CAIRN-INT.INFO : International Edition

1For several decades social sciences have critically assessed the role that the press plays in local political life. Local and regional newspapers were seen to provide a depoliticised and biased coverage of the local political reality. In solely relaying the institutional message, [1] in minimising the diversity of opinion, [2] in focusing on producing consumer news and in manipulating regional cultural stereotypes, [3] journalism was not perceived as having contributed as much as it could have to the vigour of local democracy. Several factors contribute to this phenomenon: the more or less negotiated relationships of dependency that form between the local reporter and the local authorities; [4] the professionalisation of municipal team communication; [5] and the consolidation among local press organisations, which increase commercial constraints and dependency on elected officials. [6]

2Is the manner in which local and regional press organisations make use of the internet today likely to transform the role that they play in local political spaces? If one reads the studies that have focused on how the web has transformed journalism, [7] one would be tempted to answer in the negative. The increase in work flexibility [8] and fragmentation, [9] and in imitation between journalists, [10] all lead one to believe that a growing use of the web by the local press is unlikely to challenge the current dynamics. In fact, the web might even be expected to bolster commercial and institutional logics, as Nicolas Pélissier notes: “The bland and sanitised nature of the consumer-oriented news on local web sites gives more the impression of an “upstanding regional news magazine, respectful of the institutions”, having its own shop window online, rather than a genuine force for opposition acting to democratise regional public space.” [11]

3In this article, we will show that certain operations initiated on the web by local press organisations are at odds with these analyses. By making online forums available to individuals so that they can discuss and comment on local public affairs, the role that press organisations usually play in some local political spaces is changing substantially.

4Our analysis focuses on an initiative set up by the multimedia editorial team of the regional daily newspaper La Voix du Nord at the time of the local government elections in March 2008. [12] From the summer of 2007, the newspaper’s local reporters wrote articles appraising the performance of the outgoing mayors of each of the 1,547 communes [the smallest administrative unit, or municipality] of the Nord-Pas-de-Calais region. These articles were published on the newspaper’s website and internet users were given the opportunity to leave comments on the same page. [13] Discussion forums were thus created which individuals made extensive use of to discuss their own communes – 17,028 comments were published, three quarters of these in the six weeks preceding the first round of the elections. [14]

5However, in order to conduct this analysis rigorously we had to confront a theoretical issue that seldom arises in studies concerning the web. Every commune in the region in effect represents a specific socio-political space: they all have a certain number of inhabitants; they cover a certain geographical area; the relationships between the inhabitants are specific to that particular commune; the inhabitants more or less share certain standards; power is exercised by particular political institutions; electoral competition assumes a certain form in each commune, [15] etc. To understand the role that La Voix du Nord played in each commune through such an initiative, one must first grasp how individuals rooted in such different social spaces used these online forums. In order to do this, we adopted a morphological approach, meaning that we tried to establish a link between the number of inhabitants and their dispersion within each commune and the nature of their involvement in online discussions. As defined by Émile Durkheim and Maurice Halbwachs, social morphology is a sociological approach which claims to explain social phenomena through “the mass of individuals which constitute society, the manner of their geographical distribution and the nature and configuration of the whole range of phenomena which affect collective relations”. [16] Applied to the analysis of urban sociabilities and civic involvement, following Louis Wirth, the morphological approach showed that the relationship between the inhabitants of a commune and their involvement in local public life were closely linked to their number and their dispersion within the commune.

6Conducting a morphological analysis of web forums challenged both the literature concerning online political discussion – which struggles to link the analysis of online discussions with that of individuals’ offline socio-political contexts – and more generally the analyses of web use – which are focused on the individual and their social characteristics. This also entailed methodological difficulties, since it was a question of processing a voluminous corpus of texts and grasping the links with more than 1,500 different communes. However, we will see that such an approach has revealed that these online forums, offered by the regional newspaper, are used in very different ways by the communes depending on their morphology.

7In the first part of the article, we will show the extent to which the morphological dimension is absent from current studies of internet use even though it has, for a long time now, been integrated into the analysis of the political involvement of individuals. In the second part, we will describe the methodology that we employed, which is based on quantitative processing of the corpus of comments and pseudonyms. In the third part, we will show that variations in the participation in discussions concerning the local government elections on the La Voix du Nord website can, in large part, be explained by the specific morphological characteristics of the urban areas concerned. Then, in a fourth part, we will single out three distinct forms of online discussion in relation to the morphological characteristics of the region’s communes. Lastly, we will show that such use of the web changes the relationship that a press organisation has with its regional territory.

Political discussion and social morphology

8Reading the sociological and political science literature that focuses on the civic involvement of citizens, there is consensus that morphological variables play an important role. The more a town’s population is numerous, dense and socially heterogeneous, the less its inhabitants get involved in the commune’s public affairs: they participate less in local elections, look less to get in contact with their locally elected representative, attend the local town hall meetings less frequently. [17] Several French works on local elections have for example shown that participation in local government elections is systematically greater in the smaller communes[18] and that the expectations of voters vary notably according to the size of the town where they live. [19] Inspired by Durkheim, Wirth’s theory explains that with urbanisation comes a weakening of interpersonal bonds, of primary social structures and of consensus on standards, which in turns leads to a withdrawal from social and political activities. [20] This theory has since been qualified and even contested; [21] however a consensus still exists that the number of inhabitants is a significant determinant of local civic involvement. [22]

9With the discussion of local political affairs – even if it is online – being an important part of the civic involvement of citizens, one might expect it equally to be linked with the morphological characteristics of the communes. And yet, the majority of studies on web use and online political discussion touch only marginally upon the morphology of the offline spaces in which the internet users are found.

10Within research on the digital divide, several studies have attempted to determine the extent to which certain uses of the internet are associated with greater or lesser civic and political involvement. [23] These works show that the internet does not fuel the social and political withdrawal of individuals, but rather that overall it promotes political involvement. [24] Although they may not all share the same theoretical perspective, one characteristic these works do have in common is to focus the analysis almost exclusively on individual variables: socioeconomic status, level of qualifications, age, gender, race, access to support in their environment, length of time that the individual has been using the internet, etc. Some studies include the regional dimension in addition to the nature of the environment in which the individual is located (urban/rural), [25] without however extending the analysis to include the morphology of the spaces to which the individuals are linked.

11From a different perspective, the numerous studies which focus on online political discussion are also not interested in the morphological aspects of these political conversations. A majority of works have focused on the dynamic of political discussions on the internet, comparing them to normative models of public space. As Fabienne Greffet and Stéphanie Wojcik highlight, the ways in which online political discussion and offline political spaces are linked have largely been excluded from the researchers’ questioning. [26] The reason for this is firstly because the participants often express themselves anonymously and it is very difficult for the researcher to know in which political spaces they move offline. But at a deeper level, part of this literature is based on the idea that the very existence of these debating forums is due to the fact that participants distance themselves from the socio-political space in which they move. In other words, individuals participate online – and in fact even more so – because the discussion format ensures that it is impossible to determine social, professional, cultural and geographical identities. By joining in political discussion online, participants free themselves both from the constraints of ordinary social relationships [27] and the numerous restrictions to speaking publicly which result from the way local political spaces operate. [28] Individuals are perceived to express themselves more freely when free of the constraints associated with ordinary social relationships and courtesy requirements specific to face to face relationships and, as a result, democracy improves through less constricted discussions. [29]

12Several researchers have, however, suggested that there could be a link between speaking in certain forums and belonging to a specific socio-political space. This is the case in studies that have focused on municipal forums. [30] During the debates, participants mention, for example, their local connection and how long they have lived in the commune. [31] When assessing the use of municipal forums in several communes, Vedel hypothesises that the level of participation in these forums is probably “very dependent on the size of the town and the quality of the existing social fabric”. [32] Analysing the use of municipal websites and forums by Norwegian municipalities, Are Vegard Haug hypothesises that the need is greater in the big communes where communication between citizens and their locally elected representatives is more difficult. [33] However, beyond these hypotheses, no study has systematically analysed the link between speaking publicly online and the morphological characteristics of the local space.

13A certain paradox therefore arises. While traditional studies emphasise the importance of the morphological dimension in analysing the participation of citizens in local political life, online political discussion is seen as a largely individual phenomenon. The analysis that we offer in this article, on the contrary, shows that online political discussion, whenever it concerns local affairs, is largely influenced by the morphological characteristics of the local area.


14It was during an interview-based survey of La Voix du Nord concerning innovation in regional and local online press that we learnt of the forums set up during the local government elections of March 2008. With the help of the CorText platform, [34] we conducted a principally quantitative analysis of them. This was undertaken in three distinct stages, namely the statistical processing of participation; the lexical processing of comments; and the analysis of heterogeneous data. Beforehand, using a computer programme, all the comments were automatically extracted from the La Voix du Nord website into a database. For each commune we were thus able to collect all the associated comments, as well as their date and the authors’ pseudonyms.

15A first series of processes consisted of analysing the distribution of participation in the forums – measured by the number of comments published and the number of commentators. Systematically, we linked the participation in each forum with a series of variables relating to the communal morphology (number of inhabitants, town density), to the intensity of municipal political life (changing of mayor, rate of turnout in the first round of the local government elections) and to the commune’s media environment (coverage of the commune by the local editions of La Voix du Nord). [35]

16Secondly, we undertook partially automated lexical processing. First of all, we extracted terminology from the entire database of published comments. From this automatic extraction, based on a certain number of grammatical and statistical criteria, [36] we isolated the 3,385 most frequently used expressions. In an analytic approach borrowing from the sociology of critique, we then manually encoded some of the expressions on this list (1,829 expressions, or 54% of the total) linking them to specific ways in which the discussions unfolded. [37] Our coding included seven large separate categories which enabled us to identify:

  • the way in which contributors address the other participants, the outgoing mayor and the other candidates, and the manner in which they refer to the inhabitants of the commune;
  • the partisan commitment of the participants, according to whether they show their support for the candidates (“Well done mayor”), reject them (“make way for the youngsters”), or use partisan vocabulary (“clans”, “leader”);
  • the scale of the argument, depending on whether it is based on national categories (“Nicolas Sarkozy”, “Socialist party”), local categories (“riverside residents”) or family-related categories (“small children”, “my spouse”);
  • the way in which contributors speak of local government, through their judgements on the actions of the mayor and the outgoing municipal team, the competency of the candidates, the make-up of the lists and the institutions outside the commune;
  • the forms their arguments take: are participants denouncing or exposing; are they mobilising democratic principles; are they appealing to categories of rational discussion; are they using marks of courtesy or, on the contrary, are they hurling insults;
  • the judgement that the contributors make of the online forum: for example, do they criticise the anonymity or the violence that exists in it, or complain about participants’ spelling?;
  • the different themes covered in the discussion: economy, transport, associations, finance, the commune’s reputation, crime, education, infrastructure, housing, environment, sport, culture and heritage, etc.
We then manually encoded the 6,869 pseudonyms associated with the comments so as to enable us to determine the ways the participants engage in the forum. We managed to identify several categories of pseudonyms:
  • anonymous pseudonyms (a friend);
  • first names, diminutives or initials (Franck);
  • proper names (Michel Legrand);
  • pseudonyms making reference to the locality, through mentioning the place of residence (a real Gonnehem resident) and/or a role, or an involvement in a local association (a Saint Souplet footballer), or a local responsibility (the mayor);
  • critical pseudonyms (we must change), possibly combined with a local reference (Bousies in danger) or a national reference (an angry socialist).
The next stage consisted of measuring and representing the close relationships existing between, on one hand, the number of inhabitants in the communes and, on the other hand, the semantic categories identified in the corpus of comments or the categories of pseudonyms, in order to produce heterogeneous maps. [38] Each comment can thus be interpreted as a co-occurrence event connecting the size of a commune – in other words the number of inhabitants in the commune where the comment was published – and a category of pseudonyms or the semantic categories from which the comment comes. The comparison – made with the help of the ?2 (chi-squared) test – of the actual number of these events observed with the theoretical number of co-occurrences that could be expected from a null hypothesis of independence of the categories, enables the specificity of one category to be calculated in relation to the other. [39] Thus, two entities (a category of pseudonyms and a given class of commune size for example) will be independent if the specificity is zero. A high positive specificity indicates a very close relationship whereas a negative value means a type of anticorrelation between the two entities.

17Once all these relationships had been calculated, they could be represented in the form of histograms (cf. Figures 4, 5 and 6) or aggregated in a network, the weighted links of which reflect the strength of the close associations between entities (cf. Figure 8).

The morphological dimension of participation

18As is often observed in online activity, participation in the online forums set up by La Voix du Nord is very unequally distributed. On average, 11 comments have been published in each forum, however for a little over half of them no comments have been posted whatsoever (Table 1). Just under 6% of the region’s communes account for 80% of the comments. Some communes experience very high levels of participation – as in the case of Coudekerque-Branche, a commune near to Dunkirk which has attracted 1,268 comments alone.

Table 1

Very unequal participation

Table 1
Number of comments Forums Percentage None 846 54.7 Between 1 and 5 427 27.6 Between 6 and 30 174 11.2 Between 31 and 100 60 3.9 More than 100 40 2.6 Total 1547 100

Very unequal participation

19On average, contributors make fewer than two comments under the same pseudonym (1.87 comments), the vast majority expressing themselves only once. Contrary to what can be observed on other local forums, [40] the opportunities to speak publicly do not appear on the whole to be monopolised by an ultra-active minority, even if a minority of inhabitants clearly participate in the discussion. [41]

20Can this unequal participation be explained by differences in access to the internet? In the first half of 2008, the inequalities in terms of internet access were considerably reduced and the high speed network covered a large part of the region, even though some disparities remained. [42] Internet use also varied by commune, social group and geographical area. [43] However, during the campaign, only a small proportion of the distribution of participation can be explained by inequalities in access to the internet and differences in online practices. Thus, the communes in Lille and its suburbs, which have the highest rate of internet access at home at this time (73.8% of households) experience one of the lowest levels of participation in the region (on average 2.33 contributors per thousand inhabitants) while the communes in the Cambrésis-Valenciennois area, where internet access is markedly lower (64.3% of households), experience a far higher level of participation (on average 6.24 contributors per thousand inhabitants).

High level of participation in medium-size towns

21Generally, the proportion of active forums steadily increases with the size of the towns associated. The more populous a commune is, the more the associated forum is likely to be active – defined here as having at least five comments (Figure 1). Similarly, the more densely populated a commune is, the more the associated forum is likely to be active.

Figure 1

Share of active forums relative to commune size

Figure 1

Share of active forums relative to commune size

Note: the curve indicates the proportion of communes having received at least five comments on their forum, relative to the number of their inhabitants.

22However, the majority of this participation is down to the inhabitants of small and medium-size towns. In comparing the distribution of contributors to the forums with the distribution of the regional population by commune, it is very clear that those who contribute the most are the inhabitants of the small, and particularly the medium-size communes (Figure 2). Conversely, the inhabitants of the large towns and cities far more rarely use the online forums provided by the regional daily newspaper. 60.9% of those who contribute do so on the forums of communes with between 1,000 and 10,000 inhabitants; even though these communes represent only 40.4% of the total regional population. Small communes (fewer than 1,000 residents) and large communes (more than 10,000 residents) respectively account for 17.1% and 22% of the total number of contributors – even though the large communes have five times more inhabitants than the small communes.

Figure 2

Distribution of participation relative to commune size

Figure 2

Distribution of participation relative to commune size

Note: the black curve indicates the distribution of the number of contributors to the forums (measured by the number of different pseudonyms identified) relative to commune size. The grey curve indicates the distribution of the regional population relative to town size.

Participation that decreases with the number of inhabitants

23Just as commonly occurs for elections, it is possible to measure an “online turnout rate” for each commune. We obtain this by dividing the number of different pseudonyms that are identifiable on the forum by the number of inhabitants in the town. In so doing, we find that average online participation initially grows slightly for the less populated towns, before steadily decreasing as the number of inhabitants increases. Generally, the more populous a town is, the smaller the proportion of inhabitants that comment online (Figure 3). Moreover, we notice that the highest participation rates are found in the very small communes. The commune of Nort Leulinghem, near Saint-Omer, for example, has the highest participation rate with 58 contributors per 180 inhabitants – in other words 322 contributors per 1,000 inhabitants. From the moment small town inhabitants decide to use the forums, they do so with an intensity that is markedly higher than that found in the more populous towns. Among those communes that have more than 20,000 inhabitants, the town of Coudekerque-Branche, near Dunkirk, is an exception with a high rate of 17 contributors per 1,000 inhabitants (417 contributors per 24,400 inhabitants), a rate that is still, however, significantly lower than those found in the small communes.

Figure 3

Average online participation relative to commune size

Figure 3

Average online participation relative to commune size

Note: the curve indicates the average rate of online participation for each commune size (obtained by dividing the number of different pseudonyms identified by the number of inhabitants in the commune).

24This leads one to think that offline and online participation in politics (the vote and the contribution to a political forum respectively) are not that different from each other, and, on the contrary, could in fact share a common logic. We know for a fact that in Western countries, participation in local elections steadily decreases as the number of inhabitants in the commune increases. [44] We would therefore be tempted to explain it in the same terms: that as a town grows, interpersonal bonds, primary social structures and consensus on standards weaken, so much so that individuals tend to withdraw from political activities. [45]

Intensity of local political life

25As soon as communes reach a certain size, their inhabitants use the forums all the more given that the intensity of local political life is greater. We considered the change of mayor and the turnout rate in the first round of local government elections as two indicators of the intensity of the local political situation. [46] It appears that when a town has more than 2,000 inhabitants, a change of mayor is accompanied by more active use of the forums (Table 2). This is particularly true for communes with between 2,000 and 5,000 inhabitants as well as communes that have more than 10,000 inhabitants. In these conurbations, the inhabitants have clearly found these online forums to be interesting arenas in which to debate the municipal political situation.

Table 2

Correlation between forum activity and the intensity of local political life[47] relative to commune size[48]

Table 2
Commune size Forums Change of local government Voter turnout (1st round of local government elections) Fewer than 500 inhabitants 639 - 0.032 + 0.048 Between 500 and 1,000 inhabitants 307 + 0.062 + 0.067 Between 1,000 and 2,000 inhabitants 217 - 0.062 + 0.14 Between 2,000 and 5,000 inhabitants 200 + 0.18 + 0.166 Between 5,000 and 10,000 inhabitants 102 - 0.038 + 0.123 Between 10,000 and 20,000 inhabitants 51 + 0.242 + 0.089 More than 20,000 inhabitants 31 + 0.254 + 0.256 All together 1547 + 0.0003 - 0.253

Correlation between forum activity and the intensity of local political life[47] relative to commune size[48]

Note: When the commune has fewer than 500 inhabitants, the Pearson correlation coefficient between forum activity and change of local government is - 0.032 (very weak negative correlation). When the commune has more than 20,000 inhabitants, the Pearson correlation coefficient is + 0.254 (relatively strong positive correlation).

26In towns with more than 1,000 inhabitants, the greater the turnout in the first round of local government elections, the more the residents contribute online. In these communes, when the rate of voter turnout is high, the inhabitants have felt a greater need to use the forums provided to them by La Voix du Nord.

27These different results point to participation that is strongly affected by the population and density of each commune. In the small communes, forums are used less, but when they are, a major proportion of the population participates. In the medium-size communes, participation is the greatest and varies depending on the intensity of the local political context. In the largest communes, forums are much more frequently active, even if they involve a smaller proportion of the population. Use of the forums varies depending on the intensity of the local context. [49] In order to understand the motivations behind this differentiated participation, we will now go on to analyse how the discussion unfolds in these different spaces.

28With commune size proving to be the key parameter in determining participation in these forums, [50] we will subsequently ignore the other variables (change of local government, voter turnout, density, etc.) in order to focus solely on the characterisation of the online discussions in relation to three large categories of commune size: small (fewer than 2,000 inhabitants), medium (between 2,000 and 10,000 inhabitants) and large (more than 10,000 inhabitants). [51]

Urban space and the unfolding of online discussion

29The lexicometric analysis of the comments and pseudonyms reveals specific ways in which online discussion unfolds. The method that we have employed enables us to establish a local network between the semantic categories and the size of the communes. By highlighting those categories which, all things being equal, are used more in the forums corresponding to certain sizes of town, this method reveals the major differences between the forums. As we will see, this differentiated unfolding of the conversation stems primarily from the forms of relationships between the local inhabitants, the nature of the local political institutions and the shape of the offline public municipal debate, which differ depending on the size and density of the communes.

The constraints of acquaintanceship

30The forums associated with towns of fewer than 2,000 inhabitants are to a large extent structured by the constraints linked with acquaintanceship, which dominates the relationships between the local inhabitants offline. [52] This is evidenced by the choice of pseudonyms (Figure 4). Far more than in the more populous communes, participants assume anonymous pseudonyms here (“a small insignificant guy”, “Anne Honymous”, etc.) or pseudonyms that express a connection with the village (“a villager”, “a native of Oxelaere”, “Brillon born and bred”). They more rarely identify themselves by their surname or first name – which is likely to be known by other participants – except when they are candidates or former municipal counsellors. When they join the discussion, the majority of participants therefore introduce themselves anonymously to others, often by mentioning their connection to the village. More clearly than in the other towns, anonymity here makes speaking publicly easier.

Figure 4

Categories of pseudonyms used in the small communes

Figure 4

Categories of pseudonyms used in the small communes

Note: the x-axis indicates the ?2 values corresponding to a test of independence between a given category of pseudonyms and the type of communes concerned. A value of 0 signifies that there is independence: the association between the category of pseudonyms and the small communes is identical to that which a random distribution model would give. Above 0, the association between the pseudonyms and the small communes is more frequent; below 0, it is less frequent.

31Although they may display their offline identity less often, participants in the discussion however address each other more directly. More clearly than in the more populous communes, they reply to one another – mentioning, at the beginning of their message, that it is a “response to comment no.” – and they address other contributors (“you’re right”, “that’s your right”, “you refer to”, etc.). Here, the forum is used by participants as a place for interpersonal discussion, where each contributor specifically addresses their comment – even though, more often than not, they do not know who is behind the pseudonym. The use of marks of courtesy and civility (“kind regards”, “good evening”, etc.) in the discussion is far more prevalent than in the forums of the more populous communes.

32In these forums, the only people identifiable by their personal details are often the mayor and the candidates. Here, one directly addresses “Mr Mayor” and the candidates. In these small communes, municipal power is embodied by the mayor, to whom the inhabitants are often close. It is therefore not surprising to find the mayor at the heart of the discussions. More so than in the larger towns, this results in the expression of support for the outgoing mayor in the form of thanks and encouragement: they “pay tribute” to the mayor, they say “many thanks” to him or her, wish him or her “good luck” or “happy retirement”, thank him or her for their “excellent work”. However, more than elsewhere also, this results in rejection directly addressed to the mayor: the commune needs “new faces”, the mayor must “make way”, he or she has “run out of steam”, etc. Whether it is a question of supporting or rejecting, the expression of opinions is therefore mainly centred on the mayor’s personality. The mayor must be “honest”, “competent”, “sincere”, “open minded”, “hands-on”, “courageous”, “dedicated” and “available”. It is therefore above all about the moral qualities of an individual and much less about this person’s political convictions, or their association with a political party or any local groups. [53]

33More than in the other communes, the discussion here is centred on the communal space. As the choice of pseudonyms mentioning the commune implicitly reveals, opinions are all the more legitimate when the individual expressing them has been connected to the village since time immemorial. The discussion itself demonstrates this singular concentration on the area of the commune. Firstly, the arguments refer more to the family and domestic world, while national categories such as political parties or national personalities are far more rarely mobilised than in the other communes. Next, the discussion focuses far more rarely on places of power outside of the commune – departmental or regional councils, the French National Assembly, the courts, etc. The forums of these small communes can, thus, be seen to provide a relatively inward-looking discussion space. When the outside is mentioned, it is solely in order to express the fear that the village has a bad reputation among the neighbouring villages.

34Anonymity is the major issue for the forums associated with the small communes. Although participants far more often take part anonymously, they also considerably more frequently criticise the anonymity which reigns in these forums. They complain about the anonymous comments and criticise the “writers of poison-pen letters”, those who “hide behind a pseudonym”, etc. In these forums, anonymity enables the discussion to unfold without the constraints associated with acquaintanceship offline, but at the same time it is singled out as giving rise to comments adjudged to be excessive, at odds with the norms of civility. Far more than for the larger communes, participants perceive the discussion to be too violent and too aggressive. Many of them criticise the “controversy”, the “slander”, the “personal attacks” and the “defamation”. “It is easy to criticise” they say most often to the “preachers” and “scandalmongers”.

35Thus a first type of arena appears which is closely associated with the forms of relationships between local inhabitants that dominate in the small communes. Online discussion unfolds here through allowing individuals to escape from the constraints of acquaintanceship which they otherwise live with offline, and which influence the ordinary expression of opinions. In the small communes, collective opinion exerts considerable pressure on the inhabitants, for whom online discussion offers a space in which this pressure is exerted less directly. However, once these constraints are loosened, discussion unfolds but remains very close to the structure of the communal political space: participants address one another directly, base their contribution on the familiar environment, keep away from national categories, and centre their exchanges on the personal qualities of the mayor.

Local opinion that struggles to materialise

36When communes have between 2,000 and 10,000 inhabitants, the constraints associated with acquaintanceship and with reaching consensus on norms are not as great. Interpersonal contact is more limited and acquaintanceship inevitably reaches its limits as the size of the commune grows. Moreover, it is easier for the inhabitants to express alternative political opinions. The reasons for this are notably institutional, since for communes with more than 3,500 inhabitants the law requires a complete list of candidates to be drawn up and prohibits voters from adding, deleting or changing the order of the names. This provides support to alternative opinions as it helps stabilise a municipal opposition. However, for individuals living in these medium-size communes, expressing an opinion publicly comes up against a major constraint: while it is more difficult for the inhabitants to interact with all of the commune’s other inhabitants, and dissonant political opinions are more accepted, individuals do not however have the necessary spaces in which to express their views and acquaint themselves with the opinion of the commune – local newspapers are few and far between and the regional newspapers give priority to the larger communes. [54] This explains why the online forum is used more here to take sides by addressing the town’s views.

37This is demonstrated by the pseudonyms principally used in these forums (Figure 5). Here too, contributors join in the discussion by making reference to the commune. However, far more than in the small communes, they take a partisan position. Through their pseudonyms they express an opposition (“the mayor out”, “yes we must change”), show anger (“to arms citizens”, “an angry mother”) or a partisan commitment (“opposition supporter”, “leftwinger from Auby”), make assertions (“we the disabled exist”), make reference to democratic principles (“justice”, “freedom”), display their sarcasm (“laughing out loud”, “bravo”), or present themselves as impartial observers (“an observer”, “a witness”). Far more frequently than in the other towns commenting online here is associated with taking sides.

Figure 5

Categories of pseudonyms used in medium-size communes

Figure 5

Categories of pseudonyms used in medium-size communes

38The lexicometric analysis of the discussion confirms the strong partisan dimension of the exchanges. Indeed, it is in these forums that the contributors most employ partisan vocabulary. It is a question of “clans”, of “support committees”, of “supporters” and of “partisans” committed to supporting their “champion” or their “leader” against “opponents” and even “enemies”. It is a question of “waging war”, engaging in “battle” “opposing”, with his or her “arms”, targeting “victory” or taking one’s “revenge”. It is also a question of guarding against “betrayals” and those who would like to “jump ship”, at the risk of leading to “defeat”. Contrary to the small communes, the discussion here is not centred on the personality of the mayor and the candidates, but rather on the opposition between the clans involved in a fight for local power.

39In these forums, the discussion is also organised very differently. Contributors more rarely address one another directly, rather aiming their comments at the forum’s public – in other words, the voters in the commune. The use of the imperative is particularly illustrative of this position, through which contributors enjoin the forum’s public to take notice of certain realities, to be moved by them and to act accordingly: “vote”, “dare to change”, “let’s get together”, “spot the mistake”, “be aware that the mayor”, “don’t forget”, etc. For the most part, it is a question here of appealing to public opinion, which ordinarily constitutes a reality that is difficult to access for the inhabitants of these medium-size communes. Compared to the small communes, the exchange here is less like a direct discussion between participants and more like a fight between individuals looking to convince the forum’s public of the merits of their views. In these forums, political information constitutes a considerable resource. This is why there is a lot more talk in them of the “public meetings”, the “leaflets”, the “neighbourhood meetings”, the “posters”, the “municipal newsletter” and the “professions of faith”. The participants depend on these sources of information to justify their taking a certain side and to appeal to the opinion of the town.

40Arguments here mostly take the form of denunciation and exposure: an individual contributes to the forum in order to shed light on the grim reality hiding behind the rhetoric of the candidates. They “raise doubts”, they urge the voters to “open their eyes”, they claim to “ask the right questions”, or show the candidate’s “true face”. They intend to demonstrate that the candidate’s “fine words” and “great promises” are just “smoke and mirrors”, “disinformation”, “propaganda” and “manipulation”. They want to expose the inaction or the incompetence of the outgoing mayor, which has led to “disaster” while “emptying the coffers”. They denounce the electioneering of the candidates, their “man œuvres” and more generally the hijacking of the general interest for the benefit of a small minority: they point to “clientelism”, “cronyism”, “privileges” and “skulduggery”. They appeal therefore to the voters, who they hope “will not be duped”.

41The forum that emerges here shares several features with the “affair form” studied by Élisabeth Claverie and Luc Boltanski. [55] The discussion which takes shape online can in effect be identified with a space for exposing scandals and affairs, in which one seeks to provoke outrage in public opinion in order to convince voters to adopt one’s own cause.

Local opinion, national issues

42In the large towns – those which have at least 10,000 to 20,000 inhabitants – expressing an opinion publicly does not face the same constraints. Acquaintanceship is even weaker, [56] as is the pressure for consensus. Furthermore, local political life is markedly more organised, around political parties, and opinion can be expressed through a greater number of channels, notably in the media. Aside from the regional daily newspaper, inhabitants of the largest towns also have local magazines, local radio stations (sometimes local television stations) and alternative online media. In this more varied media landscape, the forums provided by La Voix du Nord face even greater competition given that the inhabitants tend to involve themselves less in the public affairs of the commune. Yet, although it may be less often than in medium-size towns, the inhabitants of these communes nonetheless use these forums to appeal to local opinion, here more affected by national issues. [57]

43In these forums, it is also a matter of mobilising the opinion of the commune. This is demonstrated firstly by the relatively strong correlation that we have identified between forum activity in the large towns and the intensity of local political life (Table 2). Given the relatively extensive media environment, the inhabitants of these communes use these forums all the more when the political situation is more tense. The situation of Coudekerque-Branche, near Dunkirk, is particularly illustrative of this phenomenon. Its forum alone has attracted a third of all comments made on forums associated with communes which have more than 10,000 inhabitants. The political situation there was particularly tense: the once-socialist mayor, incumbent since the 1970s and implicated in several criminal justice matters, was up against a young candidate representing the socialist party – who went on to win the election on the evening of the second round.

44Furthermore, these participants more frequently use pseudonyms signalling a political affiliation which make sense in a national context (Figure 6). They display their national political preferences (“left-winger”, “socialist sympathiser”, “a former commie”, “long live the left”), make reference to contemporary figures (“Nicolas and Carla”, “Ségolène and Nicolas”) or historical figures (“Blum”, “Jaurès”, “Saint-Just”), or refer to national political categories (“lower class France”), or more general political categories (“a democrat”, “a citizen of the world”). As in the medium-size towns, expressing your views publicly online here is still therefore based on taking a partisan stance, even if this refers far more to the national rather than local context.

Figure 6

Categories of pseudonyms used in the large communes

Figure 6

Categories of pseudonyms used in the large communes

45Lastly, the discussion is based principally on a partisan reading of the local political world – participants also talk of the “battle” between “opponents” or “enemies” – and on the revelatory-denunciatory form of argument. These are two common features shared by the medium-size and large towns – even if these categories are found a little more often in the medium-size communes. Here too, it is a question of appealing to local opinion, by exposing a reality hidden behind the candidates’ rhetoric. These different elements allow us to state that, in these forums, it is also a question of the participants asserting their own, often partisan, opinion and of convincing the voters of the commune.

46However, these exchanges appear here to be considerably more swayed by the national context. In effect, it is in these large communes that the most references are made to political parties, national political figures (“Nicolas Sarkozy”, “bling bling”) and national institutions (“republic”, “state”). It is also in these forums that we find the most mentions of authorities outside the commune (the conurbation, federation of communes, regional council, National Assembly, government, courts). Participants also more frequently discuss issues of employment, transport and security – all of which are themes that fall within the remit of the large towns and cities and which are largely echoed in the national debate.

47The forum that emerges here therefore differs from that of the medium-size communes on two key points. Firstly, these forums operate within a more varied media environment and, as a result, their use depends more heavily on the intensity of the local political context. Secondly, these are arenas which are markedly less centred on the town and more open to national issues. This is due to the fact that the participants tend to be less involved in local affairs, so much so that national issues constitute for them an easier entry into the local debate. However, this is also explained by the fact that the municipalities’ responsibilites more often feature in national debates.

48A very close link appears therefore between the manner in which the discussion unfolds on the web and the morphology of the corresponding urban space. However, as has been often analysed, the morphological variable – primarily the number of inhabitants – encompasses a number of different elements of commune life which principally concern the nature of the relationships between the local inhabitants, the specificity of the local institutions, and the types of local political competition, in addition to the state of the local media landscape. Closely associated with commune size, these four elements account for how participants use these online forums.

49Now that we have identified what inhabitants typically use the forums for in relation to the characteristics of the urban space, we are in a position to understand how the uses of such a mechanism change the role that La Voix du Nord usually plays in regional democratic life.

A different relationship between press organisation and local area

50Like the majority of French regional newspapers, La Voix du Nord relies on a relationship with the local area that has its origins in the creation of local editions at the beginning of the twentieth century. [58] This model, which became established in the 1920s and further developed following consolidation of regional press organisations, consists of dividing the regional territory into zones that share both a cultural and socioeconomic identity. At the time of the local government election campaign, the daily newspaper comprised 24 local editions corresponding to as many zones. In the local agencies, the local reporters cover one or several communes and rely on a network of correspondents. The introduction of online forums in the run up to the March 2008 local government elections resulted in a change in the press organisation’s relationship with its local area. This is explained firstly by the fact that using local reporters in the operation – they wrote the appraisals of the communes with which they are involved on a daily basis – coincides with a twofold downgrading of their status. On the one hand, they are unable to intervene in order to define legitimate topics of discussion; and on the other hand, they have to leave the task of moderating comments to the multimedia editorial team, based in Lille – which in some instances could leave them helpless in the face of criticism from locally elected officials crying defamation. However, more fundamentally, this new relationship with the local area both calls into question the amount of attention the daily newspaper usually pays to each commune, and transforms the press organisation’s role in local democratic life.

Reducing the gaps in the print version’s coverage

51The newspaper does not give the same amount of attention to every commune in the Nord-Pas-de-Calais region. Analysis of the coverage of the printed editions of La Voix du Nord in the ten days running up to the first round of the local government elections show that, relative to population size, it is the medium-size communes which suffer the least amount of coverage by the regional daily newspaper [59] (Figure 7). Conversely, the largest communes experience a high level of over-coverage by the regional daily newspaper: the communes with more than 20,000 inhabitants in fact receive half of the total coverage afforded all of the towns together, yet they are home to only one third of the region’s inhabitants. As regards the small communes, they do not experience any significant under-coverage relative to the size of their population.

Figure 7

Average coverage of the communes by La Voix du Nord

Figure 7

Average coverage of the communes by La Voix du Nord

Note: the dark curve indicates the share of the coverage that the daily newspaper dedicated to the different communes of the region relative to their size during the ten days preceding the first round. The light curve indicates the share of the regional population that resides in the different communes relative to their size.

52Generally, we find that use of the forums counterbalances the varying levels of attention that the newspaper gives to the region’s communes depending on their size (cf. Figure 2). We identify here the very particular situation of the medium-size communes: their inhabitants comment far more in the forums with the effect of remedying the media under-coverage of their commune, which impacts on how local opinion is shaped.

Veil and catalyst

53Analysis of the way discussion unfolds shows that, depending on the size of the commune, La Voix du Nord assumes specific roles by providing the inhabitants with forums. In the small conurbations, the press organisation acts as a “veil”, enabling participants to enter into discussion while at the same time maintaining a veil of ignorance on the identity of each participant. The forum eases the constraints here associated with widespread acquaintanceship, thus facilitating the public expression of opinion. By acting as a veil, La Voix du Nord helps to democratise the communal space, though opening up the discussion to a larger number of participants [60] (a young person, a former expelled municipal counsellor, etc.) and to certain topics that are not usually included in the discussion of local affairs.

54In the medium-size communes, the newspaper plays another role in allowing the inhabitants to appeal to local opinion which, in a context of media under-coverage, is usually expressed with difficulty. Structurally, these communes suffer from a lack of coverage while competition for local power is often strong. By providing these forums, the newspaper is playing the role of a “catalyst” for local opinion, enabling participants to submit the communal reality to their critical judgement and to construct their local political cause. This also helps to strengthen local democracy through relaxing the constraints associated with the relationship that the regional newspaper traditionally maintains with the regional territory. The citizens, as a result, can access greater information on local affairs and the election and depend less on the information controlled directly or indirectly by the locally elected representatives.

55For the large towns and cities, the role played by the newspaper appears less unequivocal. Generally, the provision of online forums here meets a less pressing need, because the citizens enjoy a richer and more varied media environment and because they orient their political involvement more in accordance with national issues. However, the provision of the forums by the press organisation is likely to play a role in local democracy, given that local political life reaches an intensity that the existing media and local forums for debate cannot totally cater for. In this case, the press organisation plays a role that is much the same as that of a catalyst, by enabling citizens to appeal to public opinion. [61]

56Table 3 below summarises the characteristics of the arenas corresponding with the three sizes of commune, which we have updated by means of the quantitative analysis of the pseudonyms and the lexicometric analysis.

Table 3

Three distinct arenas

Table 3
Commune size Fewer than 2,000 inhabitants Between 2,000 and 10,000 inhabitants More than 20,000 inhabitants Role of La Voix du Nord veil catalyst – Nature of relationships between the local inhabitants very strong acquaintanceship weak acquaintanceship very weak acquaintanceship Relative coverage by local media low level of under-coverage high level of under-coverage over-coverage Constraints associated with publicly expressing an opinion widespread acquaintanceship local opinion that is structurally difficult to grasp local opinion that is conjuncturally difficult to grasp Dominant pseudonyms anonymous and local partisan and critical partisan and national Main modes of address interpersonal/marks of politeness and civility opinion of the inhabitants/voters Predominant purpose support/reject denunciation, exposure Nature of political competition competition between individuals clans engaged in battle competing political parties Main level of the discussion family issues – national political categories Openness of the discussion to actors outside the commune firmly closed closed open Nature of issues with the debate anonymity and violence – – Major topics finance, school housing, social security, associations employment, security, transport

Three distinct arenas

57For La Voix du Nord, the provision of these forums results in a simultaneous shift in its relationship with the regions and in the role that it plays in local democracy. For small and medium-size conurbations, such an initiative helps to democratise local political spaces, in ways which differ depending on the size of the town.

58* * *

59Certain experiments conducted in regional press organisations demonstrate therefore that use of the internet does not necessarily lead to the bolstering of institutional and commercial logics. On the contrary, we have shown that a press organisation can capitalise on the web in order to modify the role that it plays in local democratic spaces. For a good many communes, the provision of a forum eases the constraints associated with expressing an opinion publicly and in the end democratises local political life. This effect, we have found, is not uniform and assumes specific forms depending on the morphology of the towns; in other words, depending on the dominant forms of relationships between the local inhabitants, the nature of the local political institutions and the local media landscape. This mainly concerns small and medium-size communes: in providing them with such arenas, the newspaper acts as a “veil” or “catalyst”. In small conurbations, the first way of using the forum (as a veil) tends to hide the identities of everyone, which leads to the emergence of alternative views within a context of widespread acquaintanceship and of pressure to have consensus. In medium-size towns, the second approach (the forum-as-catalyst) contributes to the emergence of public opinion that is struggling to materialise, and to its partisan mobilisation.

60As several works have demonstrated, innovation in press organisations is closely linked to their organisational dimension. [62] In the case of the “mayor appraisal” experiment, the newspaper’s ability to reconfigure a proportion of the local socio-political spaces depends very much on the strength of La Voix du Nord’s regional presence and how it uses its local reporters. Managing the experiment by the multimedia editorial team, located in Lille, largely releases the forums from the game of negotiating relationships between the local reporter and the locally elected representatives. The latter are not able to influence the way in which discussions are conducted, contrary to what often happens in experiments with forums run by the municipalities in which the intervention of local government impacts considerably on participation. [63] For all of these reasons, it is reasonable to assume that only regional press organisations, due to their presence, their organisation and their position both inside and outside the local political spaces – inside due to the presence of local reporters, outside due to the centrally managed website – are in a position to provide forums that allow citizen appropriation to such an extent. [64]

61However, this analysis also leads us inextricably to the conclusion that it is essential to consider how online discussion is rooted in specific social spaces. [65] The material that we have studied in this article is particular insofar as expressing an opinion publicly online is very closely linked with being a member of the commune’s social space. However, one would be mistaken to consider that this represents an exception: generally, online discussion cannot be analysed independently of the socio-political spaces in which it unfolds. From this point of view, it is very interesting to note that the debates which focused on the anonymity or the violence of online political discussion [66] – is it a sign of a democratic regression or progress? – can be settled by analysing more closely the socio-political context within which the discussion is rooted. In one small commune, anonymity will enable people to speak freely, while in another more populous commune, it will play only a negligible role. This demonstrates the benefit of not solely focusing on the social characteristics of individuals when analysing web use, for in so doing one runs the risk of reducing the depth of explanation that social sciences are able to provide regarding this phenomenon, the social and political importance of which is ever-growing. [67]

Figure 8

Map of heterogeneous networks (commune size and semantic categories)

Figure 8

Map of heterogeneous networks (commune size and semantic categories)

Note: the presence of a link between the size of the commune and a semantic category signifies, all things being equal, that this use of this semantic category is preferred in the communes of the size indicated (and vice versa). The size of each node is proportional to the number of comments associated with the node. The final map is available online at the following address:
Table 4

Multiple linear regression (ordinary least squares model)

Table 4

Multiple linear regression (ordinary least squares model)

Note: The dependent variable here is the number of comments that have been posted on the communes’ forums. The model’s explanatory variables here are the number of inhabitants, commune density, change of local government, first round turnout, tenure of the outgoing mayor, the printed daily newspaper’s coverage and the industrial specialisation of the commune. This model explains one quarter of the total variance.


  • [1]
    Cf. Christian-Marie Wallon-Leducq, “L’urne dans le bocal. L’élection municipale partielle de Dunkerque vue par La Voix du Nord”, Mots, 25, 1990, 109-12; Loïc Ballarini, “Presse locale, un média de diversion”, Réseaux, 148-149, 2008, 405-26.
  • [2]
    Cf. Kenneth K. Wong, Pushpam Jain, “Newspapers as policy actors in urban school systems. The Chicago story”, Urban Affairs Review, 35(2), 1999, 210-46.
  • [3]
    Cf. Thibault Courcelles, “Le rôle de la presse quotidienne régionale bretonne dans la création d’une ‘identité bretonne’: étude comparative de Ouest-France et du Télégramme”, Hérodote, 110, 2003, 129-48.
  • [4]
    Cf. Daniel Murphy, The Silent Watchdog. Press in Local Politics (London: Constable, 1976); Jacques Le Bohec, “Les rapports entre élus et localiers. La photographie de presse comme enjeu de pouvoir”, Politix, 7(28), 1994, 100-12.
  • [5]
    Cf. Olivier Roubieu, “Le journalisme et le pouvoir local”, Actes de la recherche en sciences sociales, 101-102, 1994, 85-7.
  • [6]
    Cf. Cégolène Frisque, “Une reconfiguration des espaces médiatiques et politiques locaux?”, Revue française de science politique, 60(5), 2010, 951-73.
  • [7]
    Cf. Éric Dagiral, Sylvain Parasie, “Presse en ligne: où en est la recherche?”, Réseaux, 160-161, 2010, 13-42.
  • [8]
    Cf. Antony Cawley, “News production in an Irish online newsroom: practice, process and culture”, in Chris Paterson, David Domingo (eds), Making Online News. The Ethnography of New Media Production (London: Peter Lang, 2008), 45-60.
  • [9]
    Cf. Thorsten Quandt, “News tuning and content management: an observation study of old and new routines in German online newsrooms” in Chris Paterson, David Domingo (eds), Making Online News, 76-97.
  • [10]
    Cf. Pablo J. Boczkowski, News at Work. Imitation in an Age of Information Abundance (Chicago: University of Chicago Press), 2010.
  • [11]
    Cf. Nicolas Pélissier, “Si loin, si proche: l’élasticité identitaire du journaliste local usager d’internet”, paper, 2002, available at:, last accessed 20 October 2012.
  • [12]
    Survey conducted within the context of a study financed by the French Ministry of Culture and Communication entitled “Local online press and its readership: how has the relationship been reconfigured?” (English translation by M. O’Mahony of the original study title “La presse locale en ligne et ses publics: quelle reconfiguration des relations?”).
  • [13]
    The articles and comments are now partially accessible at the following address:, last accessed 20th October 2012.
  • [14]
    With participants having commented under a pseudonym which they chose themselves, one cannot be absolutely certain that they commented in the forum associated with their own commune. However, with 90% of participants having commented in a single forum, one can reasonably assume that this has predominantly been the case.
  • [15]
    Works in electoral sociology furthermore identify, since the 1990s, a trend towards the “localisation” of French local government elections, in the sense that the campaigns and results appear to increasingly depend on local particularities. Cf. Bernard Dolez, Annie Laurent (eds), Le vote des villes. Les élections municipales des 11 et 18 mars 2001 (Paris: Presses de Sciences Po, 2002).
  • [16]
    Stephen P. Turner (ed.) Emile Durkheim: Socialist and Moralist (London: Routledge, 1993), 111. For a more recent perspective, see Rémi Lenoir, “Halbwachs: démographie ou morphologie sociale”, Revue européenne des sciences sociales, 42(129), 2004, 199-218.Online
  • [17]
    Cf. J. Eric Oliver, “City size and civic involvement in metropolitan America”, American Political Science Review, 94(2), 2000, 361-73.
  • [18]
    For example, see Isabel Boussard, Jean Chiche, “Le comportement des ruraux lors des élections municipales de juin 1995”, Économie rurale, 237, 1997, 10-14.Online
  • [19]
    In their survey concerning the 1983 municipal campaign in the Amiens region, Daniel Gaxie and Patrick Lehingue demonstrate that the expectations of voters varies according to their socio-professional category, age, gender, and size of the town of residence (Daniel Gaxie, Patrick Lehingue, Enjeux municipaux. La constitution des enjeux politiques dans une élection municipale (Paris: PUF/CURAPP, 1984)).
  • [20]
    Louis Wirth, “Urbanism as a way of life”, American Journal of Sociology, 44, 1938, 1-24.Online
  • [21]
    Cf. Robert A. Dahl, Edward R. Tufte, Size and Democracy (Stanford: Stanford University Press, 1973).
  • [22]
    Cf. Vincent Hoffmann-Martinot, “La participation aux élections municipales dans les villes françaises”, Revue française de science politique, 42(1), 1992, 3-35.
  • [23]
    Cf. Paul DiMaggio, Eszter Hargittai, Coral Celeste, Steven Shafer, “From unequal access to differentiated use: a literature review and agenda for research on digital inequality”, report produced for the Russell Sage Foundation, 2004.
  • [24]
    Cf. Dhavan Shah, Jaeho Cho, William P. Eveland, Nojin Kwak, “Information and expression in a digital age. Modeling internet effects on civic participation”, Communication Research, 32(5), 2005, 531-65.
  • [25]
    Cf. Paul DiMaggio et al., “From unequal access to differentiated use: a literature review and agenda for research on digital inequality”.
  • [26]
    Fabienne Greffet, Stéphanie Wojcik, “Parler politique en ligne. Une revue des travaux français et anglo-saxons”, Réseaux, 150, 2008, 19-50.
  • [27]
    Cf. William H. Dutton, “Network rules of order: regulating speech in public electronic fora”, Media, Culture & Society, 18(2), 1996, 269-90.
  • [28]
    Cf. Paul DiMaggio, Eszter Hargittai, W. Russell Neuman, John P. Robinson, “Social implications of the internet”, Annual Review of Sociology, 27, 2001, 307-36.
  • [29]
    Cf. Zizi A. Papacharissi, “Democracy online: civility, politeness, and the democratic potential of online political discussion groups”, New Media & Society, 6(2), 2004, 259-83.
  • [30]
    Thierry Vedel, “L’internet et les villes: trois approches de la citoyenneté”, Hermès, 26-27, 2000, 247-62; Stéphanie Wojcik, “Les forums électroniques municipaux, espaces de débat démocratique?”, Sciences de la société, 60, 2003, 107-25; Marie-Gabrielle Suraud, “Le Net ou la clôture de l’espace public. Débats sur l’accident industriel de Toulouse”, Réseaux, 118, 2003, 211-36.
  • [31]
    Cf. Stéphanie Wojcik, “Les forums électroniques municipaux, espaces de débat démocratique?”.
  • [32]
    Cf. Thierry Vedel, “L’internet et les villes: trois approches de la citoyenneté”, 254. (Translation of the quotation by M. O’Mahony).
  • [33]
    Are Vegard Haug, “Local democracy online: driven by crisis, legitimacy, resources, or communication gaps?”, Journal of Information Technology and Politics, 4(2), 2007, 77-99.Online
  • [34]
    Technology platform of IFRIS (French Institute “Research, Innovation, Society”), specialised in the analysis of textual corpora for social sciences and the humanities. Cf., last accessed 22 October 2012.
  • [35]
    Data sources are detailed in the footnotes throughout the text.
  • [36]
    The task of terminology extraction relied on the grammatical labelling of the texts, from which we extracted the relevant nominal groups. These expressions (in some cases comprising up to six terms) were then lemmatised in order to form larger lexical families. Frequency and specificity criteria applied to the extracted multiterms enabled the most relevant to be identified.
  • [37]
    We based this here on the work of Luc Boltanski who studied several textual corpora in order to analyse the critical skills deployed by individuals. Cf. Luc Boltanski, Love and Justice as Competences (Cambridge: Polity Press, 2012).
  • [38]
    Cf. Jean-Philippe Cointet, “Dynamiques sociales et sémantiques dans les communautés de savoirs. Morphogenèse et diffusion”, social sciences PhD thesis, 2009, École polytechnique, Palaiseau.
  • [39]
    More specifically, in order to calculate the specificity between two entities i and j, one must first calculate the theoretical (expected) number of co-occurrences in the entire corpus which is equal to the overall number of co-occurrences of i and j divided by the total number of actual (observed) co-occurrences in the entire corpus. The value of ?2 is then equal to the square of the ratio between the difference in the number of actual co-occurrences minus this theoretical number, divided by the square root of the theoretical number. In order to retain the information contained in the sign of the deviation between actual and theoretical values, we will subsequently omit the square and use the formula equation im13, with O corresponding to the number of co-occurrences observed and E the number of co-occurrences expected if the entities were randomly distributed across the comments.
  • [40]
    Cf. Marie-Gabrielle Suraud, “Le Net ou la clôture de l’espace public. Débats sur l’accident industriel de Toulouse”.
  • [41]
    Except in cases where the participants introduce themselves explicitly as candidates or former elected representatives, we are unable to establish with certainty what proportion of the contributors themselves are actors in the municipal competition. Although one can assume that this type of actor is over-represented among the participants, the size of the participation rate in the small communes leads one to think that these forums do not only encourage contributions from these actors (cf. below).
  • [42]
    In spring 2008, 68.7% of households in the region had an internet connection at home. This rate varies by 4.8% depending on the size of the commune, by 2.1% depending on socio-professional category, and by 10.9% depending on the area. Source: survey conducted in May 2008 by the market research agency Cegma Topo entitled “Household ICT use in the Nord-Pas-de-Calais region” (English translation by M. O’Mahony of the original survey title “Usage des TIC par les ménages du Nord-Pas-de-Calais”); telephone interviews with 2,500 individuals in Lille and the Nord-Pas-de-Calais region who were representative of the regional population.
  • [43]
    22.1% of people with internet access at home participate in discussion forums. This rate varies by 6.8% depending on the size of the commune, 5.2% depending on the socio-professional category, and by 8.7% depending on the area (source: cf. note above).
  • [44]
    Cf. Vincent Hoffmann-Martinot, “La participation aux élections municipales dans les villes françaises”; J. Eric Oliver, “City size and civic involvement in metropolitan America”.
  • [45]
    Cf. Louis Wirth, “Urbanism as a way of life”.
  • [46]
    With the election taking place after the publication of the comments on the forum, it might seem more intuitive to investigate the effect of the forums on the outcome of the election, as a good number of studies have done. We reject this point of view in so far as the vast majority of the forums generate only a limited number of comments and also in order to avoid any media-centric perspective.
  • [47]
    Forum activity is an ordinal variable with five different modes (no comment; between 1 and 5 comments; between 6 and 30 comments; between 31 and 100 comments; more than 100 comments). Change of local government is a binary variable (0 = no change; 1 = change). Voter turnout in the first round of local government elections is an ordinal variable with 5 modes (less than 60%; between 60% and 70%; between 70% and 80%; between 80% and 90%; more than 90%).
  • [48]
    If one looks at all the communes together without differentiating between them in terms of population size, the correlation coefficients obtained are then subject to the intrinsic variability of the average number of comments per commune size. By nature, a low number of comments are generated in the small communes even though the voter turnout is high. Conversely, the large communes, with a low level of voter turnout, generate, by virtue of their population, a much larger number of comments. Thus the negative Pearson coefficient between voter turnout and online activity obtained by incorporating all the commune sizes together into the same group can be explained mainly by the extensive nature of the measure of online activity used and by the decrease in the rate of voter turnout with the commune populations. This coefficient remains nonetheless positive for all of the population subclasses which enable similar-sized and, in principle, more homogeneous communes to be grouped together.
  • [49]
    Other contextual variables were tested, such as the dominant economic activity within the commune. The analysis reveals a weak positive link between an industrial specialisation and the participation in the forums (cf. Table 4 in the appendices).
  • [50]
    A multiple linear regression using the ordinary least squares method shows that the number of inhabitants is the major variable in accounting for the participation in the forums (cf. Table 4 in the appendices).
  • [51]
    This partition is pertinent from the point of view of the volume of comments generated in each category, however we have also been able to check that the nature of the discussions such as we later characterise it, is sound, no matter how the commune sizes are divided up. Thus we observe, for example, that those communes with fewer than 1,000 inhabitants and the larger communes with fewer than 2,000 inhabitants are characterised by qualitatively similar registers of expression.
  • [52]
    In a survey on the local government election in a very small town in the south of France, Sylvain Barone and Aurélia Troupel show that politics there is the “extension of friendly or conflictual relationships, relationships of trust or mistrust, family or client relationships which develop within the context of local interaction”. See Sylvain Barone, Aurélia Troupel, “Les usages d’un mode de scrutin particulier. Les élections municipales dans les très petites communes”, Pôle Sud, 2(29), 2008, 95-109.
  • [53]
    We find here a result brought to light by Gaxie and Lehingue: in every small commune, the debate more often focuses on the ethical qualities of the candidates (Daniel Gaxie, Patrick Lehingue, Enjeux municipaux. La constitution des enjeux politiques dans une élection municipale, 265).
  • [54]
    We will demonstrate this point later in this article.
  • [55]
    Cf. Élisabeth Claverie “Procès, Affaire, Cause. Voltaire et l’innovation critique”, Politix, 7(26), 1994, 76-85; Nicolas Offenstadt, Stéphane Van Damme, Luc Boltanski, Élisabeth Claverie (eds), Affaires, scandales et grandes causes. De Socrate à Pinochet (Paris: Stock, 2007).
  • [56]
    While studying the relationships between the local inhabitants in the commune of Villefranche-sur-Saône (33,000 inhabitants in 1975), situated in the Rhône department in France, Michel Bozon talks however of a vague but general acquaintanceship, based on rumour, the cafe and the local press. See Michel Bozon, Vie quotidienne et rapports sociaux dans une petite ville de province. La mise en scène des différences (Lyon: Presses Universitaires de Lyon, 1984).
  • [57]
    This is in tune with the analyses of several researchers who highlight a process of nationalising local government elections. Cf. Daniel Gaxie, “Le maire entre disciplines et libertés. Remarques sur les limites du travail politique”, Politix, 7(28), 1994, 140-48.
  • [58]
    Predecessor to La Voix du Nord, L’Écho du Nord developed ten or so local editions in the 1920s. Cf. Jean-Paul Visse, La Presse du Nord et du Pas-de-Calais au temps de l’Écho du Nord (1819-1944) (Lille: Presses Universitaires du Septentrion, 2004), 211-16. For a history of the creation of the local editions at French regional press level, see Marc Martin, “L’invention des éditions locales”, in La presse régionale. Des affiches aux grands quotidiens (Paris: Fayard, 2002), 149-61.
  • [59]
    We built a corpus from all the local editions of La Voix du Nord, which is in a quasi-monopolistic position in the region. We counted, by way of a computer program, all the occurrences (regardless of the citation context) of the name of each town in the region between 22 February and 7 March 2008.
  • [60]
    Given the uncertainty that surrounds the number of contributors who are themselves actors in the electoral competition, it is difficult to measure precisely the degree of openness of the discussion to new actors. Nonetheless, the high rates of participation online attained in the small communes leads one to think that the discussion is not limited to actors directly involved in municipal politics.
  • [61]
    In splitting the corpus into two successive periods grouping together an equal number of comments in each, we noted, in the second period, that the usage profile for registers of expression in the large communes is so markedly similar to that of medium-size towns that they form a single undifferentiated cluster. This supports the argument that these arenas act as a catalyst in large communes when the deadline for local government elections approaches.
  • [62]
    For a history of online innovations in North American press organisations, see Pablo J. Boczkowski, Digitizing the News. Innovation in Online Newspapers (Cambridge: MIT Press, 2004). For an analysis of the manner in which the press organisation influences the innovation of online news formats, see Éric Dagiral, Sylvain Parasie, “Vidéo à la une! L’innovation dans les formats de la presse en ligne”, Réseaux, 160-161, 2010, 101-32.
  • [63]
    Cf. Stéphanie Wojcik, “Les forums électroniques municipaux, espaces de débat démocratique?”.
  • [64]
    One must not however under-estimate the difficulties that necessarily accompany this type of initiative within press organisations. Local journalists often reluctantly consider that their role is to give ordinary citizens a voice. Thus, they often judge that such initiatives fall outside the realm of journalism proper and must therefore remain relatively exceptional. Cf. Jacques Le Bohec, “La question du ‘rôle démocratique’ de la presse locale en France”, Hermès, 26-27, 2000, 185-98; Cyril Lemieux, “Élus et médias locaux: la nouvelle donne”, Pouvoirs locaux. Communication, médias et démocratie locale, 52, 2000, 75-9.
  • [65]
    Dominique Cardon talks of the “realist turn” of the internet. Cf. Dominique Cardon, La démocratie internet. Promesses et limites (Paris: Seuil, 2010), 28-31.
  • [66]
    Cf. Patrice Flichy, “Internet et le débat démocratique”, Réseaux, 150, 2008, 159-85.
  • [67]
    We would like to thank Andreï Mogoutov, Philippe Breucker, Éric Dagiral and Ashveen Peerbaye for their help and advice at the different stages of this research.

For decades, research has been quite sceptical towards the role journalism plays in local political life. This article shows that news organizations can take advantage of the web to play a new role in local democratic arenas. Based on a statistical and lexicometric analysis of online forums set up by a French regional newspaper (La Voix du Nord), this study shows how the morphology of municipalities impacts the way online discussion unfolds. In providing such online discussion arenas to citizens, the news organization plays the part of a “veil” in small cities – playing down the effects of strong acquaintanceship among individuals – and the part of a “catalyst” in medium-sized cities – promoting the rise of local opinion.

Sylvain Parasie
Sylvain Parasie is Assistant Professor of Sociology at the University of Paris-Est Marne-la-Vallée, researcher at the Laboratory for Technology, Territories and Societies (LATTS) and a member of the French Institute “Research, Innovation, Society” (IFRIS). He is the author of Et Maintenant une page de pub! Une histoire morale de la publicité à la télévision française (1968-2008) (Paris: INA Éditions, 2010). His current work focuses on the political issues of innovation in online journalism (Université de Marne-la-Vallée, Cité Descartes, 5 boulevard Descartes, Champs-sur-Marne, 77454 Marne-la-Vallée cedex 2, France).
Jean-Philippe Cointet
Jean-Philippe Cointet is a researcher within the Sciences in Society (SenS) unit of the French National Institute for Agricultural Research (INRA) and also a member of the French Institute “Research, Innovation, Society” (IFRIS). His work focuses on the knowledge communities defined as socio-semantic networks. In particular he is studying scientific communities and conversational dynamics on the web. He participates in the CorText platform of IFRIS (Université de Marne-la-Vallée, Cité Descartes, 5 boulevard Descartes, Champs-sur-Marne, 77454 Marne-la-Vallée cedex 2, France).
Translated from French by 
Michael O’Mahony
Latest publication on cairn or another partner portal
Uploaded on on 03/03/2014
Distribution électronique pour Presses de Sciences Po © Presses de Sciences Po. Tous droits réservés pour tous pays. Il est interdit, sauf accord préalable et écrit de l’éditeur, de reproduire (notamment par photocopie) partiellement ou totalement le présent article, de le stocker dans une banque de données ou de le communiquer au public sous quelque forme et de quelque manière que ce soit.
Loading... Please wait