Good girls don't comment on politics? Gendered character of online political participation in the Czech Republic

ABSTRACT

The article aims to contribute to the still relatively unexplored area of the relationship between gender and online political participation. Using two complementary methods – a representative, post-election survey of the adult Czech population and a content analysis of communication on the selected Czech political parties’ Facebook profiles during the campaign for the 2013 Parliamentary Elections – we attempt to challenge some established assumptions regarding the allegedly equalizing effect of the Internet and social media on participatory behaviour of men and women. While survey data discovered subtle yet statistically significant differences between men and women in some online expressive activities on Facebook, mainly commenting on other users’ statuses, content analysis further revealed that there are not only notable gender gaps among the Facebook users who commented on the campaign, but also differences in the tone of communication produced by the respective gender groups, with men posting more negative comments addressed to parties as well as to other Facebook users. We suggest that these results question the prevailing perception about the narrowing of the ‘gender gap’ in the online environment and call for a more nuanced methodological approach to different forms of online political expression.

KEYWORDS:

• online political participation
• gender
• election campaigns
• Facebook
• survey
• content analysis

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes on contributors

Lenka Vochocová, PhD, is a lecturer and researcher at the Department of Media Studies, Faculty of Social Sciences, Charles University in Prague. Her research interests cover the fields of political economy of communication, gender media studies and public sphere theories. [email: [email protected]]

Václav Štětka, PhD, is Senior Researcher and Leader of the PolCoRe research group at the Institute of Communication Studies, Faculty of Social Sciences, Charles University in Prague. His research interests encompass political communication and the role of new media, transformation of media systems, and issues of media ownership and globalization. [email: [email protected]]

Jaromír Mazák is a lecturer and a PhD student at the Department of Sociology, Faculty of Arts, and member of the PolCoRe group, Faculty of Social Sciences, Charles University in Prague. In his dissertation project, he explores issues of civic participation and new political movements. [email: [email protected]]

Notes

1. >According to Strandberg (2013), the Finnish citizens who actively follow Facebook and Twitter content in a direct connection to the election campaign challenge the traditional patterns – they are more likely to be young, less educated and female (p. 12–14). Similarly, Gil de Zúñiga, Veenstra et al., 2014 stress the importance and unexpectedness of their finding that women express ‘greater levels of [political] participation’, at least among those who read Internet blogs (p. 46), although the authors admit certain methodological limitations.

2. >A significant gender gap was recently discovered within the UK Twittersphere; according to BBC Trending, male Twitter users have been using political parties’ official hashtags overwhelmingly more often than do women, with 75% political party tweets coming from males (according to http://www.bbc.com/news/blogs-trending-32137886, last accessed 8 July 2015).

3. >The survey was administered by Public Opinion Research Centre (http://cvvm.soc.cas.cz/en/), one of the largest polling agencies in the country.

4. >The decision to analyse only the first 30 reactions to each party status was an arbitrary one (although not unprecedented – similar approach was chosen by Iñaki Garcia-Blanco in his analysis of citizens' comments of selected newspapers’ Facebook sites, see Garcia-Blanco 2014). At the same time, we intended to approximate an average volume of reactions to party statuses during the election campaign (the median number of users’ contributions per party status ranged between 21–56 for the seven parties in the sample). Given that we archived Facebook content with a two-day delay (i.e. on Wednesday we archived content from Monday, on Thursday for Tuesday, etc.), using Facebook's ‘most relevant’ feature, these 30 comments and replies did not correspond to their chronological order but instead were sorted based on their level of interactions (likes, shares). In other words, for each status we have analysed the 30 reactions which have drawn most attention by the users.

5. >The questions were not limited by a specific time frame such as ‘last month’ but referred generally to the duration of the election campaign.

6. >As ‘positive’ we coded those comment/replies that expressed support for the party or the candidate, including an agreement with their policy/programme/statement, etc. ‘Negative’ were those ones critical to the party/candidate, or expressing disagreement with their policies or programme. ‘Ambivalent/neutral' were those contributions neither decidedly positive nor negative, including those expressing both positive statements and criticism (e.g. ‘I like your videos but unfortunately you don't exclude collaborating with party X’; or ‘I would vote for you but without Mr XY on your list’). When coding the valence of the comment towards another Facebook user, we only included those statements directly addressing the user, by using either their name or pronoun (e.g. ‘you are totally wrong’), instead of those comments expressing a general reaction to the previous post without explicitly referring to another user (e.g. ‘this is true’, ‘I don't agree with that’).

7. >Intercoder reliability was tested on a sub-sample of 227 coding units (users’ comments and replies), drawn from the population of party posts by systematic random sampling (every 30th post was included; the first one was selected using a random number generator). In the absence of universally accepted sample size for intercoder reliability testing in a social media environment, we considered such sample size adequate, complying with the general recommendation by Lombard et al. (2002, 601) who suggested that a sample between 50 and 300 units should be satisfactory in most cases.

8. >It is possible that the social desirability bias influenced the replies of female participants in the survey who may have exaggerated when answering question concerning their online engagement. Dalton and Ortegren (2011), for example, conclude based on several studies that females are more prone to the social desirability response bias.

9. >Income was measured as net personal income by asking the respondents to choose from 17 income classes. The difference between women and men was 1.9 income classes. It was tested by independent-samples t-test, equal variances assumed, t = 6.24, df = 494, p < .001). Note that 33% of the cases were excluded from the analysis (24% did not answer the question about income and 9% claimed to have no net personal income).

10. >The relationship between political commenting on Facebook (binary index) and income was analysed with a t-test (equal variances not assumed, t = 0.196, df = 268, N = 486, p = .845). The relationship with age was also tested with t-test (equal variances assumed, t = −0.083, df = 717, p = .934). The relationship with education categorized into four levels was tested with Pearson chi-square test (N = 729, p = .644).

11. >The error margins in the chart should not be compared among each other as the categories are not exclusive, but each category also includes all contributors to its right.

Additional information

Funding

This research was supported by the Czech Science Foundation (GACR) [Standard Grant Nr 14–05575S] under the topic ‘The Role of Social Media in the Transformation of Political Communication and Citizen Participation in the Czech Republic'.