Navigation – Plan du site

AccueilNuméros23Géographie et sociologie électora...“It is the Context, Stupid!”. Or ...

Géographie et sociologie électorales : duel ou duo ?

“It is the Context, Stupid!”. Or is It? British-American Contributions to Electoral Geography since the 1960s

« C’est le contexte, idiot! » Ou non? Contributions anglo-américaines à la géographie électorale depuis les années soixante
Herman van der Wusten et Virginie Mamadouh


La notion de ‘contexte’ a été d’une importance fondamentale dans la géographie électorale anglo-saxonne depuis les années soixante. Kevin Cox l’a utilisée pour organiser son article fondateur dans Progress in Geography en 1969. Le présent article souligne l’importance du travail innovateur de Cox et le relie aux travaux de deux des auteurs les plus importants de la période suivante, Ron Johnston et John Agnew. Cet article retrace ensuite les thèmes majeurs de géographie électorale et indique l’importance de différentes notions de contexte à l’intérieur de ce domaine plus étendu. Cet article conclut en indiquant un programme de recherche émergent pour les géographes électoraux et ainsi qu’avec quelques notes sur la dissolution graduelle des nettes divisions antérieures entre le domaine anglo-saxon de la géographie électorale et d’autres domaines, et entre les imaginations strictement séparées des sociologues et des politologues d’un côté et des géographes de l’autre en ce qui concerne les élections.

Haut de page

Notes de l’auteur

The title of this paper summarizes the argument between geographers and political scientists by means of a snowclone – a cliché combined with a wordplay - frequently used in American electoral campaigns and political culture. It derives from “the economy, stupid!”, a key message in Bill Clinton’s 1992 presidential campaign against George H.W. Bush.

We thank one of the anonymous referees who pointed us to Jonas & Wood’s (2012) important collection.

Texte intégral


1A recent meta-analysis of not less than 90 studies on voter turnout in national elections for the 2000-2010 period only (Smets, Van Ham, 2013, p. 344-359), shows a lack of consensus on a core explanatory model and a wide variety of control variables although they are often referred to as ‘the usual suspects’. In the field of context effects in voting behaviour the situation is slightly different, but does not seem brighter. There is a semi-permanent standoff between believers (now notably concentrated among geographers) and non-believers (mostly, but not exclusively among political scientists) and the number of studies oriented to the identification of such effects seems to be faltering. The cottage industry in that field from the late 1960s on seems to have largely folded. In Electoral Studies 2009-2013 from about 150 papers published we only found three that could be considered to be more or less focussed on contextual effects. In Political Geography during the same period out of about 300 contributions 14 seemed primarily concerned with electoral geography, out of which perhaps a handful were even remotely concerned with contextual effects. At the same time amongst those who consider context effects to be a fruitful part of the explanation for electoral behaviour, no widely shared core model of explanation has emerged. Even the taste for one seems to have disappeared.

2This does not mean that the subject should be left on the wayside. Despite insistent recommendations to geographers not to pin their hopes for disciplinary recognition on the definitive identification of contextual phenomena (King, 1996), belief in the fruitfulness of further research in that direction has certainly not completely subsided. A rather comprehensive overview of the then existing literature in terms of methods and substantive results concluded in 1997 that local contextual effects were perhaps not overwhelmingly powerful, but recurrently turned up as significant in electoral studies (as they did in other kinds of participatory behaviour, e.g. church attendance) (De Vos, 1997). More recently further enrichments of the literature have been added as we will see. It is also highly suggestive that the meta-analysis on turnout picks up ‘region’ and ‘residential mobility’ – central elements in all discussions on the subject - among the few recurrent explanatory factors.

3In this paper we will move on in that direction particularly based on the experience in the field of British and American political geography. We will first of all reiterate, mark the commonalities and contrast three core authors on the subject: Kevin Cox who authoritatively put the subject on the agenda of electoral geographers in the years around 1970, his coeval Ron Johnston and John Agnew, one of Cox’s first postgraduate students. On that basis – certainly not a complete overview but providing an outline of the general development – we will then indicate the position from where to move forward elaborating on the way the context-issue sits in the current study of electoral geography.

Foundational voices: Cox, Johnston, Agnew

4A major effort to introduce the geographical situatedness of individuals and their voting decisions in electoral geography is Kevin Cox’s work in the years around 1970. Freshly arrived from Britain in the US, Cox produced a PhD on electoral matters in 1966 supervised by an economic geographer and partly inspired by a seminar offered by a political scientist in which he had taken part. He sent his dissertation to Stein Rokkan in Bergen, Norway (Johnston, Pattie, 2012, p. 35), a pivotal figure in a transatlantic network of fruitful collaboration in the field of a historically informed political sociology in that period (his still relevant, widely scattered output has been ably systematized in Flora, 2000). As a consequence Cox’s work was published in two of the major collections of the network (Dogan, Rokkan, 1969; Allardt, Rokkan, 1970). The overall aim was to renew the comparative study of the electoral institution and all it entails, notably also the way in which long-standing regional diversity affected party formation and electoral preferences. To account for the various national party systems and the long term electoral fortunes of their different parts a model of social cleavages that could in different ways be politicized, was put forward. In this circle there was a sensitivity to the different historical paths that regions and places had followed for their populations to be caught by way of ‘context-effects’. Cox published valuable work in the collections just mentioned and in journals focussing on these context-effects in explaining contrasts in the Conservative vote between inner London and its suburbs, extra support for the French communist party in the different parts of Paris depending on local concentrations of a working class and long term contrasts in voting preferences for either Liberals or Labour within Wales depending on variations in local circumstances. A complete overview is in Johnston, Pattie, 2012. That chapter also ably summarizes the state of American geography in which Cox’s initiative came to fruition.

5In a seminal paper published in 1969 in the very first issue in Progress in Geography preceded by one on ‘environment as perceived’ and followed by one on ‘diffusion research in geography’, Cox proposed a distinctive geographical perspective on elections concentrating on the act of voting by individuals putting it in what he called ‘their spatial context’. Summarizing the results of the various empirical studies done so far by the circle of political sociologists with a view to contextual impacts he tried to re-read them as manifestations of the spatial context in which each voter necessarily operated. Spatial was notably read by way of distance, and the extension of relevant environments and the question was how voters were affected by spatially modelled communication patterns and also by voters’ movements in space. In his perspective for the further development of electoral geography Cox was strongly influenced by Hägerstrand’s studies on the propagation of innovation waves, to be studied by Monte Carlo simulations of modelled communication fields that were followed by his time-geography.

6In Cox’s own further work this programme came to nothing. By 1973 he had turned his attention to other problems, now considering “that the study of voting patterns should not be seen as an end in itself. I could not see how they could be related satisfactorily to political geography as a whole” (Cox in Johnston, Pattie, 2012, p. 36). The other thing is probably that voting is (or at least was for the most part at the time) different from innovation adoption: for most people a recurrent act of participation in an institution of considerable stability in which the adoption of an innovation (supporting a newly established party?) was a rare event. For many the act of voting is more than a conscious decision moment, yet another instance of a longstanding habit, for many people also an act of secondary importance. Therefore the act of voting should perhaps not be modelled as a one shot, distinctive decision. However that may be, Cox opened the door for other geographers to the study of elections in a context dominated by other social sciences and looking for a distinctively geographical perspective.

7Ron Johnston started to write about the votes in elections and the significance of local contact patterns around that time. His first paper on such matters in his voluminous output seems to date from 1972. Cox’s paper was a major inspiration (Johnston, 2010, p. 553). In Taylor and Johnston’s Geography of elections (1979), part 3 called ‘Geographical influences on voting’ (that apparently is Johnston’s work) deals exclusively with relevant local factors and provides a pretty exhaustive overview of the literature of the moment mostly written by political scientists and political sociologists. In an impressive series of works Johnston (with various collaborators) has since then mainly focussed on the UK in dealing with regional and local biases from the national patterns of political preferences and their explanations in successive elections. He has experimented with different techniques to master the complicated problems of adequate measurements of the various effects possibly implicated under the contextual effect rubric. Thus, he has put ‘Voters in their place’ as he has summarized his efforts in his major work on the 2005 election (Johnston, Pattie, 2006). He has since then turned his attention to the 2010 case.

8Johnston has given up on the idea that there should be a distinctive, separate field for geographic electoral studies. He accepts the force of the social composition of a population for explaining the vote. But he emphasizes the importance of ‘socialization in place’ (inspired by Agnew, see below) which contributes additional explanatory power as significant place-based extras: in brief the effect of political conversation in the vicinity, the directly visible and consequential effects of economic development in one’s own local environment, the effects of local campaigning or its virtual absence. All these extras are in their turn embedded in larger scale networks. In this way and in increasingly full cooperation with political scientists and sociologists Johnston has made an extremely valuable contribution to an interdisciplinary field of Electoral Studies (the journal with that title appears since 1980).

9John Agnew (1987, 2002) has made his mark in this field in the 1990s publishing two books at the start and the end of that period both entitled ‘Place and politics…’ (with something added). Both deal primarily with electoral politics but also occasionally with parties and collective action. The first book set the programme illustrated with evidence from Scotland and the US, the second was primarily demonstration of its potential based on the Italian case. In between Agnew reiterated, sharpened and extended his views on a number of occasions.

10Agnew (1987) opens a frontal attack on political sociology as practiced and tries to develop a new perspective on political activity, e.g. voting, from a position as a geographer but in line with a number of scholars who propose a renewed epistemology for the social sciences, notably Bourdieu, Giddens and Bhaskar, plus Pred from within geography. In this new approach, structures and practices are continuously shaping each other. At one point he considers the aim of his project to be the closing of the divide that Durkheim and Vidal de la Blache opened around 1900 in their argument about prioritizing either ‘social facts’ or ‘place facts’, thus putting the sociological and the geographical imaginations on separate paths (Agnew, 1987, 229). Obviously the stakes in the end are higher than simply the field of electoral studies.

11Agnew criticizes the practice notably of American political sociology of the 1960-1980s as nearly exclusively focused on the individuals, each one caught within the confines of a nation state. The resulting psycho-social grid gives rise to nation-specific analyses where individuals are sorted in the categories indicated by nation-wide cleavages. Individual traits supposedly explain turnout and preferences based on correlations but leaving the causal mechanisms out of sight. National differences (often considered in terms of stages of modernization) should account for differences in cleavage structures and the resulting parties between national settings. Contextual effects in that perspective are only provisional, still unexplained, remnants to get rid of in the next round of analysis.

12Agnew instead proposes a geohistorical grid where people all have their place and where political engagements materialize in specific microsociological settings, not reducible to collections of individual traits. These settings are the places that count. Places are definitely not isolated cases. A major part of their identity is in their connections with other places most certainly not limited to the various national territories. Political engagements are primarily constituted in places and then possibly aggregated in larger wholes (e.g. through movements and parties). But overarching movements and parties can in their turn also significantly affect the nature of places. A contextual analysis in Agnew’s view should foreground this embedding of places in larger networks and the ways in which the resulting different scales affect each other.

13Taking off from where Cox started, Agnew and Johnston have taken somewhat different roads. In the end they find themselves at positions that are not very far apart and apparently well reconcilable as Johnston indicates by establishing a general viewpoint that uses some of Agnew’s main insights, particularly his focus on place, as their point of departure (Johnston, Pattie, 2006, p. 43).

14Cox while doing empirical work in line with the main tenets of political sociology at the time, tried to add a distinctively geographical entry point to the study of the electoral vote. It put communication channels at the centre and did away with the exclusive focus of the usual context effects on some sort of local level. He started his modelling proposal on the voting decision with an overview of studies of the context effect as done so far, nearly exclusively concerned with some version of the local context that he then put into question though not denying the importance of distance. His own empirical work was also preoccupied with those kinds of context. His methodology was in line with a more broadly supported quantitative preference at the time. This is where Johnston also started and to which he has largely subscribed ever since. It is true that Johnston has later put the usual question marks at the theoretical claims of law-like pronouncements on social reality that early work in that direction often underpinned. But he has always remained a dedicated empirical researcher and has lamented the overly turn away from quantifying work that British geography in particular has characterized during the last decades (e.g. Johnston, 2008, p. 332-335).

15Johnston has continued in the same direction as Cox’s early empirical work on the subject, always extending and refining different mechanisms at work from scale levels not too far away from the individual voter. He has studied them as closely as possible using a wide variety of survey data shared with political scientists and sociologists but also by using official sources with respect to redistricting processes and campaign expenditures for instance. He has always insisted on the role of social conversation with political elements imbued and the purposeful action of political parties particularly in the run up to elections during campaigns. He has certainly not been blind to the wider ranging processes in which voters are implicated from the world system down, but he has given priority to precisely understanding the more immediate circumstances in which people are called to vote. As he has analysed successive elections mostly one at the time his work is largely cross-sectional although he certainly pays attention to temporal features as his number of elections studied lengthens (e.g. on the transition of voters from election to election and on the changing significance of the North-South divide in Britain). In his work on elections there is only a restricted sense of the world outside the national frame.

16Agnew followed Cox in a different way. As Cox in ‘The voting decision…’ he wanted to mark a position distinctive from the main tenets of political sociology and steeped in geography. He was also inclined to do away with the focus on small scale spatial contexts. But whereas Cox put primary stress on communication channels, Agnew aimed at networks in a wider sense. He was certainly not only inspired by Cox’s example, but also by a switch in the epistemological basis underlying much research in human geography at the time (from logical positivism and unreserved quantification to poststructuralist, critical versions accompanied by a more modest ambition with respect to explanation and prediction). His aversion of the contextual effect as studied by political sociology is not only based on empirical but just as much on theoretical grounds. His approach encourages a dominant role for a contextual embedding of the individual voter at all scales. Agnew’s individual voter co-constitutes his/her place.

17Place (re-emphasized by Agnew and accepted by Johnston) is considered a combination of a locale (the setting of everyday life), location (its position within a multitude of wide ranging relevant networks) and a sense of place (attachment, identification). The difficulty remains here that the three dimensions do not come together easily and therefore the notion of the overall concept of place remains elusive. Another problem has to do with the difficulty to imagine a sensible study to see locales in operation as settings for everyday life in sufficient numbers and variety to get any sense of what they do for socialization in place with a view to political commitments. Agnew circumvents the problem in his initial elaboration by insisting that many people are socially and spatially constrained having to do with only a few locales while others have more of them, wider apart, and only jet-setters move from Acapulco to Aspen to Gstaad to Cannes (obviously the times, they may have somewhat changed since) (Agnew, 1987, p. 27). Unfortunately he does not tell us how many more the middle category has, how wide apart they are and how large a part of the population this encompasses. It looks like a plethora of micro-sociological settings is increasingly common, they may be well dispersed spatially and they may even not be place dependent at all (as Webber foresaw in 1964). And this still ignores residential migration (for many people a succession of moves), which was one of Cox’s main concerns, but definitely disappears to the background when ‘socialization in place’ is fore-grounded (Simmel’s concept of sociability that Agnew also introduces in this context lacks the pronounced sense of learning that socialization has).

18In the end Johnston uses a broad theoretical umbrella when he takes Agnew’s place concept to be the basis for his place-based considerations. Agnew in his turn has applied his conceptualization in his study of Italian politics around the turn of the century. His various chapters cover a lot of ground but it is not easy to retain a general thread. It is probably significant that both Johnston’s and Agnew’s book lack concluding chapters giving readers the gist of what the perspective has been able to produce in extra insight. While Johnston & Pattie write separate conclusions at the end of each chapter without any overall final message, Agnew concludes with four ‘themes’ of a rather general order that seem to summarize his findings. The end of contextual studies in electoral research is nowhere in sight.

Electoral geography’s research agenda with context embedded

19In the 1960s the search for and interpretation of context-effects in voting studies emerged as a part of the study of the politics in ‘Western’ democracies at the end of a prolonged period of relatively stable party structures and limited electoral volatility. Just then new movements showed up and there were signs of popular political unrest. It is perhaps telling, also from this perspective, that Cox in 1969 turned to the literature on innovation diffusion to look for guidance in the case of context effects. What about the development and the agenda of electoral geography at large?

20In a recent overview of electoral geography development Leib and Quinton (2011, p. 9) consider its rise until about 1990 (when an earlier overview appeared in Johnston, Shelley & Taylor eds. 1990) and then see it as more and more isolated in a British-American geography increasingly dominated by the cultural turn and poststructuralist approaches. Critics mention positivism, empiricism, “methodological obsession” (Agnew, 1990, p. 15 quoted on p. 9) and a lack of interest for social theory. Reviewing 222 journal articles published on electoral geography in English-language human geography journals between 1990 and 2007, they signal the dominance of the US and the UK as origins of the articles and the hegemony of the journal Political Geography, source of half the items. The main topics they encountered during this period, were voting studies and studies on electoral systems and redistricting, the first pertaining to either the spatial distribution of compositional factors or place-based contextual factors, and their relationship to voting behaviour. Compared to the earlier period the agenda had not altered.

21Electoral geography generally has often been castigated as utterly ‘empiricist’ due to an overwhelming availability of primary data (although unfortunately aggregated, often in unhelpful formats) and a lack of theoretical finesse. This does not mean that electoral geography studies have failed in these regards across the board. Since André Siegfried’s seminal Tableau politique de la France de l’ouest sous la troisième république appeared in 1913, electoral studies, though often surcharged with primary data, have at the same time contributed to deeper insights by close reading of the evidence. In addition, electoral geographers have repeatedly opened up new vistas and broadened the field of electoral studies beyond refined description of results and context effects as conceptualized in the preceding versions. Let us illustrate this point with examples from two themes.

22First of all, attention has been paid to the changing distribution of democratic polities that long have contained popular elections as an essential ingredient and possibly resulting variations in the nature of voting results. John O’Loughlin et al. (1998) has carefully documented the spread of democracy and the reversals toward autocracy across the world during 1946-1994 inserting a major part of these years in the long view as the third wave of democratization (Huntington 1991) with notes of caution added. The extension of institutions that sustain the repeated occurrence of free elections is conditioned by broader internal social developments but there are apparently also external influences in play that result in a diffusion of democracy and its opposite in spatially specific distributions through external agency and regional contagion.

23From the 1980s onward Peter Taylor (1989, p. 240-250 repeated in following editions) has underlined the importance of looking at examples of free elections beyond the set of ‘Western’ democracies to consider the possibility of significant differences in their results due to the strongly different socio-economic conditions in which they occur. Using evidence from India and Ghana (Osei-Kwame, Taylor, 1984) he suggested a much larger volatility due to governing parties’ incapacity to deliver on electoral promises and opposition parties’ unrestrained tendency to engage in utopian wish lists. This tended to result in completely reversed support patterns for parties in successive elections – a situation of ‘democratic musical chairs’. Taylor saw such contrasts as one more example of the differentiating impact of the capitalist world system in which both kinds of countries held different positions. Consequently, in terms of the introduction of free elections as well as their results, contextual factors (position in world system, propinquity to earlier adopters) are again relevant in these cases. These factors, however, work at scales beyond the finally localized contexts that have so prominently been displayed in earlier sections.

24A second interest of electoral geographers has always been the dynamic of electoral change in its short term and long term guises. Already Siegfried was not only insisting on the stability of electoral preferences in the West of France during the first part of the Third Republic. He also paid attention to the temporary incursions of what we would probably now call populist parties and the typical locations in which those incursions were most evident (Siegfried 1913 repr. 2010, p. 587-614). The extreme volatility in voting patterns considered typical for non-core countries by Taylor is one instance. Such volatility has now become somewhat less distinctive as volatility in national elections within long-established liberal democratic political systems in Europe has significantly increased since the 1990s (Mair, 2006). In Agnew’s studies of the Italian electoral landscape (Agnew, 2002), its instability even during the time that party systems were considered ‘frozen’ in the 1950s and 1960s and the shifts that occurred from the 1990s till he finished the manuscript got prime of place.

25Archer and Taylor (1981) demonstrated the near reversal of the electoral fortunes of US political parties across the map in a series of successive periods over nearly one and a half century. This long term dynamism reflected perhaps less the distinctive role of the US in the successive stages of world system evolution as the successive roles of the different parts of the country in its national history even as it became ever more entwined with the outside world (Civil War, industrialization, global city formation). If a single context had to be emphasized here it should be the national one.

26Another instance of a differentiated national scene, background to an incisive electoral shift but now in one sudden jump, is the result of the upsurge of Hitler’s party in the Reichstag elections of 1930 portrayed by O’Loughlin, Flint, Anselin (1994). Next to class and confession the authors emphasize the importance of ‘context’, regional as well as local level contexts apparently, each stormed by to different degrees mobilized regiments of true believers with messages directed to the various population categories and contexts, resulting in different levels of success also depending on the mobilizing capabilities of political opponents. Contexts, in the authors’ view, function as differently configured arenas in which struggles between political forces occur. But it should obviously not be forgotten that the Weimar Republic here clearly also functioned as an overarching context that provided a set of unique figures and circumstances finally leading to this unheard of political catastrophe.

27Later on, O’Loughlin has used the 1930 and also the 1932 Reichstag elections as examples in his contribution to an extensive Book Review Forum in the Annals of the American Society of Geographers (O’Loughlin, 2000, p. 592-601). The Forum was on the celebrated book by Gary King (1997) indicating a new way to deal with the Ecological Inference Problem. The website on the subject is still active and growing:​category/​research-interests/​methods/​ecological-inference. Other contributors were Sui, Fotheringham and Anselin. King’s methodology provided new roads to estimates of individual level relations from aggregate data. Following King’s prescriptions O’Loughlin tried to make new inroads to the resolution of the puzzle regarding the Nazi vote in terms of the cumulation of compositional and context effects in the absence of what he called ‘truth’ facts (individual survey data that get in this reading an undeserved glow of utmost respectability). In the end uncertainties remain but there are additional specific strong indications e.g. for the locally variable mobilization levels of the Nazis.

28Next to Ron Johnston John O’Loughlin has been a key figure in introducing and using spatial statistics in electoral geography. He had entered the field at an early stage (O’Loughlin, 1981). In an overview paper (O’Loughlin, 2003) he reiterated the plea for a systematic analysis of contextual effects per se next to the spatially differentiated compositional effects of sociological, political and economic factors. Whereas political scientists tend to disregard contextual effects, to see them as a residual category and to expect ever more refined models and measurements of sociological, economic and political factors to produce compositional effects to inter place differences, geographers bring the role of place as a specific factor to the fore. Local indicators of spatial association (LISA) combine global measures of clustering with local specifications and thus demonstrate the importance of studying the isotropic and anisotropic processes involved in spatial autocorrelation (O’Loughlin, 2003, p. 37). Geographical weighted regression (GWR) in which location is systematically taken on board the regression analysis (see also Shin & Agnew 2011) and exploratory spatial data analysis (ESDA) are now in use. There has also been some interest in multilevel modelling (Jones, 1991; De Vos, 1997, pp. 87-92, 184-188) that enables researchers to distinguish scalar effects at different levels simultaneously on top of the effects of personal characteristics. Contextual interaction is then part of the analysis and not a residual category. Jones has worked with Johnston and Pattie on its implementation within electoral geography (Jones, Johnston, Pattie, 1992, p. 343-380)

29As for the near future of electoral geography the interest in short term electoral dynamics in so far well established democracies will probably endure if not further increase. Successive pairs of national elections may well continue to show historically unusual levels of volatility at least partly caused by the transformation of the party systems. Traditional parties seem increasingly unable to bridge the gap between their roles as steady representatives of the popular voice and their contribution to actual governing by acting as recruitment channel for political office holders notably. Earlier, Taylor has characterized this type of gap as the politics of power versus the politics of support. Following Mair (2006) the public now becomes increasingly reluctant to become engaged in traditional political parties of whatever colour operating principally in the national arenas as shown in decreasing turnout, diminishing membership and declining more or less stable partisan attitudes. At the same time there seems to be a sustained, perhaps growing sense of civic engagement at the local level (Pattie, Seyd, Whiteley, 2004). Political professionals are increasingly oriented to policy making and actual governing without much continuous attention to their relation with the public and their sympathisers in particular. Mair (2006) suggests mutual reinforcements of these two trends thus undermining the vital functions parties have traditionally played in stabilizing democratic polities. The result is a growing gap in perception between an electoral and a constitutional (‘checks and balances’) form of democracy and increasingly technocratic forms of governance making space for populist parties and flash movements decrying the gap between ‘the elite’ and the ‘population at large’.

30Political geographers interested in elections might do well to concentrate on the identification and analysis of localized social milieus, that may provide particularly fertile ground for such configurations of popular political expression (among other types the present day equivalents of Munich in Winter 1919 so to speak). They might also further focus on the consequences of changes on the supply side of political messaging for the conduct of electoral campaigning and the ways in which campaigns now affect voters in different kinds of local milieus.

31Further contributions for an emerging agenda for electoral geography transcending the enduring focus on national elections comes from the book in which Leib, Quinton (2011) wrote their recent overview on electoral geography already mentioned. As a whole it was meant to help ‘revitalize’ the subject, that is to more closely connect it to the current agenda of human geography at large and to fully incorporate the technical and technological developments of recent years (Warf, Leib 2011). As for the desired changes they apparently had in mind, three developments already partly underway seem pertinent: 1) diversification of topics: from a nearly exclusive emphasis on national elections to a more diverse interest also in local elections and in referenda, and from a focused interest on campaign-election day- results to a more widespread interest in the electoral institution as a whole: pre-registering and postal voting; voting technology; recruitment of candidates; 2) diversification of theoretical perspectives, incorporating political economy and post-structuralism, often in tandem with spatial analysis; and 3) diversification of methods, with an increased technical and methodical sophistication: GIS, spatial statistics and mixed methods. The book has a number of case studies (grouped in a section on electoral geography outside the US - including oddly enough a case study of the vote of the Puerto Rican diaspora in mainland US in the referenda on the status of the island - and a section on electoral geography in the US) demonstrating these various diversifications.

32We close this section with three examples of recent work reflecting the shifting research agenda of electoral geography we encountered. They demonstrate the permanent concern with the contextual notion, bended and transformed as it may have been, and also the grossly changed circumstances in which the electoral institution now functions in long established democracies and elsewhere.

33The first example pertains to mixed method approaches. Brown, Knopp, Morrill have applied them in different studies. In the first one (Brown, Knopp, Morrill 2005) they undertook an exploratory study of an unsuccessful electoral initiative in 2002 to repeal a gay rights ordinance in Tacoma, Washington. They combine a discourse analysis of the local debate based on local and regional reportage on the campaign with a quantitative analysis of demographic and voting data at census block-group/precinct level. The analysis of the Tacoma News Tribune revealed a debate between two broad social goods: equality versus morality. The local ordinance was also “seen as a critical part of the economic restructuring towards a new economy” to attract new investments (in line with Richard Florida’s influential idea about the creative class). Its repeal was expected to hurt business interests. The press coverage displayed a strong divide between African Americans on the one hand and gays plus lesbians on the other (suggesting competing rights), ignoring other groups. A regression analysis of the census data on the voting results identified class indicators and household structure variables (that functioned as proxy for same-sex partnership) as important explanatory factors and to a much weaker extent race - the factor that had turned out to be so important in the discourse. They then performed a series of quantitative analyses that finally resulted in eight clusters of small neighbourhood units with different profiles in terms of voting, household structure, class and race. In their view context remains of eminent importance, but not in the traditional way as patterns of racial and economic segregation but as ideas and expressions of place making: the self-image of Tacoma and its ‘imagined future’ were at stake. More recently the same team of researchers has looked at counties in the presidential races of 2004 and 2008 that voted in anomalous ways with respect to the national trends studied by quantitative analysis . They then focus on these anomalous counties by way of secondary data like local histories, local medias, key informant reports and in some few cases visits to the field (Morrill, Knopp, Brown, 2011).

34The second example is the call for a feminist electoral geography or at least the feminizing of electoral geography which is probably the least feminized subdiscipline of political geography (a point also stressed by Leib, Quinton, 2011, p. 11). Secor (2004) sketches a research agenda for feminist approaches to electoral geography in a volume introducing feminist approaches to diverse domains of political geography (Staeheli, Kofman, Peake, 2004). Starting from the issue of women´s representation in different political systems, and the question of the existence of gender cleavages in different contexts, she also wonders whether women legislators do politics differently, and what gender gaps matter in term of political behaviour . The latter refer to the effects of differential access to mortgage or to credit for consumer goods for example. Finally she discusses how feminist approaches question the importance of formal political behaviour (elections) in the light of other political activities and the construction of the divide between public and private sphere that is characteristic of western democracy. More in general, feminist geographies engage with local contexts, as they promote more grounded and embodied electoral geographies, in line with the geo-sociological approach advocated by Agnew stressing the embeddedness of citizens in places that matter to them.

35Building on feminist geopolitics and emotional geography Schurr (2013) works towards an emotional electoral geography by combining visual ethnography with qualitative interviews to investigate “how emotional campaign practices performatively generate political communities”. In so doing for local electoral campaigns in an Amazon town in Ecuador 2006-2010, she tries to capture how speech and body acts bring into being emotional spaces that contribute to the constitution of political spaces within specific time spatial contexts. Though infrequent, feminist approaches have brought new approaches to electoral studies with a strong accent on local ethnographies.

36Thirdly, we signal the impact of neogeography on electoral geography. Neogeography is a neologism used to describe the unprecedented democratization of the production and the circulation of geographical and cartographic knowledge that occurred with the increasing speed and ease through which making (digital) maps online and on-demand has been made possible through recent technological developments (Shin, 2009). Elections are one of the domains in which neogeography has been particularly visible. As a result, electoral maps are produced and circulated and commented to an extent and at a speed unknown before. As Shin (2009) reminds us the county-by-county map published by USA Today in the days following the closely contested 2000 election was a new phenomenon. Until then such maps were the results of painstaking academic work and were published long after the elections. By the time of the 2008 presidential election, all newspapers and all news broadcast covering the elections published such maps as soon as the results were known, and the mapping of blue and red counties (and purple ones for close calls in forecasts) are now part of popular geographical literacy, even far beyond the territory of the US and the audience of US citizens proper (hence the title of the articles by Morrill et al mentioned above).

37With respect to contextual effects, candidates and parties on the one hand and citizens on the other hand are able to monitor their local and regional context in ways unknown before through online GIS. In the US situation, neogeography has also a significant influence on the awareness of the financing of campaigns, with online tools making it possible to locate the donators/sponsors of the campaigns for specific candidates or referendum propositions. In this manner, the internet has become a contextual factor of its own, that is particularly difficult to conceptualize and theorize, contributing to both a globalization, a nationalization and a localization of electoral behavior, depending on which sites an individual engages with. In comparison it makes the puzzle of residential migration for the conceptualization of socialization in place foregrounded by Cox as discussed above, a relatively easy problem to grasp.

Final note

38Although this paper focused on British and American geographers it is important to stress how porous and indefinite such circumscriptions have become. The disciplinary borders are still standing, they are sometimes emphasized to make a point, but there is a sustained interest in what happens beyond the walls, certainly from the geographers’ side. As to our focus on the British and American exemplars of the species (a fuzzy geographical circumscription with linguistic overtones), quite a few can in fact bear both hats, others share in any case an English language pedigree but ever more other people are otherwise implicated. They write in the English language journals, cooperate in studies British and American researchers do in their respective countries and they may also well engage in studies of subjects situated in the British-American realm.

39Apart from Stein Rokkan, who at the time already transgressed ethno-linguistic as well as disciplinary boundaries with gusto, we came across one recent example of a researcher based outside the British –American realm in Switzerland (also concerned with a subject located elsewhere) and writing in a major English language journal - Carolin Schurr’s engagement with emotional geography. Another quite recent example of writing in the English language journals by researchers based outside the English speaking world , this one fully in line with the traditional interest in contextual analysis, is David & Van Hamme’s (2011) paper in Political Geography. It provides a theoretically subtle, technically advanced insight in the neighborhood effect in a Belgian election. The pillars (a set of connected organizations in different spheres of life) in Belgium organized the three major ideological families (Liberals, Catholics, Socialists) since the early 20th century in a tightly held grip that also included political support. Although alternative and additional formations plus the split between the Flemish and Walloon pillars have re-formatted social life, the effects of the pillars in voting results are still locally visible. The authors trace those local effects of pillars, differentiated for the three pillars, through two mechanisms: via generational transmission of pillar values within families linked as they remain to a specific institution of those pillars (health insurance) and by the general level of local embeddedness of pillars.

40Finally, this paper may hopefully in some small measure help to narrow the divide, earlier mentioned, between the sociological and the geographical imaginations that Durkheim and Vidal articulated and that also showed up outside France. A century ago Siegfried landed already somewhere in between and was only hesitantly and slowly appreciated on both sides. It is about time the study of elections continues without unnecessary divisions.

Haut de page


AGNEW, J., 1987, Place and politics: the geographical mediation of state and society, Boston, Allen & Unwin.

AGNEW, J., 1990, From political methodology to geographical social theory? A critical review of electoral geography, 1960-1987, in Johnston, R., F. Shelley, P. Taylor (eds.) Developments in electoral geography, London, Routledge, p. 15-21.

AGNEW, J. 1996, « Mapping politics: how context counts in electoral geography », Political Geography, 15, 2, p. 129-146.

AGNEW, J., 2002, Place and politics in modern Italy, Chicago, University of Chicago Press.

ALLARDT, E. & ROKKAN, S. (eds.), 1970, Mass politics; studies in political sociology, New York, Free Press.

ARCHER, J., & TAYLOR P. 1981, Section and party: a political geography of American presidential elections, from Andrew Jackson to Ronald Reagan, Chichester, Research Studies Press.

BROWN, M., KNOPP, L. & MORRILL R., 2005, « The culture wars and urban electoral politics: Sexuality, race, and class in Tacoma, Washington », Political Geography 24, 3, p. 267-291.

COX, K., 1969, « The voting decision in a spatial context », Progress in Geography, 1, p. 83-117.

DAVID, Q., & HAMME, G. van, 2011, « Pillars and electoral behaviour in Belgium: the neighbourhood effect revisited », Political Geography, 30, 5, p. 250-262.

DOGAN, M., ROKKAN, S. (eds.), 1969, Quantitative ecological analysis in the social sciences, Cambridge Mass., MIT Press.

FLORA, P., 2000, Stein Rokkan. Staat, Nation und Demokratie in Europa. Die Theorie Stein Rokkans aus seinen gesammelten Werken rekonstruiert und eingeleitet von Peter Flora, Frankfurt am Main, Suhrkamp.

HUNTINGTON, S. P., 1991, The third wave: democratization in the late twentieth century, Norman OK, University of Oklahoma Press.

JOHNSTON, R., SHELLEY, F., & TAYLOR, P. (eds.), 1990, Developments in electoral geography, London & New York, Routledge.

JOHNSTON, R., & PATTIE, Ch. 2006, Putting voters in their place. Geography and elections in Great Britain, Oxford, Oxford University Press.

JOHNSTON, R., 2008, « Quantitative Human Geography: are we turning full circle? » Geographic Analysis, 40, 3, p. 332-335.

JOHNSTON, R., 2010, « Book review essay: What every human geography student needs to read to know the discipline? » Progress in Human Geography, 34, 4, p. 528-535.

JOHNSTON, R., PATTIE, Ch., 2012, Kevin Cox and electoral geography, in Jonas, A., A. Wood (eds.) Territory, the state and urban politics: a critical appreciation of the selected writings of Kevin R. Cox, Burlington VT, Ashgate, p. 23-44.

JONES, K., 1991, « Specifying and estimating multi-level models for geographical research », Transactions of the Institute of British Geographers. New Series 16, p. 148-160.

JONES, K., JOHNSTON R., & PATTIE, Ch., 1992, « People, places and regions: exploring the case of multilevel-modelling in the analysis of electoral data », British Journal of Political Science, 22, p. 343-380.

KING, G., 1996, « Why context should not count », Political Geography, 15, 2, p. 159-164

KING, G., 1997, A solution to the ecological problem: reconstructing individual behaviour from aggregate data, Princeton, Princeton University Press.

LEIB, J., N. Quinton, 2011, On the shores of the ‘moribund backwater’? Trends in electoral geography research since 1980, in Warf, B., J. Leib (eds.) Revitalizing electoral geography. Farnham, Ashgate, p. 9-27.

MAIR, P., 2006, Polity-scepticism, party failings and the challenge to European democracy. Uhlenbeck Lecture 24, Netherlands Institute of Advanced Studies, Wassenaar.

MORRILL, R., KNOPP, L. & BROWN, M., 2011, Anomalies in red and blue II: Towards an understanding of the roles of setting, values, and demography in the 2004 and 2008 U.S. presidential elections, Political Geography, 30, 3, p.153-168.

O’LOUGHLIN, J., 1981, The neighbourhood effect in urban voting surfaces: a cross-national analysis, in Burnett, A., & P. Taylor (eds.) Political studies from spatial perspectives: a cross-national analysis, Chichester, John Wiley, p. 357-388.

O’LOUGHLIN, J., FLINT, C. & ANSELIN, L., 1994, « The geography of the Nazi vote: Context, confession and class in the Reichstag elections of 1930 », Annals of the American Society of Geographers, 84, 3, p. 351-380.

O’LOUGHLIN, J., WARD, M., LOFDAHL, C., COHEN, J., BROWN, D., KEILLY, D., GLEDITSCH, K. & SHIN, M., 1998, « The diffusion of democracy 1946-1994 », Annals of American Society of Geographers, 88, 4, p. 545-574 .

O’LOUGHLIN, J., 2000, « Can King’s Ecological Inference Method answer a social scientific puzzle: who voted for the Nazi party in Weimar Germany? », Annals of the American Association of Geographers, 90, 3, p. 592-601.

O’LOUGHLIN, J., 2003, Spatial analysis in political geography, in Agnew, J., K. Mitchell, G. Toal (eds.), 2003, A companion to political geography, Oxford, Blackwell, p. 30-46.

OSEI-KWAME, P., & TAYLOR, P., 1984, A Politics of Failure: The Political Geography of Ghanaian Elections, Annals of the Association of American Geographers, 74, 4, p. 574-589.

PATTIE, Ch, SEYD, P. & WHITELEY, 2004, Citizenship in Britain: values, participation and democracy, Cambridge, Cambridge University Press.

SCHURR, C., 2013, Towards an emotional electoral geography: the performativity of emotions in electoral campaigning in Ecuador, Geoforum, 49, p. 114-126.

SECOR, A., 2004, Feminizing electoral geography, in Staeheli, L., E. Kofman, L. Peake (eds.), Mapping women, making politics; Feminist perspectives on political geography, New York, Routledge, p. 261-272.

SHIN, M., 2009, Democratizing electoral geography: Visualizing votes and political neogeography, Political Geography, 28, 3, p.149-153.

SIEGFRIED, A., 1913, Tableau politique de la France de l’Ouest sous la Troisième République, Paris, A. Colin (repr. by Editions de l’Université de Bruxelles 2010).

SMETS, K., HAM, C. van, 2013, The embarrassment of riches? A meta-level research on voter turnout, Electoral Studies, 32, 2, p. 344-359.

STAEHELI, L., KOFMAN, E. & PEAKE, L. (eds.), 2004, Mapping women, making politics; Feminist perspectives on political geography, New York, Routledge.

TAYLOR, P., & JOHNSTON, R., 1979, Geography of elections, Harmondsworth, Penguin.

TAYLOR, P., 1985, Political geography. World-economy, nation-state and locality, Harlow, Longman.

VOS, S. de (1997) De omgeving telt. Compositionele effecten in de sociale geografie. PhD UvA Amsterdam.

WARF, B., & LEIB, J. (eds.), 2011, Revitalizing electoral geography, Ashgate, Farnham.

Haut de page

Pour citer cet article

Référence électronique

Herman van der Wusten et Virginie Mamadouh, « “It is the Context, Stupid!”. Or is It? British-American Contributions to Electoral Geography since the 1960s »L’Espace Politique [En ligne], 23 | 2014-2, mis en ligne le 02 juillet 2014, consulté le 27 septembre 2023. URL : ; DOI :

Haut de page


Herman van der Wusten

Professor em.
AISSR/Department of Geography, Planning and International Development, University of Amsterdam

Articles du même auteur

Virginie Mamadouh

Associate professor
AISSR/Department of Geography, Planning and International Development, University of Amsterdam

Articles du même auteur

Haut de page

Droits d’auteur


Creative Commons - Attribution - Pas d'Utilisation Commerciale - Pas de Modification 4.0 International - CC BY-NC-ND 4.0

Haut de page
Rechercher dans OpenEdition Search

Vous allez être redirigé vers OpenEdition Search