Introduction
New media optimists have claimed that social technologies, such as the micro-blogging service, Twitter, stand to change political communication in positive ways. Many hope that new technologies can help level the playing field between political elites, who enjoy a number of resources to their advantage, including access to traditional media channels, and non-elites, who struggle to get their messages out (e.g., Bimber 1998; Rheingold 1993). Likewise, there is optimism that new technologies might be harnessed and used to reverse the trend of increasing apathy among citizens in liberal democracies, and particularly among youth (Delli Carpini 2000). In this study, we explore how elected officials use Twitter to interact with the public. Twitter touts itself as a technology used around the globe that is being rapidly adopted by users in almost every country. Therefore, to observe a wide variety of ways in which officials use Twitter, we have conducted a study of officials in three regions with liberal democratic governments. Particularly, we draw upon datasets used in our ongoing research (Hemphill, Otterbacher and Shapiro 2013), in which we follow the tweets of Members of the European Parliament, Korean National Assembly Members, and United States Members of Congress. The goals here, however, are to examine qualitatively the types of vertical communication taking place between elected officials and the citizens they represent and to develop a framework for analysis that can facilitate future work.
Our analysis suggests that, while Twitter provides the infrastructure to facilitate a high level of interactivity between political officials and constituents, not everyone takes advantage of these affordances. We illustrate that Twitter is being used in a variety of ways, from an essentially one-way channel for information provision from official to citizen, to a space in which genuine mutual discourse takes place. In addition, we argue that, despite its image as a “social” technology, many officials use Twitter to engage in para-social interaction (e.g., celebrity-fan interaction), rather than the more substantive human-human interaction. In such cases, officials provide just enough interaction for citizens to respond to them as people (Giles 2002), while yielding little control of the communication situation to citizens. Finally, we offer suggestions for analyzing interactivity on Twitter that considers not only the use of structural features, but also language tactics. Future work can exploit such measures in a large-scale, representative study of officials’ behaviors in order to expand on our findings below.
1. Background and Related Literature
Twitter is now being “used by people in nearly every country in the world” 1 and elected officials in many regions are adopting it as a part of their communication strategy. One way that Twitter might positively impact political communication, and the democratic process more generally, is by promoting vertical communication between elected officials and the citizens they represent. For instance, according to the website of the European Parliament (EP), social media is “revolutionizing” the way that MEPs communicate with citizens. 2 The EP views social media as a means to engage directly with citizens in order to learn what citizens are thinking about and to allow them to “question MEPs themselves.”
1.1. A Trend toward Interactivity?
Do political elites really interact with citizens? Lilleker and Malagón (2010) point out that politicians are simultaneously the party facing the greatest risk and are poised to enjoy the greatest potential rewards from such encounters. For example, interactivity can help the politician establish rapport and a sense of connection with citizens (McMillan 2002b), as she portrays herself as a responsive, capable, and well-meaning representative. However, the risks include the politician losing ambiguity in her political message (i.e., being pinned down by citizens on a particular issue), as well as a general loss of control of the communication situation (Stromer-Galley, 2000).Going a step further, Stromer-Galley (2000) not only finds that politicians are resistant to interactive, vertical communication, but also claims that new technologies allow them to present a façade of interactivity to citizens, thus reaping greater benefits with fewer risks. She distinguishes human-media interactivity (e.g., engaging with content, such as a photo or video, posted by a politician on Twitter) from human-human interactivity that is mediated by technology (e.g., messaging one’s representative and receiving a response). In addition to examining the structural properties of a technology that enable interaction, it is also important to consider its intersubjective characteristics (i.e., what communicators perceive to be the technology’s capabilities) (Van Dijk, 1999). For this reason, Stromer-Galley and Foot (2002) conducted focus groups with citizens before the 2000 elections in the United States on the subject of political candidates’ websites. They found that citizens perceive the possibility for both types of interactivity with political elites on the Web. However, citizens’ needs for interacting with politicians are largely satisfied by human-media interactions with little demand or expectation of direct (i.e., human-human) interactions.
1.2. Interactivity on Twitter
Social media are often assumed to be interactive by their very nature. However, researchers of computer-mediated communication (CMC) consider interactivity to be a variable in any communication setting, and so it is not a characteristic of the medium itself (Rafaeli andSudweeks 1997). Indeed, Twitter structures and conventions of use can facilitate interactivity (e.g., mentioning or directly addressing someone using “@”; the use of “RT” to rebroadcast another’s post). However, as we will show, the provision of these features alone does not guarantee that they will be used in an interactive manner.
The following examples of public exchanges between MEPs and citizens illustrate not only the potential risks and benefits to politicians of engaging in interactive, vertical communication, but also the challenges for researchers in terms of studying interactivity.
Exchange 1:MarietjeSchaake (MEP, Netherlands) and Faceyet (citizen)MarietjeSchaake: Anyone looking through the #SyriaFiles who finds something that needs political follow up, feel free to email me.
Faceyet: @MarietjeD66 Isn’t offering to help with “political follow up” on #SyriaFiles in effect an offer of assistance to #WikiLeaks? #EU #NATO
MarietjeSchaake: No, I’d look into it, take parliamentary action independently on a daily basis, but based on (multiple) info sources... #Syriafiles@faceyet
Exchange 2: Julie Girling (MEP, UK) and Hollicombe (@ToxicTorbay) (citizen)Hollicombe: @juliegirling as our MEP could we ask you about your views on the #hollicombe development in #Torbay, & the possible #publichealthrisk ?
Julie Girling: @ToxicTorbay Thanks for getting in touch. As this is a local planning and development matter I urge you to contact local Cllrs and your MP.
Exchange 3: Alexander Alvaro (MEP, Germany) and Caren S Wood (citizen)Caren S Wood: @AlexAlvaro Sorry me getting personal, but did already someone told you that you look like Mr. George Clooney of the EP? How refreshing!! :))
AlexanderAlvaro: @CarenSWood Life could be worse, hm?
We make a number of observations. In contrast to the latter two exchanges, the public official initiates the first one. She does not directly address anyone in her tweet, but instead extends a general invitation to citizens to get involved and to contact her. In addition, it is the longest of the three exchanges, and might be considered the most interactive by researchers who consider the number of turns taken by communicators, in which messages relevant to the original topic are exchanged (e.g., Rafaeli and Sudweeks 1997). The second and third exchanges are both initiated by citizens, and the addressed officials respond with a single message, essentially ending the conversations.
In all three cases, the officials demonstrate that they are responsive to inquiries from the public. However, particularly in exchanges one and two, their responses show that they are simultaneously trying to “save face” (Goffman, 1967) and preserve their political ambiguity. That is, when Schaake is put on the spot as to where she stands on WikiLeaks, rather than responding directly to the question, she counters that she would rely on multiple information sources before acting politically. Similarly, when Girling is directly asked where she stands on a particular is-sue, and it is clear that the citizen would like to hear Girling’s view (“as our MEP…”), Girling deflects the question, citing that it is a local issue. The third exchange, in which a fan has contacted an MEP to flirt with him, might be considered by some as embarrassing or distracting from the political message or image. However, Alvaro uses the exchange to show his sense of humor. In summary, these examples show that Twitter provides an environment that enables vertical, interactive communication, and that some politicians are using it to engage their constituents.
1.3. Evaluating Interactivity in CMC
Much research on interactivity in CMC takes one of two approaches: analyzing structures provided by the medium or users’ perceptions of its capabilities (Van Dijk 1999). For instance, in the first camp, researchers have focused on the extent to which interactive features are included in politicians’ websites and have tried to understand how communication approaches correlate to party and demographic characteristics (e.g., Jankowski et al. 2005; Braghiroli 2010; Lilleker et al. 2011). In contrast, others have argued that interactivity is not only a variable in terms of the structures provided by a medium, but is also a psychological factor of a given technology (Kiousis 2002). Accordingly, some have approached interactivity from the perspective of users’ perception of a technology’s facility for fostering interactivity (e.g., Stromer-Galley and Foot 2002). Downes and McMillan (2000), taking the user-driven approach, have determined that there are two key dimensions to interactivity: the direction of communication that may take place between senders and receivers of messages (i.e., one-way versus two-way) and the level of control that the message receiver has. McMillan (2002a) subsequently developed a four-part model of cyberinteractivity, which is summarized and related to the case of Twitter in Figure 1. We argue that Twitter use by political officials might fall into any of the four quadrants and use this framework of cyber-interactivity to guide our exploration of politicians’ interactions with citizens via Twitter.
Figure1: Twitter political communication in relation to McMillan’s (2002b) model of cyber-interactivity
Three of the modes of interactivity – monologue, feedback and responsive dialogue – would be described as human-media interaction by Stromer-Galley (2000) and Stromer-Galley and Foot (2002), since, in these exchanges, the citizen interacts with content rather than the politician her-self. Only mutual discourse would be considered as human-human interaction. As previously mentioned, Stromer-Galley suggests that politicians prefer to avoid this type of interaction and tend to use human-media interaction in order to present an image of interactivity to the public. Therefore, we must consider the extent to which political officials engage in mutual discourse on Twitter.
1.4. Para-social Interaction
It is important to note that even the weaker forms of interactivity may have significant effects on audiences. McMillan (2002) describes how para-social interaction can occur as a result of human-content (or human-media) interaction, claiming that, even when there is limited ability for genuine human-to-human interaction, message receivers can develop a feeling of being close to message senders (i.e., in the current study, citizens and political officials, respectively).
The concept of para-social interaction comes out of mass communication research, where the senders are media personalities and the receivers are the audience members. Describing the phenomenon in the context of televised news broadcasts. Levy (1979)has claimed that over time, media users may begin to develop a sense of shared experiences with personalities (i.e., news anchors in the context of his work). Similarly, Norlund (1978) notes that para-social interaction depends on the development of regular and dominant media personalities. Perhaps emotional attachment between television celebrities and audience members is manipulative and is merely an attempt to achieve audience loyalty (Rafaeli1990).
In the context of political officials on Twitter, we should question whether elites use the medium as a space to have mutual discourse with citizens, or if they use Twitter in order to further present a façade of interactivity, as suggested by Stromer-Galley (2000). It may also be the case that officials purposefully remain in the modes of feedback and responsive dialog, in order to create a sense of para-social interactivity, while they limit their use of mutual discourse, or avoid it all together.
2. Goals and Research Questions
While much previous research considered political officials’ use of interactive features in their websites, we are not aware of studies that seek to examine interactivity between political elites and citizens on Twitter. Therefore, using this new medium, our study seeks to explore and characterize the types of interactivity in which politicians are engaging. Inspired by the review of related literature, we propose two research questions:
• RQ1: What is the level of cyber-interactivity of politicians on Twitter?
• RQ2: Do they engage in mutual discourse or do they avoid it?
3. Data and Method
We consulted our datasets of public officials who use Twitter in their communication strategy.For each region (Europe, South Korea, and the US), we identified five officials3 who had been active on Twitter from 1 January 2012 to 1 July 2012. We also considered diversity with respect to gender and political party. Details on the officials selected for the study are provided in Table 1.
Table 1: Public officials, party, gender and lifetime Twitter activity statistics
The dataset comprises nearly all4 of the officials’ tweets during the six-month timeframe. This is a multi-lingual dataset with the American officials tweeting exclusively in English and Korean officials in Korean. In contrast, four of the EU MEPs we studied tweeted in at least two languages, with English being used as a lingua franca. All non-English tweets were translated to English using Google Translate5, and were verified by a speaker of the source language to ensure the appropriateness of the translations.
First, we analyzed the official’s activity and use of Twitter’s structures that enable interactivity:
• How many tweets did the official post during the six-month period?
• How often did the official mention others?
• How often did the official reply to others’ tweets?
• How often did the official retweet?
We also considered the posting of additional content (photos and videos) that fosters human-media interaction. Finally, we read through official’s tweets to find illustrative examples of how officials interact with citizens. In particular, we considered the official’s use of direct reply in an effort to understand whether or not these replies are to citizens and, if so, what they concern.
4. Analysis
For each group of officials, we first present their activity (i.e., number of tweets posted during the first six months of 2012). We also summarize their use of structures: mentioning another user using “@,” retweeting, and direct reply to another user6 . As a measure of human-human interactivity, we also compute the percent of tweets that are replies. Finally, we characterize each official’s interactions with citizens, providing illustrative examples of typical behaviors.
4.1. Members of Parliament (European Union)
4.1.1. AlexAlvaroAlexAlvaro, a German MEP who tweets in both English and German, is very responsive to citizens. As shown in Table 2, he focuses more on engaging directly with others, rather than posting content. We observed many cases where he invited citizens to discuss contents with him and thus engaging in mutual discourse, such as the following:
Table 2: EP officials’ use of structural features supporting interactivity
AlexAlvaro: Why do you need to stay anonymous to be able to express yourself? Must admit that I don't understand that concept... #eurodig
ronpatz: @AlexAlvaro because blowing the whistle can get you in jail or just because your opinion is valid without a name, too. #eurodig
AlexAlvaro:@ronpatz I would disagree about the value of a nameless opinion and...not everyone is a whistleblower (for those I understand) #eurodig
4.1.2. juliegirlingjuliegirling’s interactivity is mainly feedback, with rare direct exchanges. She frequently posts links to content including her monthly newsletter or pictures of events she attended. Even her direct interactions with citizens focus on information provision rather than critical discussion:
treiziemeetoile: @juliegirling quick question if I may: is MrsGirling participating in the EP ASEAN delegation to Burma this week, meeting Aung San SuuKyi?
juliegirling:@treiziemeetoile yes that's correct. Information about the visit to follow in the coming weeks.viahttp://www.juliegirling.com
4.1.3. Rodi_KratsaRodi_Kratsa’s use of Twitter includes monologue and mutual discourse, and she tweets in both Greek and English. She often positions herself on current issues and does not post much additional content. More than 20% of her tweets are direct replies, and we observed several exchanges in which citizens ask her to do something:
billhicks6: @Rodi_Kratsa Could you bring up the case of the suicides in the EP, and ask that they be investigated as a case of murder by negligence or intention? [Referring to the increasing rate of suicide in Greece during the financial crisis.]
Rodi_Kratsa:@billhicks6 Soon, I will be taking other initiatives on the matter of the suicides.
4.1.4. NiccoloRinaldiNiccoloRinaldi tweets primarily in Italian with an occasional English tweet. He typically tweets in the feedback mode, posting content to engage constituents, such as photos and videos. Almost 10% of his tweets are replies, which is consistent with the mutual discourse mode. Of interest was his interaction with citizens during the recent ACTA vote in the EP, such as this exchange with a student:
antodicarlo:@NiccoloRinaldi#A TA . What is this?
NiccoloRinaldi:@antodicarlo Seehttp://www.niccolorinaldi.it for more information. It’s an anti-counterfeiting agreement that would affect Internet freedom as well as access to medicines.
4.1.5. marietjed66marietjed66 is the most active and interactive official we observed. She tweets in both English and Dutch, using feedback and mutual discourse. As previously noted, she often poses questions and invitations to engage citizens. She is also responsive to unsolicited inquiries, such as the following:
AmQamar: Hello! May I ask what you are doing at the European level to solve the problems of the persecutdahmadiyya community? @MarietjeD66
MarietjeD66: @AmQamar we highlight it in our human rights work on Pakistan etc[con’t]
AmQamar: @MarietjeD66 Dont want to hold u up, but just to inform u that I have also visited your website and seen ur work. I appreciate your work.
MarietjeD66:@AmQamar thank you.
While many officials post quick, one-off responses to citizens’ questions, @marietjed66 frequently has extended exchanges. In interactions such as this one, it appears that citizens appreciate the time officials take in responding to their questions.
4.2. National Assembly Members (Korea)
Table 3: Korean officials’ use of structural features supporting interactivity
4.2.1. kangnaraCitizens often initiated interaction with kangnara, who always responded politely. We also noted that he frequently responded with generic replies, such as the following:
kangnara: Thank you for your encouraging words. I’ll do my best to lead.
This is not surprising given that direct interaction with constituents is time consuming (StromerGalley, 2000). Another characteristic of kangnara is that he often posted pictures; we counted a total of 100 photos. Typical photos included him attending official and family events and sports. We characterize kangnara’s use of Twitter as featuring both mutual discourse and feedback.
4.2.2. jinpyokimjinpyokim is very open to answering questions from constituents; nearly a third of his posts are replies. Citizens often want information from him, as illustrated in the following exchange:
sununiv_in:@jinpyokimThe Yeongtong subway construction is often taking too long. Do you care to comment?
jinpyokim:@sununiv_in:By the end of this year, the train line should be constructed all the way to Mangpo Station. Announced in 2000 with construction not beginning until 2006 for various reasons, with no budget problems, there’s been progress. Please be patient for just a little while longer.
We also observed him interacting with students who asked him to complete a survey, and he quickly responded to their request; jinpyokim spends a good deal of his time in mutual discourse, and engages in the feedback mode of interactivity as well.
4.2.3. KimYoojungKimYoojung is very prolific, typically posting several tweets each day. In addition to professional activities, she often mentions day-to-day details, which add a personal touch, such as the following comment:
“I’ve got to have a strong, sugary cup of coffee! Even when there’s a lot going on, a strong cup starts the day!”
Over half of her tweets are responses to others in mutual discourse. Many of these interactions are, similar to those of her colleagues, words of encouragement and appreciation. For example:
lafe12: @KimYoojung: Senator, the last four years have been difficult and filled with anxiety. Thank you for your hard work.
KimYoojung: @lafe12: Thanks~^^ We’ve missed you! How have you been?
4.2.4. KwonYoungGhilWhile KwonYoungGhil is quite prolific, he exchanges very few messages with others. He fosters human-media engagement by occasionally posting photos and videos. He tweets about strikes and economic injustices, positioning himself in relation to the events or issues, as follows:
KwonYoungGhil: hildren are often referred to as “the treasures of our country”. Where to spend money if not on them? Free childcare should not be interrupted.
KwonYoungGhil’s use of Twitter falls mainly into two modes: monologue and feedback.
4.2.5. YoungahParkYoungahPark is the least prolific of the Korean officials we studied. Her tweets often focus on issues of education and her own views, without prompting a reply from citizens:
YoungahPark: Teacher evaluation in the Teacher Evaluation Bill is now being discussed at the curriculum general meeting. It’s a shame that it was unanimously supported three years ago at the meeting, but it still has not been passed. The situation is very frustrating. I apologize to those who are waiting for this bill to pass.
We observed YoungahPark using monologue, feedback and to a lesser extent, mutual discourse.
4.3. Members of Congress (United States)
Table 4: US officials’ use of structural features supporting interactivity
virginiafoxx exhibits a press agency style of use rather than an effort to interact directly with citizens. For instance, we observed the following message multiple times:
virginiafoxx: Help me reach 2,000 likes on Facebook! If you follow me on Twitter, be sure to check out & like my Facebook page! [URL]
She had no direct exchanges with anyone. She often mentions and retweets other political elites; however, we were unable to find a single mention or retweet of a citizen. virginiafoxx tweeted several photos, which often depicted visits to companies and other organizations. It is clear that virginafoxx avoids mutual discourse,and her primary mode is feedbackwith limited responsive dialog.
4.3.2. SenGillibrandSenGillibrand’s activity takes a personal tone but still primarily is in feedback mode. Her tweets are generally written in the first person, and she often posts photos of professional and family activities, as in the following:
SenGillibrand: Last night, I took Henry, Theo & a friend to Congressional Night at the Natl@AirandSpace Museum. They loved it! pic.twitter.com/pNswZejr
While SenGillibrand does avoid mutual discourse, she often uses mentions to give kudos to civic groups and individuals involved in work and causes that she supports:
SenGillibrand:Congrats @ReshmaSaujani on the amazing @GirlsWhoCode project in #NYC& its new partnership w/@Twitterhttp://bit.ly/MQOvok #offthesidelines
4.3.3. clairecmcSimilar to SenGillibrand, clairecmc’s tweets have a personal tone, generally written in the first person voice. She occasionally posts photos, often with family members. While clairecmc had relatively few replies to others (6.2% of her posts), we observed some interesting interactions with citizens. For example, in one instance, she defends herself against a citizen’s criticism:
mrsdeedum:“@FSMidwest: @clairecmc Got her GAME 6 ticket signed [URL] but who paid 4 it? You or lobbyist?
clairecmc: @mrsdeedum I paid for my own ticket. Always do.
In summary, claircmc is primarily tweeting in the feedback mode, with some mutual discourse.
4.3.4. ThadMcCotterWhile ThadMcCotter interacts primarily with other politicians and the media, he engages in exchanges with citizens, which he often initiates, such as the following example:
ThadMcCotter:Lunch with one of Michigan's finest at USAG-Yongsan. #TM12[URL]
AndrewHemingway: @ThadMcCotter killer bow tie! If @repschock will stop hogging GQ I think you have a chance
Similar to SenGillibrand and clairecmc, ThadMcCotter tweets about both professional and personal interests. However, whereas the former often tweet about their families, ThadMcCotter often had exchanges about TV shows or his home sports teams. His tweets are primarily written in the first person voice, and he makes extensive use of both mutual discourse and feedback.
4.3.5. SenSandersSenSanders discloses that his staff members generally tweet for him. Tweets written by him personally are clearly denoted as such. We counted 65 tweets that were written by Senator Sanders (7.9%).Like virginiafoxx, SenSanders does not use direct replies. The dominant voice of the posts is the third person, likely because staff members usually write them. SenSanders is also fond of posting questions that provoke citizens to think about an issue, and engages them with additional content:
SenSanders:The CEOs of 15 top U.S. and European banks got an average raise of 12% last year. Did you get a raise last year? [URL]
SenSanders extensively uses the feedback mode of interacting; however, he also makes use of responsive dialog, as in the following examples in which citizens are invited to participate in polls:
SenSanders: Should the US continue to subsidize the fossil fuel industry? Let Bernie know here: [URL]#Energy#Oil#Gas
5. Discussion
Twitter provides a number of features designed to facilitate interaction. Some promote humanmedia interaction (e.g., posting a URL or photo) while others enable direct, human-human interaction (e.g., mentioning, which often leads to a reply). Structurally, Twitter has the capacity to put citizens in direct contact with their representatives. However, the provision of the functionality alone does not guarantee that the medium will be fully exploited.
We identified several officials who regularly engage in mutual discoursewith citizens. Many of them exhibit a willingness to answer inquiries in a polite and timely fashion. Even more encouraging, some, in particular AlexAlvaro and Marietjed66, explicitly invited citizens to discuss issues with them and engaged in more than simple, one-off exchanges.
5.1. Para-social interaction
On the other hand, we observed other officials who remained in the feedback and responsive dialog modes. For instance, virginiafoxx and SenSanders had no direct exchanges with others. Both promoted their Web presences elsewhere in order to drive traffic there (e.g., Foxx’s Facebook) or to collect feedback from the public (e.g., Sanders’ polling site). Many also used Twitter in monologue mode, simply posting updates and views on current events and issues.
It may be that many politicians, despite having adopted Twitter, have no desire to engage in mutual discourse. What do these officials gain by using social technologies in ways that are less than fully interactive? As previously mentioned, even in communication situations when there is limited ability for human-to-human interaction, message receivers can develop feelings of being close to message senders. Thus, some politicians interact just enough to get constituents to identify with them, without having to yield much control in the exchange, and without having to invest the energy necessary to sustain mutual discourse.
5.2. Analyzing Interactivity
We examined the extent to which officials use interactive features of Twitter, but our qualitative analysis revealed a need to further explore how officials use language in conjunction with Twitter’s structures. We observed how MarietjeD66 and AlexAlvaro posed provocative questions or invitations to encourage constituents to interact. Likewise, we saw the importance of “thank you’s” issued promptly in response to citizens’ inquiries. Given Hyland’s (2005) claims that writers relate to readers through linguistic tactics that foster engagement, we detail three of his five key features of textual engagementin Table 5 and includeexamples of the features of engagements from tweets of political officials.
Table 5: Linguistic features of engagement (adapted from Hyland (2005))
Our future research will systematically study how officials use these tactics, as we expect that the combination of engaging language use with Twitter’s structures is crucial for understanding how officials directly interact with citizens. As shown in the first example, a strategic use of pronouns might be used in order to create a sense of solidarity with citizens. Likewise, directives that ask citizens to do or think about something might enhance the feeling that the official is communicating in a more personal way with the citizen, rather than simply conveying information. Finally, the use of provocative questions, as indicated in the third example, might be used as a means to incite citizens to discuss an issue. In our future work, we plan to consider how often politicians use such linguistics features in their tweets, and how the patterns of use correlate to their dominant style of interaction (e.g., primarily mutual discourse versus feedback).
6. Conclusion
Politicians’ Twitter use varies in terms of how interactive they are. We observed interesting cultural differences that warrant further study: American officials were significantly less interactive as compared to Europeans and Koreans. This can be confirmed with a large-scale study of such patterns.
In conclusion, Twitter has much potential for promoting interactive, vertical communication. Of course, it’s unrealistic to expect that all or even most officials will use Twitter in a highly interactive way, and we observed officials who did not exploit its interactive potential. Previous research, focusing on politicians’ use of websites, suggested that citizens are aware of the interactive potential of new technologies, but do not demand interaction with their representatives (Stromer-Galley and Foot 2002). An update on this work is much needed in light of the use of new social technologies. Hence, future work should also focus on citizens’ expectations and experiences in engaging their representatives. We hope that such further work might show positive examples of interactive communication that will, at the very least, encourage citizens to try to engage officials via new media, such as Twitter.