative (-6)". Party was re-coded into a dummy variable for Government versus all Opposition parties. l6 The items were as follows: 'Can you say whether the following have gone too far, not far enough or are they about right: (a) Welfare benefits that are available to people today (WELFARE); (b) Attempts to give equal opportunities to women in Britain (WOMEN);(c) Government spending on defence (DEFENCE); (d) The building of nuclear power stations (NUCLEAR); (e) Attempts to give equal opportunities for black people and Asians in Britain (RACE);(0The availability ofabortion on the NHS (ABORTION)'. I' Pearson correlations between attitudes on these items were significant (p >0.01 in each case. N = 585). Welfare Defence Nuclear Race Defence Nuclear Race Abortion

1 12 Notes and Comments The bivariate analysis suggests that, as expected, party proved to be most strongly associated with conservative-liberal attitudes. Nevertheless, women were more liberal within each party, including the Conservatives. In terms of parties, there was a predictable ranking, with the clearer distinction being between Conservative candidates and the Opposition parties (see Table 3). The results also demonstrate that within the Alliance the leadership divisions which emerged during the last election between David Owen and David Steel reflected wider differences in their parties, with the Social Democratic candidates closer to Conservatives while the Liberals were nearer Labour. The Pearson correlation analysis suggests that the candidate's gender, occupational class, education and age were also significantly related to attitudes. Since these factors were inter-related, for example more women candidates in the last election stood for the TABLE 3 Analysis of Mean Conservative-Liberal Index Mean CON-LIB Rating No. of Candidates Pearson Corr. (r) Standardized Beta Coefficient Party Conservative SDP Liberal SNP/PC Labour Gender Occupation Prof./Managerial Other Age group Younger Older Education Graduate Non-grad. Region Northern Southern Multiple R Adjusted R2 - Note: Attitudes were codes from +6.0 (most liberal) to -6.0 (most conservative). For the Pearson correlation and multiple regression, parties were re-coded into Government (1) and Opposition (0). For the items see fn. 16 (p >0.01 = **). Age groups were classified as 'Younger' (under 40) or 'Older' (over 40). Region was classified using the twelve standard census regions collapsed into two categor- ies: 'Northern' (North, North-West, Yorkshire and Humberside, Scotland and Wales); 'Southern' (all the other regions). Sourer: British Candidate Survey, 1987.

Notes and Comments 113 Labour party than for the Conservative, we entered the independent variables in a multivariate analysis using ordinary least squares multiple regression. For the analysis we transformed into dummy variables the candidate's party (Conservative/ Opposition), region of the constituency (Northern/ Southern), occupational class (professional and managerial/ other) and education (graduate/ non-graduate) with mean conservativeliberal score as the dependent variable. The results confirmed the earlier conclusion that party was the strongest predictor of liberal attitudes, but gender, age and education continued to prove significant using standardized beta co-efficients. The factors of region and occupation dropped out of the analysis as insignificant, since these were strongly related to the candidate's party. Overall the model showed a high level of goodness of fit with an adjusted RZof 0.54. The results therefore suggest that gender is one of the factors which distinguishes attitudes among Parliamentary candidates: within each party women tended to prove more liberal than men. As shown in Table 4, gender differences were particularly marked in the Liberal and Nationalist parties, where women tended to be as liberal as the men in the Labour party. The picture becomes more complex if we turn to the specific items within the index. The results suggest that on a range of issues the sharpest divide was between the Conservatives and the Opposition parties (see Table 4). On the provision of social services 94 per cent of Conservative candidates felt that welfare benefits had gone too far or were about right, while in contrast 92 per cent of Labour candidates believed that they had not gone far enough. The candidates were also polarized by party on the question of equal opportunities for women, with 89 per cent of Labour and 80 per cent of Alliance candidates in favour of further change compared with only 14 per cent of Conservatives. Attitudes towards defence spending and racial equality showed a similar party divide. TABLE 4 Mean Liberal-Conservative Scores by Party and Gender Lib-Con score N Conservative SDP Liberal PC/SNP Labour - -- - -- -- - Note: Attitudes were coded from +6.0 (most liberal) to -6.0 (most conservative). For the items see fn. 16. Source: British Candidate Survey, 1987.

114 Notes and Comments TABLE 5 Mean Liberal-Conservative Attitudes to Change Con women Con men - SDP women SDP men Lib women Lib men Lab women Lab men Gamma by gender Gamma by party Abortion Welfare Race Nuclear Defence equal benefit equality power spending!yore: Attitudes were coded from + 1.0 (most liberal) to - 1.0 (most conservative). Gamma coefi- cients olassociation (chi square, p > 0.01 = **), N = 586. For the items see fn. 16. Sourc,e: British Candidate Survey. 1987. Yet the results suggest that within each party there was a significant gender gap on many of the issues, although considerable caution is needed when interpreting the results given the smaller number of cases in each sub-category. Among Alliance candidates women tended to be slightly more liberal than men across every item, with the greatest differences on defence spending, abortion and equal opportunities for women. In the Labour party the gender differences were less marked or consistent, except on reproductive rights where women were more strongly pro-abortion. But on this issue, just as in the electorate, women's attitudes are polarized; women candidates do not speak with one voice. In the Conservative party a higher proportion of women than men felt that the availability of abortion had gone too far and none supported further liberalization. Feminists cannot assume, therefore, that women candidates will necessarily support 'women's issues' more strongly than men in their party: it depends on the nature of the issue. candidates across all parties strongly supported female equality of opportunity but abortion proved more divisive, as indeed it does in the women's movement. DISCUSSION AND CONCLUSIONS Overall our results provide support for the substantive argument. Among Parliamentary candidates women tend to be more sympathetic than men to liberal concerns in their parties. If these attitudes affect legislative behaviour then the entry of more women into Parliament could influence the direction of public policy. The extent of the gender difference should not be exaggerated. candidates remain party politicians first and foremost. The evidence in this survey is necessarily limited and further research needs to examine elite attitudes towards a wider range of issues, particularly feminist concerns. Nevertheless, the results are significant and provide tentative support for the view that in Britain women politicians may prove substantively different to their male colleagues.

Notes and Comments 115 Whether the gender difference is sufficient to matter is open to debate. A number of reasons suggest that it may. Party discipline might preclude backbench attitudes from having a major effect on the legislative process in Britain, yet there are free votes in Parliament where backbenchers can play a decisive role. Elite attitudes can influence the non-legislative activities of Parliament through written and oral questions, private correspondence, select committees and adjournment debates. Gender differences may prove even more significant if women move into policy-making positions within government. Lastly, some believe that gender differences could transform public policy if enough women were elected to reach a 'critical mass' in Parliament. It is argued that if women remain below 10 per cent of the legislature gender differences in attitudes will fail to make a major impact. Those women who manage to enter the House of Commons will feel that they have to play by the 'rules of the game' to succeed. If they have different priorities or values they will not feel free to articulate them. Once women candidates are more successful, as in the European and Scandinavian countries where 20 per cent to 30 per cent of the legislature is female, women may be able to transform the agenda. Until that critical mass is reached the odds against women politicians qua women having a measurable influence are so high that almost any indicators that they have affected the policy agenda must be taken seriously. Neither the substantive nor the symbolic arguments for increasing the number of women politicians may be safely discarded. Community Cohesion and Voter Turnout in English Parliamentary Constituencies MUNROE EAGLES AND STEPHEN ERFLE Voting turnout varies both over time and across space. In Britain there has been a secular trend in the postwar period for decreasing turnout at parliamentary elections (from a high of 84.1 per cent in 1950 to 75.4 per cent in 1987, with a low point of 71.8 per cent in 1970). Such temporal variations in turnout are dwarfed in scale, however, by differences in turnout across constituencies at the same election. In the 1970 election, for example. turnout ranged from a low of 44.9 per cent in Stepney to 85.3 per cent in Cornwall North. Though diminished slightly, variation in constituency turnout rates remained significant in the June 1983 election (from a low of 51.8 per cent in City of London and Westminster South to a high of 81.1 per cent in Leicestershire NW). Given the underdeveloped state of empirical research on participation generally,' it is not surprising to find few studies of variations in turnout across constituencies. In his recent geographical analysis of the 1983 election, for example, R. J. Johnston had difficulty developing hypotheses about turnout due to the paucity of existing re~earch.~ Consequently, although the importance of community characteristics as determinants of Department of Political Science, St Mary's University, Halifax; School of Social Sciences, University of California, Irvine. The authors acknowledge helpful comments from Ivor Crewe. David Denver, Martin Harrop and C. L. Sullivan on an earlier draft of this Note and financial assistanc'e from the University of California, Irvine. ' Anthony Mughan, Party and Participation in British Elections (New York: St Martin's Press. 1986), pp. 1-21. However, research on political participation in Britain is showing signs of increasing vitality as seen in the recent work of George Moyser and Geraint Parry, 'Class, Sector and Political Participation in Britain'. Manchester Papers in Politics (Manchester: Department of Government. University of Manchester, 1987). pp. 147. R. J. Johnston, The Geography of English Politics (London: Croom Helm. 1985). p. 135.