Changing spaces of political encounter and the rise of anti-politics: Evidence from Mass Observation's General Election diaries

Negativity towards the institutions of formal politics is currently a concern across much of the demo- cratic world. It is generally agreed that such negativity increased among British citizens during the second half of the twentieth century. In this paper, we analyse a novel dataset not previously used to study this topic: Mass Observation's General Election diaries. Since diarists wrote mostly about politicians, political campaigns, and associated media coverage, we ask speci ﬁ cally what the diaries can tell us about increased negativity towards politicians and its relationship to developments in political communication. We take a postholing approach to sampling of the diaries, enabling comparative-static analysis between the middle and end of the twentieth century. We view the diaries in a geographical framework derived from contextual theories of social action. This gives us a focus on spaces of political encounter, modes of political interaction, performances by politicians, and judgements by citizens. We argue that prominent spaces of political encounter changed over the period from long radio speeches and rowdy political meetings to televised debates and associated expert commentary. We demonstrate how these latter settings for political interaction afforded less opportunity for politicians to perform virtues to citizens, and for citizens to calibrate judgements of politicians.


Introduction
'Anti-politics' has been used to describe many things, one of which is citizens' negativity towards the institutions of formal politics (Clarke, 2015). This negativity can be observed e or heard e in talk data from interviews and focus groups, or, more indirectly, in survey data on things like confidence, trust, approval, and satisfaction with such institutions. This negativity should be disaggregated by 'object of political support' (Norris, 1999;following Easton, 1965). Citizens may feel disaffection towards politicians, or parties, or parliament, or government. Scholars have identified anti-politics as an important phenomenon of the current period in parts of Europe, North America, Australasia, and elsewhere (e.g. Boswell & Corbett, 2015;McDowell, Rootham, & Hardgrove, 2014;Saunders, 2014). We need to understand more about where this phenomenon came from.
It is widely accepted that anti-politics became more prevalent in many democracies during the second half of the twentieth century (Dalton, 2004;Norris, 1999;Nye, Zelikow, & King, 1997;Pharr and Putnam 2000;Torcal & Montero, 2006). For countries like Britain, this was a period of transformation from a time of relative political support immediately after the Second World War to a time of relative political disaffection since the late twentieth century (Stoker, 2016). It is also widely accepted that such a long-term and complex historical development is likely to be explained by multiple factors. Citizens changed during this period. They became wealthier, better educated, and more critical (Inglehart, 1997;Norris, 1999); less aligned to the main parties (Dalton and Wattenberg 2000); and more consumerist in their approach to politics (Stoker, 2006). Politics also changed. Over the period, governments performed less well against an expanded set of criteria (Mulgan, 1994). Power was distributed away from national governments and towards other actors (Hay, 2007). Politicians and parties became less distinguishable in ideological terms (ibid). Finally, political communication changed. Politics became increasingly mediated and journalists increasingly framed politics in negative terms (Cappella & Hall Jamieson, 1997). Political campaigning became nationalised, professionalised, and increasingly focused on controlled situations (Lawrence, 2009).
In a global study, Norris (2011) tested such 'demand-side', 'supply-side', and 'intermediary' factors, and found that all of them, to some extent, help to explain patterns of political support (or, more accurately, the withdrawal of political support). The current challenge, therefore, is not so much to sort these explanations further e though productive debate continues on this point (see Baldini, 2015;Marsh, Vines, & Halupka, 2016;Richards & Smith, 2015) e as to understand more about each factor and how it works.
In this paper, we analyse General Election diaries kept by volunteer writers for Mass Observation (MO). During the period in question e what we might call the 'long' second half of the twentieth century, from the end of the Second World War, through the post-war period, through the late twentieth century, to the attacks of 11 September 2001 e such diaries were kept on three occasions: 1945, 1987, and 2001. We argue that such diaries help us to understand more about anti-politician sentiment in Britain (since most of the writing in the diaries is about politicians, as opposed to the other institutions of formal politics) and its political communication explanations (since most of the writing in the diaries is about how citizens' received political campaigning and associated media coverage during the period). In doing so, these diaries help us to answer Corbett's (2014Corbett's ( , 2015 call for more research on antipolitician sentiment and particularly its history, including questions of continuity and change. We draw on geographical insights from contextual theories of social action (Thrift, 1983(Thrift, , 1996 to argue that settings or locales in which politicians and citizens encounter one another e 'spaces of political encounter' e are important for the kinds of political interaction they shape and the kinds of performances by politicians and judgements by citizens they afford. We argue that prominent spaces of political encounter changed during the second half of the twentieth century. Long radio speeches and rowdy political meetings became less prominent. Televised debates and media reporting of polling results and expert analysis became more prominent. The strength of the MO diaries is that we can see in them how such changed spaces of political encounter were related to changes in modes of political interaction, performances by politicians, and judgements by citizens (including a move to more negative judgements regarding politicians).
In making these arguments, we aim to supplement existing studies of electoral geography. We study the relationship of political campaigning not to voting for particular parties (e.g. Cutts, Webber, Widdop, Johnston, & Pattie, 2014;Johnston, Pattie, Scully, & Cutts, 2016), nor to voter turnout (e.g. Fisher, Fieldhouse, Johnston, Pattie, & Cutts, 2016), but to judgements of politicians in general (as one object of political support). We also study this relationship not for the current period, during which constituency campaigning has made something of a return, but for the second half of the twentieth century e that long period of increase for anti-politics in Britain, and a period characterised by the nationalisation of political campaigning.
We return to this literature, the current period, and the question of constituency campaigning in the concluding section of the paper. But let us clarify our main argument at this point. We argue that prominent spaces of political encounter changed between the immediate post-war period and the late twentieth century. A part of this argument is that political campaigning became nationalised. Local political meetings became less prominent. National media campaigns and associated coverage became more prominent. But this is not the full argument. After all, a prominent space of political encounter in the earlier period was the speech on BBC radio. Our argument is more that political interaction became increasingly mediated and indirect during the second half of the twentieth century. Rowdy political meetings allowed citizens to challenge politicians directly. Radio speeches may not have allowed this, but let politicians speak relatively directly to citizens for quite some time e in a way that exposed them and their programmes, and so challenged politicians in a different way. By contrast, televised debates at the end of the century involved questions and interruptions by journalists and other politicians, so that politicians would usually only have to speak for a short period of time and could get away with avoiding topics or not answering questions. Associated media coverage now also gave less voice to politicians and citizens, and more to journalists, pollsters, and expert analysts.
In making these arguments, we also seek a contribution to the revitalisation of electoral geography (Leib & Warf, 2011). We contribute more social theory by our engagement with contextual theories of social action. We contribute a conceptualisation of space as context, setting, situation, locale. We contribute a new qualitative dataset: MO's General Election diaries. We contribute a new topic of concern through our focus on the geography of political campaigning as it relates to anti-politics (as opposed to, say, electoral success). Finally, we contribute a post-positivist form of argumentation that works towards empirical plausibility e as opposed to proof e by demonstrating 'logical connections among phenomena which can be described concretely' (Sennett, 1977, p. 43). We discuss this approach further below. But first, we review and extend the relevant literature on political communication and interaction.
Political communication, political interaction, and spaces of political encounter As we have seen, the rise of anti-politics describes a long-term development and complex problem likely to be explained by many interconnected factors. Some of these factors have been termed political-communication or intermediary factors. Research in this field has focused on how politics became increasingly mediated during the second half of the twentieth century, how media came to frame politics in negative ways, and how this framing came to have negative effects on political support among citizens (e.g. Cappella & Hall Jamieson, 1997). Alternatively, research has focused on how political campaigning became modernised during this period (e.g. Rosenbaum, 1997). In a context of limited candidate spending, the expanded franchise, and changing media, political campaigning moved from the local to the national scale; from uncontrolled meetings to controlled press conferences, rallies, and photo opportunities; and from a focus on party platforms to a focus on personalities (especially party leaders). Political campaigning became professionalised and dependent on polling, marketing, advertising, and public relations. As such, it became more negative (because 'knocking copy' has been shown by these professionals to work), more focused on agenda-setting (with certain issues purposefully avoided), and more targeted on floating voters in marginal seats (to the exclusion of other voters).
The close relationship between these two sets of developments e in media coverage of politics and political campaigning by parties e is captured by the concepts of mediatisation (see Str€ omb€ ack, 2008) and political communication (see Blumler & Kavanagh, 1999). Media became the most important source of political information for citizens (the mediation of politics). Then media developed their own commercial logic (simplification, polarisation, personalisation, visualisation etc.). Then political actors began adapting to this media logic (often reluctantly at first). Then political actors internalised this media logic e for example, by valuing policies in terms of their newsworthiness and potential for explanation and justification within media formats. The result was movement from a party-dominated political communication system after the Second World War, to a limited-channel, nationwide, television-dominated system from the 1960s, to a new era of political communication since the 1990s e characterised by conflicting forces including professionalisation of political advocacy and increasing competitive pressures on journalists (ibid).
Much of this literature focuses on structural change in the media and political campaigning sectors. There has been less research on how structural change affects micro-scale political encounters between politicians and citizens, and how such encounters affect politicians' performances to citizens and citizens' judgements of politicians. Here, the concept of political interaction is useful, as used most notably by Jon Lawrence. For Lawrence (2009), the nomination hustings of the eighteenth and nineteenth centuries brought candidates face-to-face with an often irreverent, disrespectful public. They were physical ordeals that 'tested the mettle' of politicians. For citizens, they were fun and entertaining. In the name of accountability, politicians had to humble themselves before their heckling, mocking, derisive constituents. They had to display 'the common touch' and be 'a good sport'. During the twentieth century, this 'spirit of the hustings' was gradually replaced by party organisation, mobilisation of core supporters, and selective campaigning. Television became more important and local face-to-face campaigning was replaced by national mediated campaigning. Political interaction was reduced to choreographed and managed photo-opportunities, press conferences, and rallies.
It is this concept of political interaction that we seek to extend in this paper. Theoretically, it can be extended by drawing on geographical insights derived from contextual theories of social action. The category of 'contextual theories of social action' comes from Thrift's (1983) early work on the determination of social action. For Thrift, time and space are central to the construction of social action. Human activity is not only compositionally determined e by families, schools, workplaces etc. e but also contextually determined by the immediate spatial and temporal setting. These settings might be thought of as locales: containers of opportunities and constraints on action; contextual fields made up of geology, hydrology, climate, organised production, social divisions, forms of the state, forms of sociability, forms of knowledge, and so on (ibid).
'Locale' was a key concept of the early 1980s. Giddens (1984) viewed space as a medium through which social relations are produced. It provides settings or contexts or situations e locales e for interaction. These are both enabling, providing resources like knowledge on which action can be based, and constraining, providing a limited milieu of rules and institutions in which knowledge can be interpreted and used. So people act within locales, drawing on knowledge and working within institutional rules. Of course, they also interpret, reproduce, and change that knowledge and those rules (structuration).
Since the early 1980s, these ideas have been developed in various ways. For example, there has been a greater focus on the spatial extensivity of contexts (Thrift, 1996). In a globalising world, characterised by new information and communications technologies, containers of opportunities and constraints e that still might be called locales, or at least translocales e are often stretched across space (as with mediated politics). Despite these developments, or perhaps because of them, the central insight of contextual theories of social action remains highly relevant today. Contexts are productive and constitutive. They provide resources and affordances. They provide orientations to action (ibid). We might also say: they provide orientations to interaction. In this paper, we focus on the settings or situations providing orientations to political interaction e between citizens and politicians e which we term 'spaces of political encounter'.
We define political interaction as communication between politicians and citizens (speaking, listening, performing, judging etc.). And we define spaces of political encounter as contexts that may be more or less spatially intensive e as in the local political meeting or the televised political debate e and shape political interaction by providing resources, knowledge, rules, affordances, and orientations to citizens and politicians. In the following sections, we extend these concepts empirically, presenting evidence for: 1) changing spaces of political encounter; 2) associated changing modes of political interaction; 3) associated changing performances of virtue by politicians; and 4) associated changing judgements of politicians by citizens. This evidence comes from volunteer writing in the Mass Observation Archive.

The potential and limitations of Mass Observation
Mass Observation (MO) was a social research organisation established in 1937 to record the everyday lives of ordinary people in Britain. In its original incarnation, MO collected material by two general means: a team of 'mass observers' who recorded observations, overheards, survey responses, interview responses, and ephemera between 1937 and 1960; and a panel of volunteer writers, between 400 and 1000 strong depending on the year, who kept monthly diaries (1939e65), completed day surveys (1937e38), and replied to quarterly open-ended questions or 'directives' (1939e55). In 1981, the Mass Observation Archive founded the Mass Observation Project, reviving the panel of volunteer writers (last used in 1955). To this day, directives are still being sent three times a year to approximately 500 respondents.
Let us describe and justify our approach to sampling and analysis by addressing three main limitations of the MO material. First, the observations, overheards, and ephemera collected by the inexperienced, untrained mass observers e in an era before the professionalisation of social science in Britain e probably tell us as much about the prejudices of these mass observers as they do about the everyday lives of ordinary people in Britain (MacClancy, 1995). For this reason, we focused on the panel of volunteer writers and their responses to MO directives e what Sheridan (1994) calls 'the most unmediated layer' of the archive. Second, the social constitution of this panel has often been criticised. For Jeffrey (1978), the original MO was a social movement of the radicalised lower middle class. For (Hinton, 2013), while not all the original panellists were lower middle class, that group was certainly over-represented, along with people from London and the South East, and people of the Left. To address these concerns, we sampled within the panel, following the example of Salter (2010). We sampled 60 respondents for each directive we looked at, seeking to fill quotas for age group, gender, region, and occupational category (the four classifications made possible by MO records). Ultimately, we sought to include a range of people with a range of social and geographical positions in British society.
Importantly, our analytical approach also lessened the importance of the social constitution of the panel, as it did for Nettleton and Uprichard (2011) who used a similar approach. Almost by definition, and regardless of social position, the MO panellists constitute a rather strange group of individuals who volunteer for a social history project e and so are particularly dutiful, engaged, reflexive, and critical (Hinton, 2013). As such, we did not approach them as representatives of people in general; or of particular genders, age groups, occupational categories, or regions. Rather, drawing on cognitive anthropology (see Lakoff, 2002) and poststructuralist theory (see Fischer and Gottweis 2012), we sought to establish the cultural resources e the categories, storylines, and folk theories e panellists use to construct and express their understandings, expectations, and judgements. And we focused on the cultural resources they share with each other, regardless of social background, and, plausibly, with other citizens too in their families, friendship networks, workplaces, and so on.
A third limitation of the MO material is that MO has not collected data from panellists continuously since 1937. Rather, there was a gap between 1955 and 1981. As such, the material enables comparative-static analysis, as opposed to diachronic analysis (Hay, 2002). This form of analysis is commonly used in research concerned with historical change but lacking full timeseries data. As with any form of analysis, it has its own strengths and weaknesses. Specifically, it helps to establish the extent and direction of change between two distinct periods (e.g. the immediate post-war period and the late twentieth century), but not the pace of change over an extended period (e.g. the second half of the twentieth century).
In this paper, we follow the postholing approach to comparative-static analysis, particularly as practised by Sennett (1977). We chose General-Election years that were possible to choose because MO had asked panellists to keep diaries during those campaigns. For the (long) second half of the twentieth century, this gave us a choice from 1945, 1987, 1992, 1996, and 2001. Then we chose a set of years e or postholes e that would be small enough to allow in-depth analysis of what is qualitative data, and large enough to capture different moments or periods of political communication. 1945 was the only General-Election year during the immediate post-war period when panellists were asked to keep diaries. It serves to tell of political interaction in what Blumler and Kavanagh (1999) term the party-dominated political communication system of the period. 1987 was the next time MO asked panellists to keep General-Election diaries, by which time the system had become dominated by television. In this paper, 2001 serves to tell of the situation at the end of the twentieth century and also the beginning of Blumler and Kavanagh's third age of political communication.
It should also be acknowledged that diaries analysed in this paper tell of political communication specifically during General Election campaigns. In one sense, these periods are exceptional. Outside of them, citizens and politicians might encounter each other in other contexts e.g. constituency surgeries. But in another sense, General Election campaigns are not so unusual. There are unofficial 'long' campaigns that spill well beyond the official 'short' campaign period (there are even so-called 'permanent' campaigns). In the wider project from which this paper arose, we also analysed other writing from MO (i.e. writing on the topic of politics not contained in General Election diaries and written at other times). In general, panellists wrote of the same kinds of encounters and interactions, whether in General Election diaries or in this other writing. Taking the example of constituency surgeries, these were not mentioned once in over 2000 pages of writing by the 840 panellists in our sample (only around 400 of which constituted the General Election diaries). All this leads us to make claims on the basis of the diaries analysed in this paper about political communication during General Election campaigns, but also, more tentatively, about political communication in general.
Finally, as mentioned in the introduction, we followed Sennett (1977) by using postholes but also focusing on the logical connections between phenomena made visible by the diaries. In our framework taken from contextual theories of social action, this gave us a focus on relationships between contexts of political encounter, modes of political interaction, the performances of politicians, and the judgements of citizens. These relationships can be observed in the MO diaries, where diarists wrote of encountering politicians in particular contexts, interacting with them according to the rules, norms, and resources provided by those contexts, and their impressions of politicians formed on the basis of particular encounters, interactions, and performances. We now consider these relationships for each of our postholes.
1945: long radio speeches, rowdy political meetings, good speakers, and best candidates One view of 1945 is that it was a particularly unusual General Election. It was the first to be held for a decade. The war had not quite finished, the register was out of date, and turnout figures are generally thought to be unreliable and incomparable (Denver, Carman, & Johns, 2012). Labour won its first ever overall majority. The war had apparently moved citizens to the left because of the suffering and sacrifice they shared (or should have shared) and the government controls that were perceived to have worked (Pugh, 1982). But there is another view of the 1945 General Election (Fielding, Thompson, & Tiratsoo, 1995). Labour won a landslide of seats but not votes. They did so for some of the 'usual' reasons. Voters were disappointed with the incumbents, who they associated with depression, appeasement, war, and foot-dragging over implementation of the Beveridge Report. Labour campaigned in response to the practical concerns held by voters regarding wages and housing.
If there are multiple views of the 1945 General Election, there is one thing on which most commentators agree: that 1945 was 'the radio election', when 45% of citizens listened to nightly political broadcasts (Rosenbaum, 1997). But what did citizens make of these broadcasts and the other means by which they encountered politicians during the campaign? In 1945, MO asked panellists to 'report at intervals on the election campaign' (Directive SxMOA1/3/ 86 e see Appendix A). The diaries they returned describe a range of political encounters and interactions. Most of the panellists received election addresses, leaflets, and pamphlets through their letterboxes. Many were visited at home by canvassers. Newspapers make regular appearances in the diaries. But the two most prominent means by which panellists encountered politicians in 1945 e about which panellists wrote most frequently and at greatest length e were speeches on the radio and local political meetings.
Panellists wrote of listening to and hearing politicians on the radio: 'I turned on to listen to Mr E Brown last evening' (Panellist 1980 e see Appendix B); 'listened in to all the wireless talks up to now' (2576); 'I've just heard Churchill's second broadcast' (1346); 'heard Eden last night' (3207). They wrote of attending political meetings outside pubs; at the local works; on greens, squares, and market places; in schools, Co-op halls, town halls, gardens, village rooms, and ambulance halls. Alternatively, they wrote of failing to attend such meetings in a way that suggests attendance was something approaching a norm of the period: 'I really should be attending more meetings' (3207); 'the girl conductor on the bus was bewailing the fact that she had not been able to get to any election meetings owing to late duties' (1048).
What did these contexts afford by way of political interaction? They allowed for the testing of politicians, for politicians to demonstrate virtues (and vices), and for citizens to know, judge, and distinguish politicians. Long radio speeches were a test for politicians. Consider the following two diary entries: My wife felt that half an hour is too long, or anyway, very few speeches were good enough to last that length of time. This was shown by Sir W Beveridge's speech. In many ways it was good and I was biased in that long before hearing it I expected it to be good, but it was too long. It seems difficult to make a well-knit speech ranging over a variety of topics and perhaps there should have been more concentration on one thing. (1165).
Not much heard about Tom Johnson. His delivery was rather hard on the ear, although his matter was good […]. Sinclair is speaking on the liberal policy as I write. His delivery is vile and irritating with too much emphasis, mostly in the wrong places. Nothing but platitudes and nothing to offer. Earnest Brown was another washout. Nothing to say and neither personal charm nor sincerity […]. Sinclair still speaking. Small men cannot get over on the air. Neither can insincere men. The voice cannot be disguised in a long talk and character comes out. (3648).
These nightly radio broadcasts asked politicians to speak for 30 minutes or more with little interruption. They exposed politicians on two fronts. 'Delivery' had to be easy on the ear; neither monotonous nor irritating. 'Matter' had to describe a well-knit argument and not just platitudes. Over the course of these speeches, politicians could demonstrate their character (or lack of character).
Political meetings were still more of a test. These meetings were participatory for citizens who laughed, applauded, and donated money, but also moaned, booed, jeered, heckled, shouted, and howled. Reports of rowdy meetings appear in many of the diaries: The meeting was literally up-roarious. No slight, innuendo, misrepresentation, or sneering remark was allowed to go unchallenged […]. Another speaker, announced as an industrialist, was heckled about profits and cartels […]. Any reference to Churchill being indispensable was greeted with moans of dissent […]. The Tory candidate came in and there were some boos for him […]. My hands were sore with clapping and my face was still with laughter. (1048).
Last evening, I attended a meeting at the town hall […]. There were quite lively questions asked. At one time it got hot and one young man started to attack another and had to be called to order by the chairman […]. One local man kept the candidate arguing about the question: was Mulberry produced by government or private enterprise? He said private enterprise. (I heard today that probably he had had a drop and that made him talkative). (1980). Went to Moore's meeting tonight […]. I arrived at the hall near the end of Moore's speech. The local miners in a solid block at the back of the hall were giving him a hard time. They were rude and occasionally funny […]. Moore made as much as he could of the Attlee/Laski business but at every reference to it there were loud howls of 'Beaverbrook'. (3207).
Politicians were challenged at these meetings. Citizens would ask questions, express their views, fight among themselves e their confidence boosted by alcohol in some cases. From the perspective of the previous century, these meetings may have seemed less rowdy and more sober (Lawrence, 2009). But from the perspective of the late twentieth century, as we shall see, they appear to be relatively participatory and challenging contexts for political interaction.
Politicians were challenged in such contexts, but were also able to pass the test and demonstrate virtues. One prototypical category of the time was 'the good speaker': a type of politician who thrived on the long radio speech and the rowdy political meeting. Politicians were good speakers when they communicated policy in a pleasant voice: 'I heard Lord Woolton speak on the radio and thought it the best speech on the Conservative side as he did not abuse anyone but said why he had joined Churchill's Government and what the future policy would be. He is an excellent speaker with a very pleasant voice' (3426). They were good speakers when they refrained from abuse and mud-slinging, communicated reason and authenticity, and answered questions: Liberal meeting, town hall, Hunstanton. Began with excellent speech by a young women who had social work in the East End of London. Then speech by candidate Penrose. Both very authentic e Beveridge and a rising party and very reasonable. Absence of mud-slinging […]. Good question intelligibly answered and admission of difficulty. Penrose a good debating speaker. (2794) By contrast, there were bad speakers who babbled, seemed to have no policies, and failed to answer questions. Consider this report of a Conservative meeting at the town hall in Hunstanton: Chairman tried to close without questions. They had had enough experience of heckling with the RAF as reported by The News Chronicle, but we did not. Barrage of questions from soldiers followed. McCullough could not answer them but avoided them quite clearly. Had to admit he has no policy at all […]. McCullough a poor speaker. (2794) Or consider this report of another town-hall meeting: At 5.30pm I went in to the Town Hall and found Sir Bedford Dorman in the chair and just starting to speak. He made a babbling speech which I in the gallery found difficulty in hearing […]. By this time, the Labour and Commonwealth people poured in and began to heckle. The poor speaker was getting more and more hot and bothered when the candidate arrived amid mixed clapping and booing. I had been told that Sir Thomas Dugdale was not a good speaker but really he did not do at all badly and was easily heard. (3426) This panellist was able to distinguish between good and bad speakers, and to be impressed by good speakers (even when not primed to be so). This is the final point to make about political interaction during the 1945 campaign: long radio speeches and rowdy political meetings allowed citizens to know, judge, and distinguish politicians. They allowed the virtues of some politicians to be heard. 'I did listen to the end of Stafford Cripps and thought him very good. What a pleasant voice he has and he speaks with sincerity and conviction' (1980). 'Samuel was very able and made a good case for liberalism ' (3310). 'Went to the liberal meeting in the evening […]. Candidate didn't turn up until 9.15 […]. However, he gave a straightforward, unpretentious speech e he was an unpretentious sort of man' (3351). Some politicians could be judged on the basis of these political encounters as sincere, able, unpretentious; in possession of a pleasant voice, conviction, and a good case. They could also be distinguished from lesser politicians. For one respondent: 'I heard last night the best speech over the radio that I have heard in this contest. Noel-Baker' (1980). For another panellist: 'Samuel impressed well as the best speech yet and I heard two people say they would vote Liberal as a consequence' (1325). For a third respondent: 'As regards personality, I feel that Dr Taylor is by far the best of the three candidates. At the only political meeting I attended he impressed me very much by his obvious sincerity and high standard of values' (2675).
It was possible in 1945 to describe politicians in comparative and superlative terms. There were good and bad speakers, better and worse speeches, and best candidates. We can see in the diaries of MO panellists that such judgements e often positive e were possible at least in part because of logical connections between spaces of political encounter, modes of interaction, and performances of politicians. Citizens encountered politicians most prominently through speeches on the radio and local political meetings. These settings afforded speaking on the part of politicians and listening, hearing, reacting, and challenging on the part of citizens. This political interaction tested the material and delivery of politicians. It oriented them to performances of virtue e of sincerity, ability, character e by which citizens could judge and distinguish them. How would any of this change by our next posthole?

1987: professionalised campaigns, media overkill, and fed-up citizens
Between 1945 and 1987, much changed in political communication (Kandiah, 1995;Rosenbaum, 1997). The BBC television service recommenced in 1946. The 1955 General Election was labelled 'the first television election' by journalists at the time e with reason, in that viewing figures at least matched listening figures during the campaign. Initially, from 1951, television just showed party political broadcasts. But the Television Act of 1954 made provisions for commercial television and the new actors this brought pushed at existing rules and conventions. The Conservative Party Conference was covered by television in 1954. The '14-day rule' e banning coverage of issues to be debated in Parliament within a fortnight e was allowed to lapse in 1956. Two years later, the Granada Network was the first to cover an election campaign (the Rochdale by-election of 1958).
Politicians responded to this increased television coverage with media training. The Conservatives established a broadcasting school in 1950 and, in 1952, a mock television studio at Central Office. After a period when control seemed to be lost e the era of 'ordeal by television' (Rosenbaum, 1997) e politicians wrestled back control of political communication. They employed pollsters and advertisers. They used press conferences to set agendas. Their rallies became ticketed and free of questions from the floor, at first to stop journalists from foregrounding hecklers in their reports, but later in response to security concerns after the murder in 1979 of Conservative Northern Ireland spokesman Airey Neave.
Professionalised campaigning was a key theme of the 1987 General Election (Crewe & Harrop, 1989). The Conservatives and Labour were thought to have run professionalised campaigns characterised by integrated communications and effective media events. Labour was thought to have run an especially good campaign, well-controlled from the centre by Gould and Mandelson's newly formed Shadow Communications Agency, and shaped by a strategy of 'one-issue-per-day' (to keep the focus on social issues e Labour's main strength). Still, when the votes were counted, Thatcher's Conservatives won a third term in office. Economic issues had proven decisive, as had divisions both within the Labour Party, between Kinnock and Militant, and between Labour and the Alliance e the third party of Britain's new three-party system (Butler & Kavanagh, 1988). The Conservatives won 43.3% of the vote on a 79.5% turnout (adjusted e see Denver et al., 2012) e an expected turnout for the 1980s, well below the high of 1951 (90%), and well above the low of 2001 (61.8%).
The 1987 campaign was so professionalised that journalists complained at the time of all-ticket rallies, stage-managed photoopportunities, and a refusal by politicians to debate the key issues (Butler & Kavanagh, 1988). But how did citizens perceive the campaign? How did MO panellists respond when asked to keep 'a log', to 'make notes on the campaign', and to 'report on [their] experience' (SxMOA2/1/22)? The panellists wrote of receiving leaflets, newsletters, and other literature through their post-boxes. But many of them commented on the lack of canvassers, meetings, posters, stickers, and ribbons. A storyline appears to have circulated at the time that local campaigns were quiet, even dead, and certainly 'low key'. 'In Skipton and Ripon constituency, the election was a very low key affair' (B1915). 'Door to door canvassing has been low key' (C1420). 'There's been very little door to door canvassing. A Tory woman did call the other night but that's all we've come across yet.
The other three parties seem to be very low key' (E1408).
Instead of encountering politicians via local campaigns, including political meetings, panellists now encountered politicians almost exclusively through national media: newspapers, radio, and especially television. Panellists wrote mostly of political interaction that was uni-directional e from politicians and journalists who spoke and wrote, to citizens who listened, watched, read, and had little opportunity to respond, to ask questions, to test politicians. Nevertheless, it does seem that party political broadcasts and the media campaigns in which they were embedded provided at least some opportunities for politicians to perform virtues or vices to citizens, and for citizens to calibrate judgements of politicians.
Let us take Labour's campaign as an example. As we have seen, it was found to be notable for its professionalism by commentators at the time. It was also the campaign panellists wrote most about in their diaries. For some, it was to be commended for its professionalism. It was modern: 'Labour has a modern, highly efficient, attacking campaign style' (H598). It was polished: 'It was undoubtedly a TV campaign, with Labour presenting a more polished, professional image than their opponents' (B1752). It was glamorous: 'the Labour Party have had a very glamorous campaign with very effective packaging of Mr Kinnock' (D1526). For one respondent, it made good use of advertising services: 'The last election, the Tories had an advertising agency giving them a public image. They kept the same agency anyway. The Labour Party also this election have an advertising agency and they have made a very good job of it. I feel they have had the best of the TV battle' (L1422). For another respondent, the campaign made good use of marketing techniques: 'Labour marketing certainly raised their profile and gave them a little chance' (W1675).
At the end of the 1980s, terms like 'professional', 'modern', 'polished', and 'glamorous' could still be used positively as complements to describe political campaigning. But the reception of professionalised campaigns was mixed among panellists. For one respondent: 'The videos were more likely of use to the English tourist board than promoting a political party' (C706). For another panellist: I felt that all three parties erred, nationally, on the side of what the pundits called a 'presidential' style with too much emphasis on the leader and less on policy. This was particularly so with the Labour Party, which was more concerned with showing us what a good and caring guy Neil Kinnock was than telling us what we really wanted to know. (W633) While some were impressed by Labour's professional campaign, others objected to its use of promotional video, its emphasis on Kinnock's image, and its use of marketing and advertising techniques more generally. Whatever the view taken of Labour's campaign, it does seem to have been possible in 1987 for citizens to calibrate judgements of politicians based on their party election broadcasts and associated media appearances. Again, if we take the example of Kinnock, judgements flowed from watching him on television. 'Labour's campaigning was some of the most culturally stimulating we've ever had in this country. I don't think anyone could doubt Neil Kinnock's sincerity or his belief that his policies were the best available' (S1964). 'Neil Kinnock started off as a 'light weight' politician, but during the campaign has grown in stature. His honesty, sincerity and compassion, his caring attitude for the underprivileged, compares favourably with Mrs Thatcher's authoritarian outlook' (C1420). Specifically on Labour's party election broadcasts, one respondent wrote: 'I must confess to shedding many tears at the Labour Party's brilliant election broadcast on the first Thursday of the campaign. I felt emotionally drained. The broadcast re-emphasised my opinions about him that he is a genuine person who backs fair play for all' (B1752). Another panellist wrote: The most impressive party election broadcast to date was the Labour one on Neil Kinnock. Beautifully put together: it can only have done him good. I've always felt Neil Kinnock had a tendency to try too hard […] There was a touch of insincerity in his manner. The election broadcast bio-pic managed to avoid that and he came over very well. (J291) Of course, views on the Labour leader were also mixed. While some judged him on the basis of his media performances to be sincere, honest, compassionate, caring, and genuine, others found him to be blustering, aggressive, too emphatic, and/or weak. The point here is that, in the late 1980s, citizens and politicians encountered each other most prominently through national media, political interaction was largely uni-directional, but politicians could still perform virtues and vices through professionalised media campaigns, and citizens could still calibrate judgements of these politicians.
However, this point should not be pushed too far and the final claim to make about the 1987 diaries is that, for all the judgements made about Kinnock and other politicians, the vast majority of judgements found in the diaries are about media campaigning and media coverage of politics. There was too much of it and it went on for too long. It was repetitive, tedious, and boring. A storyline circulated of media overkill: 'The general opinion seems to be that people are sick of the election from the word go. Many of the people I have spoken to have no interest in it and are thoroughly bored by the media overkill' (J1949). 'I think that everyone's opinion, including my own, was e total overkill. On TV there seemed to be no actual news for 2e3 weeks, merely extended election bulletins' (T1927). 'I do feel increasingly irritated with overkill in the media, especially monopolising TV. Shall be glad when it's all over' (M355). Finally, another storyline circulated about citizens who were fed up. They were 'fed up with the election' (B36), 'fed up with these politicians waffling on all the time' (S1857), 'fed up of all the election coverage on TV' (D1833), and 'fed up with it all' (G1909). In 1987, people were fed up not least because while they could switch off the television e and many did just that e there were few channels to choose from and people could not choose 'normal' programming instead (M1519).
The General-Election diaries from 1987 provide an account of low-key local campaigns, national media campaigns, and citizens fed-up with media overkill. Politicians and citizens encountered one another most prominently through televised party election broadcasts and news coverage, as opposed to the long radio speeches and rowdy political meetings of 1945. Political interaction was more uni-directional and citizens had less opportunity to test politicians themselves. Through televised party election broadcasts, party leaders could still just about perform virtues and vices to citizens, and citizens could still just about calibrate judgements of this limited group of politicians. But citizens' judgements mostly focused on the media campaign and media coverage. No prototypical categories like 'the good speaker' or 'the best candidate' can be found in the 1987 diaries. Compared to 1945, panellists used fewer comparative and superlative terms to describe politicians.

2001: televised debates, opinion polls and expert analysis, frauds and buffoons
If the packaging of politics increased during the 1980s, it became obsessive during the 1990s (Franklin, 2004). The launch of Sky News in 1989 heralded a multiplication of media outlets and an expansion of news and current affairs programming. The parties responded. For example, in 1989, the Conservatives began using people-metering to identify 'power-phrases' for repetition across their communications. Labour followed soon afterwards (Rosenbaum, 1997).
By the 2001 General Election, the main parties had grids setting out their agendas for each day's press conferences, leaders' tours, and broadcasts (Butler & Kavanagh, 2002). They followed these plans closely, tweaking them in response to nightly focus-group results as necessary. Spokespeople and candidates were primed with messages of the day and agreed lines by fax and e-mail. Journalists responded by portraying the election as boring (ibid). They focused on opinion polls that predicted a landslide for Labour on a low turnout. Or they focused on events that unmasked and embarrassed the parties, as when Sharon Storer berated Tony Blair outside a Birmingham hospital for the cancer care received by her partner, or when John Prescott punched a man for throwing an egg at him.
When polling day arrived, Labour won a second majority and lost only six seats. The incumbent party had a good story to tell. It had spent the last Parliament securing economic stability. It now had plans to invest in schools and hospitals. Meanwhile the Conservatives were stuck with few options. They retreated to a core vote strategy focused mostly on Europe e an issue of low salience to the majority of voters at the time (Butler & Kavanagh, 2002). But the big story of 2001 was the lowest turnout since expansion of the franchise at 61.8% (adjusted). Numerous explanations were suggested for this low turnout, including that Labour were expected to win comfortably; the campaign was not active by historical standards (especially in the non-targeted seats); and there was a perception among citizens that both main parties had become similar, making the choice between them less important (ibid).
In this context, MO asked its panel once again to 'record [their] reactions' and 'keep a diary' during the campaign period (SxMOA2/ 1/62/2 and SxMOA2/1/63/3). As in 1987, panellists wrote of receiving literature through the door but seeing few posters or canvassers. They wrote of media coverage, which they found to be too much; too long; too tedious, puerile, and boring. But two new components of national media campaigns and their coverage are particularly prominent in the 2001 dairies. The first is televised debates between politicians, journalists, and other politicians. Panellists watched these debates: 'On the Record with John Humphries, 2.00pm. Today's debate on the economy with Andrew Smith, Labour, Michael Portillo, Conservative, and Matthew Taylor, Lib Dem' (C1939). Or they failed to watch them when usually they would have done: 'I have not followed the election as closely as I usually do. In fact, I have not watched any debates and have only watched the news' (F1634).
The main storyline shared by diarists about these debates was that they were stage-managed so that politicians could avoid answering questions or addressing topics of concern to voters. 'Today, we had Tony Blair with Jonathan Dimbleby […]. The debate went well, but felt it was a bit stage-managed e.g. a lady in the audience put the question "why was the French Health Service better than ours". The question was quickly pushed aside and not answered' (C1939). 'I turn off the radio if Ann Widdecombe comes on, or Hague, or any of them actually. They do not listen to the question asked and they do not answer the question either' (J2891). 'I avoided party political broadcasts but watched some Newsnight debates. Jeremy Paxman not at his usual standard. It seems that problems were swept under the carpet. Foot & mouth, for example […]. The fuel crisis also seems to have been forgotten' (R2862). It may be that Paxman did usually push candidates to address certain topics. But one respondent preferred to do this directly as part of a rowdy audience in post-war political meetings: Television, by greedily and self-importantly hogging all the political activity, ruined the opportunity for 'ordinary' people to become involved. The days when people could take part by attending local political rallies and meetings, making demands, asking questions, and heckling candidates are long gone. And political commentators seem mystified as to why there is such a level of disenfranchisement! Take the 'spinners' of national television and much more out of the arena during the elections and people will surely retake possession of events and issues. (H1541) What should be noted here is that televised debates played a significant role in structuring political interaction between politicians and citizens in 2001, and these debates were perceived to be stagemanaged, to be spun, to be insulated from the demands and questions of 'ordinary' people.
The second 'new' component of political campaigns and their coverage, particularly prominent in the 2001 diaries, was reporting of opinion polls and expert analysis. These were mentioned by panellists in 1987, but they were commented on much more frequently and at much greater length in 2001. 'The opinion polls have the Conservatives trailing badly with Labour ahead with some 55% of the poll' (G2089). 'I have to admit that in the last two months before it was called, I thought that maybe the Labour majority would be down […]. But if we are to believe all the polls and the experts, I am wrong because they are saying that not only will Tony Blair win but he will have an even bigger majority' (F1634). The storyline circulated widely at the time that, based on the polls, the General Election was a 'foregone conclusion'. For one respondent: 'The result was a foregone conclusion […]. This was the first time I never sat up to watch the results on TV' (H260). For another panellist: 'In 2001 it was just a foregone conclusion. Nevertheless, I did watch the results come in until about 4.00am. There are very few people I know who bothered' (M2933). For a third respondent: As for the pre-election lies and misrepresentations mouthed by the three main parties, to be daily thrust down our throats by the media, and if we are further to take into account the opinion polls, then we are heading for the biggest yawn-producing event of the new century on June 7th. That Labour are still too far ahead to be overtaken by the Conservatives, no matter the promises and bribes made to voters, makes all a foregone conclusion and hardly worth going to the polls. (R1418) A foregone conclusion makes for no exciting surprises. It leads citizens to feel powerless. Some panellists expressed concern about polls and expert opinion along these lines. 'Both sides are worried about voter apathy. Labour because the polls put them so far ahead and the Tories because they are being given no chance […]. I feel that obsession of the media with the thought of voter apathy only increased rather than decreased that apathy' (B1426). 'I would like to see opinion polls banned […] and also television pundits to refrain. They try to be impartial but are not. Politicians should have their time on tele and hold meetings so that people can judge for themselves' (J1481).
Let us turn to this question of citizen judgements. Given the most prominent contexts of political encounter in 2001 e the stage-managed televised debate plus media coverage of polling results and expert analysis e what performances were possible for politicians and what judgements were possible for citizens? Were politicians able to demonstrate virtues? Were citizens able to 'judge for themselves' and distinguish good politicians worthy of support? As in the 1987 diaries, we find few if any prototypical categories of politician (equivalent to 'the good speaker' or 'best candidate' of 1945). We find few characterisations and judgements of individual politicians at all. Instead, as we have seen, panellists wrote about poll results and expert opinion. They appear to have delegated their judgement to others.
Nevertheless, if shared categories of politician can be found in the 2001 diaries e more sporadically and less consistently than before e they are 'the fraud' and 'the buffoon'. Consider this from one respondent: I would have found it impossible to vote for William Hague as Prime Minister e he just comes across as a buffoon, not a real person. Tony Blair, on the other hand, is too smarmy for me e a bit too media savvy and image conscious. I believe that he is capable of something like what Bill Clinton did e staring straight into the camera and telling a complete lie. (G2776) And this from another panellist: 'I stayed up quite late to watch the elections. It was quite fun. Ann Widdecombe managed to prove her insanity and the other politicians slimed their way through the various questions' (J2891). The fraud was media-savvy, smarmy, slimy, image-conscious e represented best by Tony Blair: Tony Blair ridiculously launched the election at a girls school […]. When they started singing hymns, he looked into the middle distance, all trembling chin and watery-eyed. He's such a fraud! On Sunday, David Frost had a go at him and again, you look into his face and you can see he just looks as if he doesn't believe what he's saying. (P2819) The buffoon was mad, insane, pathetic, puerile e represented best by William Hague or John Prescott. 'Well, well, well e what about John Prescott!! He punches a man because he gets an egg thrown at him […]. The man's mad. Just as I said above, a buffoon' (P2819). 'Two things remain in my memory of the election campaign: 1) John Prescott's straight left to the jaw of his assailant; and 2) Wee Willy waving a pound coin in the air triumphantly, as if he's just won the lottery or scored the winning goal in a cup final. Puerile and pathetic' (R1389).
In 2001, we see the same logical connections we saw in 1945 and 1987. The most prominent spaces of political encounter were now the televised debate and associated media coverage e especially reporting of opinion polls and expert commentary. These contexts afforded particular modes of interaction. Citizens mostly watched politicians on television. A minority of citizens were given voice indirectly through polls or focus groups. An even smaller minority found opportunities to aim questions, opinions, or eggs at politicians on tour. But journalists were now expected to challenge and test politicians on behalf of citizens. These modes of interaction afforded particular performances by politicians: the avoiding of questions, the avoiding of issues, the sticking to the script (or the making of gaffes by going off script). In the diaries, all this leads to judgements of politicians as frauds or buffoons. Alternatively, it leads to no judgements at all e no comments on the qualities of particular politicians, their performances, the virtues or vices they displayed, their standing in relation to other candidates; but instead just the noting of judgements by pollsters and expert commenters regarding which politicians and parties were up and down, almost as if the General Election had little to do with the judgements of voters themselves.

Conclusion
Anti-politics is a current concern for scholars, politicians, and many others (Clarke, Jennings, Moss, & Stoker, 2016). Negativity towards the institutions of formal politics increased during the second half of the twentieth century in countries like Britain. This long-term and complex development is likely to be explained by multiple factors. In this paper, we asked what previously under-utilised evidence from MO's General Election diaries could tell us about two specific parts of this phenomenon: citizens' negative sentiment regarding politicians; and political communication explanations for the rise of such negativity.
When viewed in the framework of contextual theories of social action, the diaries allow us to see logical connections between spaces of political encounter, political interaction, performances by politicians, and judgements by citizens. They allow us to see these relationships for different moments of the period e 1945, 1987, and 2001 e corresponding to three different ages of political communication (Blumler & Kavanagh, 1999). They allow us to make the plausible argument that prominent spaces of political encounter changed during the period, and these changed contexts afforded changed modes of political interaction, performances by politicians, and judgements by citizens.
The focus of this paper was on General Election campaigns during the second half of the twentieth century. This focus leaves at least two questions. First, are certain prominent spaces of political encounter and associated modes of political interaction missing from the MO diaries that just focus on General Election campaigns? We discussed this question briefly earlier in the paper. It is possible that citizens encountered politicians outside of these campaigns by other means (e.g. constituency surgeries). Our initial impression from reading MO writing produced at other times is that little if anything of relevance and significance is missed by the diaries. But a more systematic comparison of political communication during election campaigns and at other periods e using similar materials to the MO diaries (with their advantages of capturing citizens' perspectives, written in their own terms, for different historical moments) e would help to confirm this.
Second, we focused in this paper on the period when commentators generally agree that anti-politics increased most in Britain. But what about developments since 2001? If the second half of the twentieth century was characterised by the nationalisation of political campaigning, then recent years have seen a revival of constituency campaigning. We know much about how this has affected votes for parties (e.g. Cutts et al., 2014;Johnston et al., 2016). We know much less about how it has affected political interaction, the performances of politicians, and the judgements of citizens (about politicians in general). Similarly, recent years have seen the rise of political communication by internet. Research on the relationship between internet use and anti-politics is ongoing and has produced mixed results to date. For example, some have found the anonymity and flexibility of the online world to encourage uncivil debate that decreases political trust and efficacy (Åstr€ om & Karlsson, 2013), or that internet use does not increase political interest, efficacy, and knowledge (Richey & Zhu, 2015). Others have found that internet use leads citizens to adopt stronger democratic preferences (Stoycheff & Nisbet, 2014), or that 'everyday celebrity politicians' can use social media to perform authenticity and reconnect with disaffected citizens (Wood, Corbett, & Flinders, 2016). Keane (2013) captures this mixed picture by his dual focus on the promise of media abundance (digital democracy, cybercitizens, e-government etc.) and the risk of media decadence (censorship, spin, echo chambers, rumour storms, cyber attacks, online gated communities etc.). More research is needed on developments in the current period, including on how encounters between politicians and citizens via the internet shape political interaction, performance, and judgement.