You are on page 1of 38

Micro-participation: The Role of Microblogging in Planning Jennifer S.

Evans-Cowley, PhD, AICP Ohio State University Greg Griffin, AICP Capital Area Metropolitan Planning Organization Abstract: Planners are challenged to engage the public in meaningful ways to shape planning processes. There has been rapid growth in the use of online social networking sites, such as Facebook and Twitter, and planners are struggling to understand how best to use these technologies to engage the public. This study examines more than 49,000 microblogs posted on Twitter and other social networking sites, tracked by Social Networking and Planning Project (SNAPP) to determine public engagement in the Strategic Transportation Mobility Plan (ASMP) in Austin, Texas. Using a mixed methods approach, relevant microblogs were examined to determine sentiment, microblogger engagement, extent of engagement, and impact on the decision-making process. This study found that there are methods that can be used to analyze micro-participation and that micro-participation can be effective in generating participation, but it faces substantial technical, analytical, and communication barriers to influence decision-making. Keywords: public participation, social networking, transportation planning, Twitter, microblogging, sentiment analysis Source for the research: This research was supported through a Federal Transit Administration Public Transportation Participation Pilot Program research grant. FTA PTP Cooperative Agreement #TX-26-1002 Authors: Jennifer Evans-Cowley, PhD, AICP (cowley.11@osu.edu) is an Associate Professor and Head of City and Regional Planning at The Ohio State University. Her research focuses on the use of technology to enhance participation in planning. She has written extensively about the role of e-government in planning. Evans-Cowley served as the external evaluator on the SNAPP project. Greg Griffin, AICP (greg.griffin@campotexas.org) is a senior planner with the Capital Area Metropolitan Planning Organization and an adjunct faculty member of the Department of Geography at Texas State UniversitySan Marcos. Griffin represented the regional metropolitan planning agency as a member of the SNAPP advisory committee.

Electronic copy available at: http://ssrn.com/abstract=1760522

Micro-participation: The Role of Microblogging in Planning Introduction In its recent report, The Future of Cities, Information and Inclusion , the Institute for the Future asks a critical question: How can social networks and social media be leveraged to engage broader participation in building smart infrastructures that meet the needs of diverse users? (Townsend et al., 2010). While raised in the context of information technology infrastructure, it could also be asked of other types of infrastructure. The City of Austin, Texas asked this question as it undertook its Strategic Mobility Plan (ASMP) process in advance of the citys November 2010 transportation bond referendum. Effective public participation in planning projects shares three common purposes (APA 2006): 1) involving citizens in planning and design decision-making processes, 2) providing citizens with a voice in planning in order to improve plans and decision-making, and 3) promoting a sense of community by bringing people who share common goals together. Like most cities, Austin faced challenges in engaging with the public to achieve those three purposes, including getting the representative public to attend public meetings, getting the word out about ongoing planning projects, and encouraging meaningful participation. Over time, Austin tried a wide variety of engagement techniques, including online social networking (CAMPO, et al 2009, City of Austin 2009, Evans-Cowley 2010; Goodspeed 2010) and was again looking to social networking as part of its larger engagement strategy.1 The City wished to experiment with online micro-participation, a method to engage many, unconnected individuals (Haythornthwaite 2007, 1) while minimizing time and opportunity costs to personal involvement. Through a grant from the Federal Transit Administration, the City, Capital Metro, and the Capital Area Metropolitan Planning Organization (CAMPO) partnered with the Texas Citizens Fund to create Social Networking and Planning Project (SNAPP). SNAPP was charged with piloting, tracking, and evaluating the use of an integrated array of tools to build relationships through online social networking to increase the quantity and quality of participation as part of the Austin Strategic Mobility Plan (ASMP) process between April and October of 2010.
1

This article focuses on SNAPPs online engagement. There were public meetings and other forms of public engagement, but these were part of their standard engagement process and are not the focus of this article.

Electronic copy available at: http://ssrn.com/abstract=1760522

The City through this experiment wanted to know whether something intelligent, something worthy of mindshare *can+ occur in the space of 140 characters (Carr, 2010). This research study sought to answer that question. Specifically it asks whether the results of microparticipation could be measured and analyzed to create understanding of the publics view on transportation topics. Was micro-participation equitable? Did the results of the microparticipation effort result in meaningful input that was utilized to support decision-making among policy makers in Austin, Texas? This article begins with a discussion of microparticipation, which is followed by an explanation of SNAPP, presentation of methods, sharing of the results, and a discussion of the implications that this case study has on the future of micro-participation.

Engagement Through Micro-Participation Efforts to mandate public participation have been effective in increasing participation, but there is still room for improvement (Brody, Godschalk, and Burby 2003). As described by Arnstein (1969), the rungs of public engagement efforts vary based on the actual use of involvement to affect outcomes. As a ladder that must be climbed beginning with the bottom, Hanna (2010) recommends the first few rungs be strong with information on which to base discussions for citizen participation and intergovernmental information sharing. As planners we agree on the need for participation and information, yet it is a challenge to get people to participate. Laurian (2004) found that many planning processes suffer from low participation simply because citizens do not know about the issues or public meetings. Transportation is one area where the struggle for participation still exists (Bullard et al., 2010). One of the key dimensions that planners must consider when looking at participation is the required investment of time and/or personal research that a citizen must invest for a given participation tool. For example, a public meeting may require an hour of time, plus travel and opportunity costs. On the other extreme, the microblog requires only a few seconds. In reality, planners should consider several participation methods that offer citizens a choice of their time investment, illustrated in Table 1. Bailey and Grossardt (2010) propose that one way to create a stronger base of information to support participation and to bridge the Arnstein Gap

between degrees of tokenism and citizen power can be achieved through online tools, such as public participation geographic information systems (PPGIS) (Tang and Waters 2005). [Insert Table 1 about here] Online social networks represent one type of tool that can be used to support interaction between groups of people who share a common interest, for example in this project using online social networks to support transportation planning. One of the features of many social networking tools2 is that they allow participants to microblog, the posting of short content, such as phrases, quick comments, images, or links to URLs, photos, audio, or video. For example, in both Facebook and Twitter, a user can post a microblog, such as Im stuck in traffic on 315, which can stand by itself or be tied to additional information, such as a picture or video of the traffic jam or a link to a news story about traffic problems. This allows the user to share information to those within their social network friends on Facebook and followers on Twitter. Microblogging has tremendous popularity, Facebook has more than 500 million users, while Twitter has more than 100 million users, because of their ease of use and ability to share information in real time either on a computer of a mobile device (Boyd et al, 2010; Facebook, 2011; Java et al, 2007; Kwak et al., 2010; Smith and Rainie, 2010; Wright, 2009; Zhao and Rosson, 2009). For example, 40 percent of Facebook users and 76 percent of Twitters users post from their mobile devices (Lenhart and Fox, 2008; Facebook, 2011). Because of the widespread use the demographics are becoming more representative of the population. A 2008 study of Twitter found a median user age of 26, older than the users of Facebook and MySpace, and Twitter users are more diverse than the population as a whole (Lenhart and Fox, 2008). See Table 2 for examples of common microblog terminology. One challenge for the planner is how to build an online relationship with a segment of hundreds of millions of users that would then lead to micro-participation, the capture of microblogs and then reaching out to the microblogger to engage in dialogue. Facebook is an example of a closed social network that relies on users to make individual determinations of whether they wish to participate. For example, SNAPP has a Facebook page called snappATX.
2

There are a wide variety of social networking tools available, including Facebook, LinkedIn, and Twitter, among others.

Individuals in Austin can learn about SNAPP, decide that they wanted to go to the snappATX Facebook page, and then become a fan. At that point, the fan will see SNAPPs microblogs. Then, the individual could take the initiative to comment on a particular microblog post. For a microblogger to engage they must have a strong attachment to community, from either a highly local community of contributors, a larger community of interest around a topic, and a societal community committed to open access (Haythornwaite, 2009). In addition, a planner cannot see microblog posts of a fan unless the planner is part of the fans network of friends. This makes it impossible to identify citizens who may be microblogging about specific planning-related topics on closed social networks. Previous research found significant challenges for planners attempting to use Facebook because it is a closed social network and planners have not yet been effective in creating connections with the public (Evans-Cowley and Hollander, 2010; Evans-Cowley, 2010). In a study of microblogs, researchers found that the connections that underlie the declared set of friends and followers is a key driver of their engagement choices (Huberman et al., 2008). People also participate in online because they enjoy it (Bonaccorsi and Rossi, 2004; Ghosh, 1998; Hars and Ou, 2002; Hertel et al., 2003; Lakhani and Panetta, 2007). They also to contribute for altruistic reasons (Lietsala and Joutsen, 2007). Planners have to find ways to connect with the public providing opportunities that are enjoyable and tie into their altruistic nature. There is growing recognition that Twitter can be a powerful tool to engage the public (Anon 2009, Berg and Halbur 2010, Carr 2010, Cheng et al 2009, and Honey and Herring 2009). Twitter is an open social network that allows users to send messages, known as tweets, from a computer or mobile device. Twitter users can post microblogs of no more than 140 characters, which are shared with others who have signed up to receive them (Lenhart and Fox, 2008).3 In addition, all of the tweets can be searched unless a user puts privacy settings in place. For example, if a user microblogs Traffic on Mopac in A ustin is heavy, then a planner could find this microblog by searching for Austin and traffic. The planner could then respond asking how often they experience traffic on Mopac. One of the important elements of Twitter is that a user can follow anyone with little expectation of reciprocity, meaning that
3

Because of the limited number of characters, users often use shorthand to communicate in their microblog; for example, shortening Austin, Texas to ATX.

people routinely follow microbloggers whose tweets are often full of links to incredibly vital, timely information (Carr, 2010). Planetizen, for example, has created a list of Twitter users to follow who post about planning issues (Berg and Halbur, 2010). This means that Twitter users have an expectation that people they do not know may be following them. The ability of a planner to read, see, or listen to what everyday citizens are saying about their community offers an opportunity to engage the public in a very different way. One could compare reading and responding to microblogs to eavesdropping, where a planner might be in a coffee shop and overhear a casual conversation between friends about the challenges they encountered in walking to meet each other for a drink. The ability of planners to tap into the extensive world of microblogging offers a rich opportunity to reach an audience that may not otherwise participate in planning for the future of their community. Planners want to be able to engage microbloggers in dialogue and to share information. Retweeting represents the forwarding or diffusion of information to new audiences (Boyd et al. 2010). Retweeting is often associated with sharing information with a specific audience, commenting on someones tweet, agreeing with someone, or saving tweets for future access. Tweets with URLs, which adds additional content, and hashtags are the most likely to be retweeted and the most popular topics are headline news (Suh et al., 2010; Kwak et al., 2009; Zarella, 2009). Diffusion is significant and almost instantaneous, with retweeted microblogs reaching an average of 1,000 users (Kwak et al., 2009). If planners are trying to share important information, our ability to have our followers retweet our microblogs to other networks of followers has significant value. However, it should be noted that retweeting is difficult to directly measure because while RT is a standard notation for retweeting, not all users use RT. In addition, it can be difficult to find all the subsequent retweets of an original tweet (Suh et al., 2010). The communication style a planner would use in micro-participation may differ from approaches used in other forms of participation. The best Twitter users are able to communicate with extreme economy and precision, creating significant meaning in the minimum number of characters (Carr, 2010). For example many public engagement efforts generate more heat than light and skillful mediation is needed to generate beneficial outcomes

(Forester, 2006). This is a particularly important consideration when it comes to microblogs, where there are so few characters with which to convey meaning and it can be easy for parties to read emotions into the messages. Using sentiment analysis on more than 150,000 posting, researchers were able to determine sentiment about brands (Jansen et al., 2009). In fact, the aggregation of microblogs in three countries accurately predicted the winner of national elections (Suenami & Yutaka, 2010; Tweetminster, 2010; Turmasjan et al., 2010). But are microblogs in aggregate representative of the sentiments of the public as a whole? One study found that the sentiments of tweets on politicians or political parties mirror the off-line political sentiments (Tumasjan et al, 2010). Although online discussion exhibits some deliberative characteristics, it is often inequitable, with a focus on non-substantive issues, and unconstructive engagement between bloggers (Jansen, 2009). How do we effectively engage the citizenry in an equitable and constructive dialogue? Novek (2009) argues that in order to be effective in online engagement, engagement must have four factors: asking the right questions, asking the right people, designing the process for the desired end, designing the process for groups and focusing on policy making rather than websites or web tools themselves. Design matters, in that the project design must be created in a way that fosters a collaboration system that brings together the stakeholders and the policy makers. The researchers expectation of SNAPP is that they would be successful in generating significant public engagement through micro-participation.

SNAPP Platform In order to support micro-participation, SNAPP created a website that served as a platform that centralized the social networking tools. SNAPP used the http://snappatx.org website as a launching point, with links to its Twitter, Facebook page, and a blog that integrated an automatic feed from partners blogs, as well as e-mail (conversation@snappatx.org), surveys, and external resources.

SNAPP captured all comments from SNAPPs blog, any comments on SNAPPs partner blogs4 that were tagged with #snappatx or @snappatx, and automatic (Atom and RSS format) feeds to news blogs that primarily deal with transportation. The SNAPP website includes an application that processes incoming requests for recent microblogs. The main Conversation page of the website lists the latest snappsthe microblogs that include the keyword snappatx. A pull-down menu allows further searches of microblogs by the transportation mode topics buses and rail, congestion and cars, and walking and bicycling. Both users and administrators are given a real-time, relevant Internet search of transportation issues particular to the Austin region in an easily understood format. SNAPPatx used standard conversational protocol in Twitter to engage in microparticipation. Upon reading a relevant microblog, a SNAPP facilitator would respond by as shown in the following example: Microblogger: I miss the GW parkway bike path in DC. Austin's great, but nothing like an 18 mile paved, straight-ish, scenic, safe (no cars) bike path. SNAPP Facilitator: @ajepst What's the best surface to ride on? Does paving make a big difference to you in bike trail design? #snappatx The SNAPP facilitators add the hashtag #snappatx to every microblog they post to enable searching for relevant conversations. SNAPP established partnerships with some of these third party sources, who agreed to tag the posts they would like SNAPP to capture in its feed with the keyword snappatx. SNAPP developed a capture tool that takes the microblogs (based on keywords) and stores them in a database for analysis. The SNAPP administrator can then export the microblogs meeting certain criteria into a Microsoft Excel spreadsheet for conversion and analysis. SNAPPs communication team focused on keeping ASMP in discussions through incrementally educating its component parts and pushing out information relevant to transportation-related topics. For example, the focus in April and May was on educating about

Partner blogs included Austin Urban Rail, Alliance for Public Transportation, Capital Metro, Placemaking Institute, East Austin Voice, Gentrified Transit, Fare Enough, Statesman (Ben Wear), Austinist, Community Impact News (Central Austin), Community Impact News (Northwest Austin), Community Impact News (Southwest Austin)

multimodal transportation impacts on placemaking, while the focus in June and July was on exploring ASMP objectives on a week-by-week basis. In August, participants were informed about council deliberations on whether to call for a bond election. In September and October, the focus was on keeping participants aware of the debates arising from the bond election. There are a limited number of instances that the ASMP was directly referenced. SNAPP decided that the direct use of ASMP was not effective because most people did not know enough about it to use it as a label to represent shared understanding. Additionally, the City of Austin was responsible for branding the ASMP, so there was a desire to separate the Citys efforts from SNAPP. SNAPP felt that it was equally effective to raise and encourage discussion of the issues and topics that underpin the ASMP. This included weekly facilitation strategies that explored each of the ASMP objectives (Glenn Gadbois, personal communication, January 10, 2011). All of these items were achieved by engaging microbloggers. In addition, two micro-surveys were completed to focus on questions that the City staff wanted answered in regards to the ASMP and the proposed package of projects that could be included in the bond package.

Methodology SNAPP provided the researchers with 49,421 raw microblogs, 5 as well as the processed microblogs, results of surveys, and any other data requested. In addition, extensive correspondence between the principal researcher and the SNAPP administrator occurred between December 2010 and January 2011 to clarify data collection procedures, communication protocols, and communication between SNAPP and the City of Austin. The researchers then conducted further analysis. This study seeks to answer the question of whether micro-participation can be analyzed to help understand the publics views on transportation issues. In order to answer this ques tion, this study uses a multi-part mixed methods analysis of the microblogs. Was micro-participation equitable, with a minimal number of heavy users? Did the results of the Austin microparticipation effort result in meaningful input that was utilized to support decision-making among policy makers in Austin, Texas? The methods are described further below.
5

The microblogs were harvested from SNAPPatxs Facebook and Blogger pages. Tweets were collected through searches on Twitter.

Coding Protocol SNAPP systematically collected 49,421 microblogs that contained both Austin (atx or other abbreviation) and transportation-associated terms.6 SNAPP reviewed each of the microblogs for relevancy. The principal researcher of this article examined and commented on the protocol described below in April 2010, when the first month of microblogs were collected to ensure that the coding protocol was appropriate. SNAPP determined that microblogs were irrelevant if they were not related to both Austin and a relationship to the planning objectives of the ASMP. For example, Austin and I are taking the train to Boston is irrelevant because the reference is to a person named Austin rather than the city. The microblogs were reduced to 11,500 relevant microblogs, including 8,308 from general microbloggers, 2,173 from SNAPPatx (1,007 facilitating conversation and 1,166 sharing information), and 1,019 from media sources. The 8,308 microblogs from general microbloggers were coded on five variables, including type, theme, topic one, topic two, and positive/negative sentiment. Type applies to the type of comment, which can be sharing, engagement, or analytic. Sharing microblogs are those that share information, such as RT @foxaustin: Listen up UT students. City of Austin cracking down on E-bus riders that become unruly on the bus. http://bit.ly/alXE1x . Engaging microblogs are those where a person is inviting a response, such as $7 fa a 24hr bus pass..... how much is it in austin???? exactly..... Analyzing microblogs are those that offer a comparison or analysis of transportation, such as transit: Austin red line vs. Twin Cities Hiawatha line - I like both. Only similarity? single route. Theme relates to the general subject matter of the messages. The themes included such ASMP indicators as economic development, regional integration, land use, and transit system.
6

The following terms were used as keywords: capital metro, cap metro, capmetro, bus, rail, dillo, transit, train, streetcar, road, street, car, motorcycle, school or, traffic, congestion, highway, freeway, bicycle, bike, cycle, lane, path, sidewalk, walk, pedestrian, bicyclist, and crosswalk. Additional not terms were included to exclude irrelev ant tweets, including Austin_crime, 4sq.com, #job, gowal.la It should be noted that bond and Austin Strategic Mobility Plan or ASMP were not included as search terms. Not including these search terms was an oversight on the part of SNAPPs administrator. This may have led to missing some microblogs that discuss relevant topics without including other transportation-related terms.

There are two topic classifications for each microblog since, in many cases, each includes multiple topics. For example, RT @fitcityleblanc: Where do most of Austin's bicycle -motorist collisions occur? Check this map! http://bit.ly/52X2g. In this example, the theme is safety while topic one is bicycle and topic two is car. The selection of topics was based on the common topics discussed by microbloggers. Finally, SNAPP assessed each microblog for either positive or negative sentiment. Sentiment analysis is used in reference to the analysis of evaluative text and tracking of predictive judgments (Pang and Lee, 2008; Tausczik and Pennebaker, 2010; Yu et al., 2008; Tumasjan et al., 2010). SNAPPs analysis exclusively included examination for positive or negative polarity. In some cases, interpretation was necessary where sarcasm is detected. For example, in stuck again, gotta love this Austin traffic the microblogger used love sarcastically to convey a negative sentiment. Caution was taken where ambiguity existed, and these posts were coded neither as positive nor negative. Approximately 45 percent, or 4,203 of the general microblogs, expressed sentiment.

Sentiment Analysis For a more detailed sentiment analysis, the researchers objectively and systematically analyzing the general microblogs to assess emotional, cognitive, and structural components of the text. An automated sentiment analysis was undertaken to evaluate the general microblogs. A text analysis software, Linguistic Inquiry and Word Count (LIWC, 2007), assessed emotional, cognitive, and structural components of the tweets using a psychometrically validated internal dictionary. LIWC calculates the degree to which a text sample contains words belonging to empirically defined psychological and structural categories. The software determines the rate at which certain cognitions and emotions are present in the text. It calculates the relative frequency with which words related to the psychological dimension occur. For example, the software looks for words such as annoyed or hate along with 182 other words that characterize anger. LIWC has been used in a variety of fields, including political science, psychology, and linguistics (Yu et al., 2008). LIWC analyses have been used to examine short

texts, such as instant messages and tweets (Tausczik and Pennebaker, 2010; Tumasjan et al., 2010). The analysis considered multiple (such as future/past/present orientation, sadness, anxiety, anger, tentativeness, certainty, work, achievement, and money). In prior studies as best suited to profile messages covering both emotions (e.g., anger over being stuck in traffic) and content (e.g., how one gets to work). Using Yu et al.s (2008) methodology, all of the microblogs were aggregated into a single text sample for evaluation by LIWC. The aggregation occurred in two topics: microblogs related to mode of travel and microblogs related to the bond election and the ASMP.

Analysis of Micro-Participation Dialogues The researchers also examined the micro-participation dialogues. A microblog can be limited to a simple statement that shares information. For example, #xPD Austin TX police officer suspended 1 day after rear-ending car stopped at light while reading her dashboard comp... http://is.gd/drROf. This microblog shares information about a police officers traffic accident. In moving to micro-participation, it is important that microbloggers exchange ideas and enter into dialogue. Exchange of ideas on Twitter is in part achieved through the use of the @ sign. Honeycutt and Herring (2009) find that the vast majority of @ signs are used to direct a tweet to a specific addressee. A review of the 8,308 relevant tweets was undertaken to determine the portion that include an @ sign, RT (retweet), and/or a URL. These markers are indicators of micro-participation, which can be conversations between a microblogger and others in his or her social network or a conversation with a particular microblogger in the broader community. The markers help to understand the intent of the microblogger who is engaging in micro-participation. Measuring micro-participation between two individuals is particularly difficult. In microparticipation between SNAPP and microbloggers, the facilitation only picked up once SNAPPatx had responded because there had not been a #snappatx tag in the microbloggers initial communication. When SNAPP facilitators engaged in a conversation, they used the #snappatx tag to allow for easy tracking of conversations. It is important to note that as conversations

extend, it becomes more difficult to fit all of the #, @, and RT tags into the ongoing conversation, so microbloggers will chop off all or a portion of these tags.7 For analysis purposes, this practice it makes it very difficult to track conversations, and there is a known undercounting of facilitated conversations in this study. This analysis uses the 1,007 microblogs that contain the SNAPP side of facilitation messages and then searches for the missing pieces of the conversation. For example, one of the SNAPP microblog messages contained @calebsimpson So... how did the first-time bike commute go this morning? #snappatx From that point, the larger microblog database was searched to identify the microblogger and the relevant date to determine if there were other relevant microblogs that can be matched to the conversation. In this case, the whole conversation follows: Calebsimpson: Good morning world. Planning on bike commuting in Austin for the firs t time, today. SNAPPatx: So... how did the first-time bike commute go this morning? #snappatx Calebsimpson: It went great, flew down S. 1st street without a problem! RT @SNAPPatx: So... how did the first-time bike commute go this morning? #snappatx All conversations occurring within one week are counted as completed dialogues. The researchers identified 157 completed dialogues and examined each for its content and the way that SNAPP engaged the microblogger and guided the conversation. The extent to which conversations were continued with others is also tracked through either a mention or a retweet. For example, a portion of a conversation in another microparticipation dialogue between Calebsimpson and SNAPP was then retweeted by AustinCyclist: RT @calebsimpson: Biking the Austin Hill Country, my first experience (since moving here) and thoughts on the bike community - http://bit.ly/905TF3 This counts as an extension of the conversation by a third party.

Twitter recently added a feature that preserves a reference to the original tweet. During this project period, it was not possible to make the changes necessary to take advantage of this new Twitters capability. Twitter has also made a change so that messages no longer show RT in the message, although if you respond logged on from your web account (as opposed to via smartphone) it tracks conversation threads. This will help with some of the challenges mentioned, but is also makes straightforward searches on RT less complete.

SNAPP attempted to engage with specific microblogs, but there were instances in which there were no responses (or none that could be found because of possible stripping of tags). 8 For example, responding to the microblog Austin traffic. never fails, SNAPPatx poste d, @_bnicole08 ATX traffic never fails to what? Disappoint? Occur? Frustrate? Give you quality alone time? Where'd you get stuck? #snappatx. However, the microblogger did not provide a further response. SNAPP had a total of 217 attempted but not completed engagements. SNAPPatx also attempted to stimulate discussion by posting microblogs about ASMP topics. These microblogs were could be seen by followers/fans and by those who visited SNAPPs website. The degree to which people responded to the questions indicates the success/fail rate of facilitation attempts by SNAPPatx. SNAPP would post microblogs to stimulate discussion on both Twitter and Facebook. Given the difficulty of connecting microblogger posts to SNAPP engagement attempts, this piece of the analysis was restricted to the Facebook SNAPPatx fan page. This analysis examines the frequency and type of responses received. For example, SNAPP posted: SNAPP: Texas get's a "starter" grant to plan for High-Speed Rail. With Central Texas as a key hub for H-S Rail, will the many entities with an interest (city transportation departments, counties, CAMPO, Lone Star Rail, Austin District of TxDOT, etc) build a collaborative plan and partnerships? What do you think? Microblogger 1: High hopes, low expectations. On this project we not only have to fight the SUV drivers, but probably Southwest Airlines, too. Microblogger 2: Why to McAllen instead of the Houston, SATX, Austin triangle? Is this the dreaded Trans Texas Corridor raising it's head again? While this analysis excludes engagement using Twitter around the stimulation questions, it does provide an understanding of the degree to which the use of stimulating questions led to engagement among fans of SNAPP.

Equity of Microblogger Participation


8

If the microblogger initiated the conversation by including the #snappatx tag and then SNAPPatx responded, the conversation was counted as complete. A conversation was also considered complete if SNAPPatx initiated the conversation and the microblogger responded.

In the off-line world of participation, there are the planning regulars, citizens who regularly attend public meetings and participate. Within meetings, there are those who sit quietly and others who dominate the discussion. Is the same thing happening online, where only a small group of people tweet about transportation in Austin? We compared the total number of microblogs in the period analyzed. Equality of discussion within microblogs would be exhibited if there was relative equality in the number of posts among microbloggers. By contrast, inequality of participation would exist if just a few microbloggers contributed a large number of posts, effectively dominating the discussions taking place with SNAPP. To determine equality of participation, an examination of the individual microbloggers and the number of relevant microblogs contributed was undertaken. The microbloggers were categorized as onetime contributors, light contributors (2-9 microblogs), medium contributors (10-19 microblogs), heavy contributors (20-49 microblogs), or very heavy contributors (more than 50 microblogs), both on a monthly and an entire-period basis. This allowed the researchers to determine whether particular microbloggers were dominating discussions in either a given month or over the entire seven-month period. Following SNAPPs principal micro-participation process, SNAPP invited every user who had participated or whose microblog was captured to participate in a survey about their experience. This survey, which attempted to capture some demographic data about a sample of the micro-participation, asked for the age and gender of participants. This study recognizes that we know little about microbloggers aside from their avatars and any other information that they wish to share about themselves (Nayar, 2004). It also recognizes that the microbloggers may or may not be citizens of Austin or part of the electorate. What is of principal interest to this study is the content of the microblogs and the extension into micro-participation, not the person writing the content. This is why the analysis of the microbloggers is limited to the equality of participation.

Influence on Decision-Making In order to understand the degree to which micro-participation influenced decision making, SNAPP contracted with a consultant to undertake independent pre-project and post-

project interviews with key stakeholders. Interviews were conducted with nine people, including the Citys Communications Director, Assistant Director of Transportation, Mayor Pro Tem, two council members, a transportation planner with the Metropolitan Planning Organization, a leading blogger in Austin, and two consultants for the ASMP. The pre-project interviews took place in March of 2010. The participants responded to a series of questions about their experience levels with social media, how effective social media is for pushing information out, how effective social media is for engaging people, what optimal participation would be, and how participation in social media should be measured. The post-project interviews were undertaken between December of 2010 and January of 2011. The interviewees were asked five questions: How well did SNAPP help in pushing information out to the public? How well do you think SNAPP engaged the public in decision making? How did SNAPP perform versus your [previously stated] ideal of public engagement? How do you think SNAPP helped with decision making? What do you think was frustrating about making decisions with information from SNAPP? Beyond the pre- and post-project interviews, the researchers interviewed the SNAPP administrator and a staff member in the Citys Transportation Department. This allowed the researchers to better understand the decision-making process beyond the questions asked in the pre- and post-project interviews. The central question from the researchers: why wasnt SNAPP more influential in the decision-making process?

Results The results are explained below, starting with an analysis of micro-participation, followed by a discussion of equality in participation and then influence on decision-making.

Analysis of Micro-participation How can micro-participation be analyzed to create understanding of the publics views on transportation topics? To start, it is important to understand where the microblogs come from. Of the 8,308 relevant microblogs, the vast majority (99.7 percent) came from tweets. Facebook represented 0.1 percent of the microblogs and comments on blogs represented only

0.2 percent. This result can be attributed to the method by which microblogs were obtained. Due to privacy settings, only microblogs posted on SNAPPatxs Facebook page could be collected. As described in the methodology section above, each microblog was examined to determine its content. Despite their brevity, the microblogs express substantive issues that are relevant to the ASMP. SNAPP searched for microbloggers already writing about topics of interest to the City as part of the ASMP. The broad nature of the topics that relate to the ASMP, such as traffic and bicycling, meant that the public could be microblogging about these topics from many different perspectives and in ways that might not be covered by the city or the media. For example, one of the microblogging topics was sightings of a local cyclist who wears only a thong while riding his bicycle. Is microblogging about the thong bicyclist relevant, and does this focus on substantive issues related to the ASMP? The public discussed 173 different topics, from accidents to Willie Nelson Boulevard. Figure 1 includes the ten most discussed topics. These topics indicate that the public was routinely discussing topics that are directly relevant to the ASMP, however discussion was skewed heavily dominated by traffic. [Insert Figure 1 about here] Sentiment Analysis When microbloggers discuss transportation topics, in what sort of light were they raised? Microbloggers discussed a variety of transportation topics, and the sentiment varied from topic to topic. To explore sentiment within topic, we aggregated the tweets about each topic area and then mapped the degree to which sentiments varied as a whole by topic. Figure 2 illustrates the frequency of positive/negative sentiment by topic. [Insert Figure 2 about here] In order to measure sentiment based on mode of transportation, we used the relative frequencies of LIWC category word counts related to multiple sentiment dimensions. These sentiment profiles represent the aggregated linguistic profile of messages associated with each mode, such as bus. Figure 3 illustrates the sentiments which are largely similar across the different modes, with some exceptions. Bicycles were more often mentioned in association

with leisure than any of the other modes. Pedestrians were more often mentioned as related to space. [Insert Figure 3 about here] [Insert Figure 4 about here] This project was ultimately about the mobility plan and the bond election. There was minimal emotion associated with the bond election and some level of positive motion. When people tweeted about the mobility plan, they used words that described the plan as an achievement.

Micro-Participation Dialogues What about the content of the microblogs? Approximately 33 percent of the microblogs shared information, while 38 percent engaged and 20 percent analyzed. Micro-participation is achieved through dialogue, which can be achieved in part through the use of the @, RT, and URL. These can represent conversations between a microblogger and others in his or her social network or with particular microbloggers in the boarder community. Approximately 57 percent of microblogs in this study contain an @ sign. This indicates that people are not just using microblogs to share their opinions, but also engage in interactive discussions. This study found a high degree of retweeting, 24 percent among the general microblogs, demonstrating that the microbloggers were retweeting information about transportation in Austin. In part, SNAPPs role was to push information out that could be shared with others. The information provided on social networks by SNAPP was retweeted 79 times. These retweets contain information that the sender finds worthy of noting, such as links to other websites. A total of 60 percent of messages contain a link to a website. These results indicate that people are finding transportation information that they then share with their network of followers. While 57 percent of the microblogs in this study use @ to indicate an effort to engage in dialogue, it is difficult to document conversations. Of the 374 documentable SNAPP attempts to engage in dialogue, 42 percent received a response from the microblogger. This is a high rate of participation when considering that SNAPP was effectively cyber-stalking, attempting to drop in on conversations that did not include SNAPP. The 157 completed conversations between

SNAPP and microbloggers that could be documented for analysis were primarily brief but specific. Each conversation was unique and highly contextual, limiting the ability to generalize. If a person complained about traffic, the tweets quickly moved to the precise location where the person often experiences traffic. In the following brief example, SNAPP engaged with a bus rider who had problems with frequency of bus service. Microblogger: Seriously the bus system in Austin needs major work. SNAPP: @Katshead42 What about Austin's bus system isn't working for you now? How could it be made better? #snappatx Microblogger: @SNAPPatx my bus was 15 minutes early so i had to wait at the stop for an hour for the next one to come. They drive by stops all the time SNAPP: @Katshead42 It sounds like more frequent buses might help ease the pain if u miss a bus that's running early, yes? Anything else? #snappatx Microblogger: @SNAPPatx that would help. If the buses ran later into the night or early morning that would help too. SNAPP: @Katshead42 Excellent. Voicing your opinion about problems/solutions is the best way to make change happen. We hear you! #snappatx In longer conversations, the efficiency of dialogue is impressive. One of the goals of SNAPP was to educate the public about the ASMP. In the exchange below, the microblogger begins with an analysis that compares Nashville and Austin. SNAPP is then able to engage by providing a specific picture of a downtown transit station. The microblogger then shares another picture of a station in Oregon, allowing the conversation to extend into viewing links to pictures. SNAPP then shifts the dialogue to a discussion of Austins proposal for urban rail. The microblogger then asks a question to get more clarification about the urban rail proposal. SNAPP then points the microblogger to the details of the urban rail proposal. In this example, its clear that there are pieces of the dialogue that have been lost. That said, it is also cl ear that SNAPP was successful in engaging this microblogger, who expressed an initial interest in bus rapid transit, to educate him or her about the urban rail proposal. Microblogger: So why has no 1 talked abt Nashville BRT? Yes it's BRT-lite, but it means that Nashville > Austin bus service! http://bit.ly/bK6H9o #transit

SNAPP: Okay #Austin, #Nashville has you beat again. An amazing downtown transit station. http://bit.ly/9YDjGU #snappatx Microblogger: @SNAPPatx really impressive station. Like that waiting room. Only downtown transit station I've been to: Eugene, OR. http://flic.kr/p/7gB3cj SNAPP: @btx91 ATX is looking at a combo of BRT and streetcar for the 2012 urban rail project but no mention of a great station like TN. #snappatx Microblogger: @SNAPPatx yeah, Congress Ave. acts as a transit mall of sorts though. Is this BRT the MetroRapid or something diff? #snappatx Microblogger: @SNAPPatx and latest on urban rail? Is the streetcar going to be at-grade, mixed-traffic like Portland, or with some kind of separation? Microblogger: @SNAPPatx Tacoma's LRT is an example of what I mean by slight grade separation #snappatx SNAPP: @btx91 Check out all the deets on the ATX urban rail project here www.austinstrategicmobility.com/resources/urban-rail-project #snappatx SNAPP: @btx91 We admit, it looks and sounds good. Have you been there to experience it in person? If so, how was it? #snappatx In the following dialogue, SNAPP was able to educate and receive input on potential solutions. The microblogger starts by telling a fellow microblogger his or her thoughts about Austin, noting that it is car dependent and lacks light rail. SNAPP then provides information that urban rail is part of the bond election and asks how the microbloggers would improve mobility. While there is a sense of humor about getting rid of the Interstate highway, it is also clear that land use is a critical issue. From this dialogue, we learn about a residents commuting patterns and his or her ideas for the future of transportation. @gary_hustwit Austin. Good: nice public outdoor spaces. Bad: Very car dependent, no urban light rail. #Urbanized @compactrobot Urban rail is an item on the 2012 transport bond so keep an eye out. How else would you improve Austin mobility? #snappatx @SNAPPatx reduce the need for mobility to begin with. More VMU. Lessen the grip of NAs. @SNAPPatx oh yeah, also nuke I-35 from space.

@compactrobot Well, that might create a different sort of traffic jam... Where are your worst I-35 trouble spots? #snappatx @SNAPPatx I avoid it, frankly. I just don't like the way it's sliced downtown in half and isolated the east side from the city. @SNAPPatx it's great for trucking companies and horrible for Austin residents. and it's a giant eyesore. @compactrobot All fair points. Do you successfully take local routes to avoid I-35? Do you feel similar ire toward Mopac too? #snappatx @SNAPPatx I only take 35 if I'm eating on the east side, & only after rush hour. otherwise I'll use airport, Lamar, or Guadalupe & cut over @SNAPPatx Mopac's not as bad. but then I don't have to use it to daily to go to/from work. The conversations are professional, but they also find ways to connect with microbloggers and encourage participation. This following very lengthy dialogue and only the relevant excerpts are included. This dialogue demonstrates that microbloggers are skilled at sending posts in parts to complete an idea. In this case, after some initial conversation, the microblogger turns to a discussion of a problem with a specific road and then goes on to describe the specific traffic problem and their specific solution to the problem. SNAPPs responses then lead to comments about a transportation improvement that the microblogger appreciates. Microblogger: If the westbound right lane of Riverside Drive were a dedicated right turn lane at south 1st street it would solve some problems. #ATX SNAPP: @Jordan_Stewart Can you describe the problem that a dedicated right turn lane there would solve? #snappatx Microblogger: @SNAPPatx #SNAPPatx From my observation, WB Riverside traffic proceeding straight fills up the right lane due to the upcoming roundabout... Microblogger: @SNAPPatx part 2 Right-turn traffic also jams into that lane because it has to. Often I've seen RL backed up thru Barton Springs...#snappatx Microblogger: @SNAPPatx Part 3 while the LL has maybe 3 or 4 cars (who have to merge after the light anyway to proceed past the roundabout). #snappatx

Microblogger: @SNAPPatx #snapp re-striping lanes: place thru traffic in underused LL, freeing up lane for turning traffic, ending chronic backup thru B.S. Microblogger: @SNAPPatx further helping: With dedicated RT lane and RT signal, cars could turn freely while LT arrow activated for SB S. 1st st. #snappatx SNAPP: @Jordan_Stewart Nice description Jordan! We hear you & thanks 4 sharing. Keep advocating for that change & hopefully it will come. #snappatx Microblogger: @SNAPPatx Good to hear. It took a few tweet so I'm glad it got through. I'm a full-time Austin resident again so I'll be involved. #snappatx SNAPP: @Jordan_Stewart Well welcome back to ATX full time! We look forward to hearing more of your thoughts on ATX transit and mobility. #snappatx Microblogger: @snappatx Really like the new bike lanes on S. Congress between Oltorf and Ben White. Buffer between bike and car lanes is smart. #snappatx SNAPP: @Jordan_Stewart Separated bike lanes do seem safer and more effective for both cars and bikes. Let's hope time proves that out! #snappatx SNAPP was able to generate a significant following on Twitter and Facebook. It attempted to stimulate discussion by posting microblogs about ASMP topics. SNAPP posted a total of 282 stimulating microblogs on Facebook. SNAPP received responses to 55 percent of its microblogs on Facebook. Where there was a response, on average two microbloggers would post a message. The stimulating microblog that generated the most responses, 20, stated Notice the Red Line traveling around today? It will continue until midnight. Why? Cap Metro is testing whether to add midday and Friday evening service to the red line. Would you ride the rail during the proposed expanded service hours? The responses ranged all over with different microbloggers commenting on timing, days, routes, and other attributes of whether service should be expanded. These examples illustrate SNAPPs ability to engage the public in specific and detailed dialogue around specific transportation issues. SNAPP was able to direct the conversation to issues specific to the ASMP and gain feedback from the public that could then be conveyed back to City officials.

Equality of Microblogger Participation The total number of microblogs per microblogger was measured to determine the equality of participation. Table 3 illustrates the rate of participation. These results indicate that there were a wide variety of participants. In part this may be because SNAPP was an observer of and engager in conversations already happening. When examining the rate of messaging, it is apparent that heavy users account for a very small portion of the messaging. The heaviest user is Austin_Crime, a microblogger who reports on crime issues. SNAPP picked up tweets related to transportation-related accidents or other transportation-related content. Most came from one-time users (57 percent). The onetime and light users accounted for more than 75 percent of the messages. Given the number of tweets received and the number of microbloggers, on average there would be expected to be 1.3 microblogs per microbloggers. The chi-square is 16,0004.636 with four degrees of freedom and a p-value of 0. [Insert Table 3 about here] The post-participation survey of participants indicated that 58 percent were male and that the largest segment of respondents (57 percent) were between the ages of 23 and 44. The participants are daily users of social networking and primarily use these tools for social activities, such as networking, keeping in touch, and socializing. Two-thirds of the respondents indicated that they use social networking when communicating with friends and 9 percent indicated that they use it when communicating with colleagues. The demographics of microbloggers may not be representative of either the Austin citizenry or electorate. While statistics on users in Austin are not available, one can reasonably assume that Twitter and other microblog users are not the majority of the population.

Influence on Decision-Making SNAPPs end goal was to use the meaningful input obtained through micro-participation to influence the decision-making project. Despite the pervasive discussions introduced by SNAPP to the Austin Strategic Mobility Plan public involvement process, City staff from the Austin Transportation Department report that SNAPP input was not formally presented to the

Austin City Council as the transportation bond language and project scopes were being developed (K. Villalon, personal communication, January 5, 2011). However, the projects results were communicated individually to some members of City Council by the SNAPP administrator. Villalon explained that the successful language for provoking on-line discussions was not always compatible with the Citys effort to remain neutral, which created a challenge in working with SNAPP. As one City official put it one of the overarching issues SNAPP faced was the temptation to lead into advocacy, which we cannot do with taxpayer money (K. Villalon, personal communication, February 1, 2011). For example, an October 1, 2010 tweet by SNAPP read, possible solution to air pollution: Renewable energy. So why hasn't it passed through Congress? http://ow.ly/2N7HH #snappatx. This fact may have been effective in inciting comments, but it may not have been perceived as completely neutral. The City staff pushed back because they felt the surveys SNAPP developed were leading or too short to provide an accurate picture or choice. Staff felt that this was a short coming of the micro-survey approach and as a result the survey results had little influence on the Citys decision making process. Although the project staff were advised by the projects advisory committee to strive to retain neutrality, the sheer volume of microblogs being pushed out by SNAPP to incite feedback became difficult for the City and the advisory committee to monitor closely. This led to concern about the degree to which the public comments should be considered. One City staff member described the results as anonoymous and was little more than a finger in the wind (K. Villalon, personal communication, February 1, 2011). Prior to the launch of SNAPP, interviewees had an expectation that optimal public involvement would include broad involvement. From the pre- and post-project interviews, there was a high expectation of the ability of social networking to inform the public. However, there was significant skepticism about SNAPPs ability to engage the public in decision making. The interviewees were impressed with how much SNAPP helped to engage the public. They were particularly impressed with the proportion of comments in the engaging and analyzing categories. The respondents also appreciated the sentiment graphs. While impressed, decision makers pointed to the need for more metrics that would help them validate the public input and understand the stories behind the data.

Respondents indicated that they were frustrated that the reports provided lacked specific policy direction and did not address authenticity. They wanted to have direction on specific policy decisions, but also recognized that it would be unrealistic to garner detailed commentary on policy options. For example, what are people specifically saying about Car2Go that would allow leaders to make a decision? The issue of who participates, and who they represent, was central to some respondents. Respondents indicated that greater visibility of participant information might alleviate these concerns in the future.

Discussion Ultimately, the voters of Austin approved the transportation bond, with 55 percent of voters supporting it. SNAPP used micro-participation to educate and engage the public around the ASMP that led to the bond election. The highly experimental nature of the project meant that there was a learning curve for both SNAPP and the City of Austin in learning how to use the results of public engagement effectively. The results indicate that information on microblogs can be aggregated in a meaningful way, that SNAPP was able to achieve more equitable participation, and that micro-participation was not effective in supporting decision-making in this case. The sentiment analysis demonstrated that it is possible to aggregate microblogs to create meaning. City officials reported that having sentiment analysis was particularly helpful in understanding the perspectives of the public. Furthermore, previous research suggests that the observed tweet sentiment are representative of the sentiments of the broader public (Suenami & Yutaka, 2010; Tweetminster, 2010; Turmasjan et al., 2010). Our research is based on the LIWC text analysis software, which is not specifically tailored to classify short tweets, common acronyms, and emoticons used. This means that the analysis may not have fully represented the true sentiment expressed in the microblogs. Future research could develop an analysis tool specifically designed for microblog analysis. SNAPP was effective in generating engagement. SNAPP was able to generate 203 fans, seven times the average number found in previous research of Facebook in planning projects (Evans-Cowley, 2010). This was in part achieved through a strategy of friending organizations

and then encouraging them to invite their members to join SNAPP. However, finding and engaging each individual through Facebooks closed network required substantial effort. SNAPP generated 366 followers on Twitter. SNAPP exceeded the number of users of 98 percent of Twitter users, all of which have fewer than 300 followers (Cheng et al, 2009). This success probably results from SNAPPs ability to directly engage with Twitter users. Retwee ting is most effective when concerted efforts to limit tweets to a single topic (Cha et al., 2010), as was the strategy of SNAPP. SNAPPs efforts resulted in an average of 45 microblogs retweeted per week. Based on research that showed that a retweet reaches an average of 1,000 users, SNAPP was potentially reaching 45,000 people per week (Kwak et al., 2009). Some fans/followers were media outlets, such as the Austin Chronicle newspaper, Fox7 television, the University of Texas public radio. These connections allowed SNAPP to reach traditional news media, having the potential to extend the reach of the communication without requiring press releases and potentially attracting fans/followers. While the diffusion of microblogs is important, what is the informational value of what is being shared? SNAPP found that only 23 percent of the microblogs were relevant, similar to another study on the value of tweets (Pearanalytics, 2009, Naaman et al., 2010). The City was particularly interested in microblogs that were analyzing or engaging, which represented 57.5 percent of the relevant microblogs. The City of Austin was concerned that some people would not comment and others would dominate the on-line discussion. We found that when it comes to equality SNAPP achieved better results than found in other studies, with most contributors contributing one time (Jansen and Koop, 2005 and Tumasjan et al., 2010). This may in part be attributed to SNAPPs direct engagement with the public, rather than relying on the public to co me to the website or blog to provide their comments. That said, the real-world identity of microbloggers participating with SNAPP is unknown, and it is impossible to determine the extent to which the microbloggers are new participants in the planning process or members of the public who have participated in past public meetings or other engagement venues. While public officials might want to know who wrote a statement, it isnt really necessary. The microbloggers are people who have experience in Austins transportation system. They are providing their opinions on the current and future of transportation in the city. Why would this input be any less valuable

than input received from a person with a detailed identity? Further study should examine the representativeness of participation and the degree to which the online views echo the broader population. One of the challenges for elected officials and transportation staff members was how to use the analysis provided by SNAPP to support decision making. Data coding and analysis is labor intensive resulting in a typical lag of 15 to 30 days to provide monthly analysis. City officials reported that sentiment analysis was a very helpful way to understand what the public was saying about a topic. In the post-project survey, these officials reported that it was difficult to understand the relationships within the data (Ewen, 2010). For example, how could one drill down to understand what is behind sentiment? Continued experimentation will be necessary to find best practices. The promise of micro-participation is that it provides an opportunity to get nearly real-time tracking of public input. Yet, further development of analysis techniques are needed to enable timely reporting that will be most useful to public officials. The City had unrealistic expectations of what the use of microblogs could deliver, in part because the micro-participation was not integrated into a comprehensive engagement plan. The City is accurate in that the users identity is not fully known and what SNAPP captured is a finger in the wind. However, the City was provided with much more information about what people are saying about transportation in Austin than what they were receiving through more traditional forums. The City simply didnt underst and how to effectively use this information and the SNAPP team was unable to deliver the input in a format that City found highly usable. As one City official stated This was an experiment for the City and I think we learned the time and attention needed to support Twitter dialogue, beyond informational tweets, is not currently feasible, even with automated analytics (K. Villalon, personal communication, February 1, 2011). In this case, decision-makers asked for help in understanding the story behind the data. One of the key weaknesses of this project was the inability to translate the results of the engagement into stories that could resonate with decision-makers. This research has experimented with different methods for analyzing the data that could with further refinement led to the ability to influence decision making in future micro-participation efforts.

Further research should empirically investigate the way in which information can be weighted from microblogging sites. Planning is not leading the call for new technologies in social media, but it is riding a wave of rapid change in global and local communication networks. The growth in use of location-aware social media, such as geo-tagged microblogging, has the potential to extend planning participation to citizens, who could digitally tag such planning issues as in this case the location of traffic congestions, areas where bike paths are needed or other transportation related issues. Smartphone technologies have the potential to further democratize planning by allowing participants to join the planning conversation from their regular locations and on their own terms. We contend that micro-participation provide new and valuable opportunities for public participation that should be integrated into a broad-scale participation process. Future research is needed to develop a model for using micro-participation for effective planning engagement.
References American Planning Association (2006). Planning and urban design standards. Wiley graphic standards. John Wiley & Sons, Hoboken, New Jersey. Anon (2009) Twitter Blog: What's happening? http://blog.twitter.com/2009/11/whats-happening.html. Accessed: December 19, 2010. Arnold, M. (2003). Intranets, community, and social capital: The case of Williams Bay. Bulletin of Science, Technology & Society, 23(2), 78-87. Arnstein, S. R. (1969). A Ladder Of Citizen Participation. Journal of the American Institute of Planners, 35(4):216-224. Bailey, K. and Grossardt, T. (2009). Toward Structured Public Involvement: Justice, Geography and Collaborative Geospatial/Geovisual Decision Support Systems. Annals of the Association of American Geographers, 100(1):57-86. Berg, N. and T. Halbur. 2010. Planetizen Picks: Top Twitter Feeds on Urban Planning. Planetizen. Accessed December 10, 2010: http://www.planetizen.com/twitterfeeds Bonaccorsi, A. and C. Rossi (2004) Altruistic Individuals, Selfish Firms?: The Structure of Motivation in Open Source Software, First Monday 9(1), Accessed July 22, 2010 at: http://firstmonday.org/htbin/cgiwrap/bin/ojs/index.php/fm/article/view/1113/1033 Bordieu, P. and L. Wacquant. (1992). An Invitation to Reflexive Sociology. University of Chicago Press.

Boyd, D., (2007), Social Network Sites: Public, Private, or What? Knowledge Tree, n13. Accessed September 14, 2009: http://kt.flexiblelearning.net.au/tkt2007/wpcontent/uploads/2007/04/boyd.pdf Boyd, D. (2004, April 24-29, 2004). Friendster and publicly articulated social networks. Paper presented at the Conference on Human Factors and Computing Systems (CHI 2004). Vienna. Boyd, D. Golder, S, and G. Lotan. 2010. Tweet, Tweet, Retweet: Conversational Aspects of Retweeting on Twitter. Proceedings of HICSS 2010: 1-10. Brody, S. D., Godschalk, D. R., and Burby, R. J. (2003). Mandating Citizen Participation in Plan Making: Six Strategic Planning Choices. Journal of the American Planning Association , 69(3):245-264. Bullard, R. D., Johnson, G. S., and Torres, A. O. (2004). Highway robbery: transportation racism & new routes to equity. South End Press. Burt, R.S.. 2000. The Network Structure of Social Capital. JAI Press, City. Capital Area Metropolitan Planning Organization, Group Solutions RJW, and Estilo Communications (2009). People, Planning, and Preparing for the Future: Round 2 Public Involvement. Technical report, Capital Area Metropolitan Planning Organization, Austin, TX. Carr, D. 2010. Why Twitter will endure. The New York Times. January 1. Accessed December 19, 2010: http://mcsd-blog.mcbarons.manheimcentral.org/zimmermank/files/2010/01/why-twitter-willendure1.pdf City of Austin (2009). Making Austin: Public Participation in a New Comprehensive Plan. Technical report, City of Austin, Austin, TX. Cha, M., Haddadi, H., Benevenuto, F. and K.P. Gummadi. 2010. Measuring User Influence in Twitter: The Million Follower Fallacy. In Proceedings of the 4th International Conference on Weblogs and Social Media, 2010. Accessed December 19, 2010: http://www.aaai.org/ocs/index.php/ICWSM/ICWSM10/paper/download/1538/1826 Cheng, A., Evans, M. and H. Singh. 2009. Inside Twitter: An In-Depth Look Inside the Twitter World. Accessed December 19, 2010: http://files.altaviaspindigital.com/tools/etudes/reseaux_sociaux/Inside_Twitter_BySysomos.pdf Day, P. (2002). Designing democratic community networks: Involving communities through civil participation. In M. Tanabe, P. van den Besselaar, & T. Ishida (Eds.), Digital cities II: Second Kyoto workshop on digital cities, LNCS Vol. 2362 (pp. 86-100). Heidelberg, Germany: Springer. Day, P., & Schuler, D. (Eds.). (2004). Community practice in the network society: Local action/global interaction. London: Routledge. Donath, J., & Boyd, D. (2004). Public displays of connection. BT Technology Journal, 22(4), 71.

Ellison, Nicole, Steinfield, Charles, and Cliff Lampe. (2006). Spatially Bounded Online Social Networks and Social Capital: The Role of Facebook. Paper presented at the Annual Conference of the International Communication Association, June 2006. Accessed January 29, 2009: https://www.msu.edu/~nellison/Facebook_ICA_2006.pdf Evans-Cowley, Jennifer S. (2010). Planning in the Age of Facebook: The Use of Social Networking in Planning Processes. Geojournal. 75(3): 407-420. Evans-Cowley, Jennifer S. and Justin Hollander. (2010). The New Generation of Public Participation: Internet Based Participation Tools. Planning Research and Practice. 25(3): 397-408. Ewen, C. 2010. DRAFT Post-Project Qualitative Interviews- Summary of Work and Results Social Networking and Planning Project. Facebook. (2011). Facebook Statistics. Accessed January 8, 2011 at: http://www.facebook.com/press/info.php?statistics Fallows, D. (2004). The Internet and daily life. Washington, DC: Pew Internet & American Life Project. Forester, J. (2006). Making Participation Work When Interests Conflict: Moving from Facilitating Dialogue and Moderating Debate to Mediating Negotiations. Journal of the American Planning Association, 72(4):447-456. Foth, M., & Brereton, M. (2004, November 20-24). Enabling local interaction and personalised networking in residential communities through action research and participatory design. In P. Hyland, & L. Vrazalic (Eds.), Proceedings of OZCHI 2004: Supporting Community Interaction. Wollongong, NSW: University of Wollongong. Ghosh, R. A. (1998) FM Interview with Linus Torvalds: What Motivates Free Software Developers?, First Monday 3(3): http://firstmonday.org/htbin/cgiwrap/bin/ojs/index.php/fm/article/view/1475/1390 Gilchrist, A. (2004). The well-connected community: A networking approach to community development. Bristol, UK: Policy Press. Goodspeed, R. (2010). The Dilemma of Online Participation: Comprehensive Planning in Austin, Texas. Accessed February 10, 2011: http://web.mit.edu/rgoodspe/www/papers/RGoodspeedAustin_Online_Participation_9-19-10.pdf. Granovetter, M. S. (1982). The strength of weak ties: A network theory revisited. In P. V. Mardsen & N. Lin (Eds.), Social structure and network analysis (pp. 105-130). Thousand Oaks: Sage Publications. Hampton, K. N. (2003). Grieving for a lost network: Collective action in a wired suburb. The Information Society, 19(5), 1-13. Hampton, K. N., & Wellman, B. (2003). Neighboring in netville: How the Internet supports community and social capital in a wired suburb. City and Community, 2(4), 277-311.

Hars, A. and S. Ou (2002) Working for Free?: Motivations for Participating in Open Source Projects, International Journal of Electronic Commerce 6: 25-39. Hertel, G., S. Niedner, and S. Hermann (2003) Motivation of Software Developers in the Open Source Projects: An Internet-Based Survey of Contributors to the Linux Kernel, Research Policy 32(7): 1159-77. Honeycutt, C. and S.C. Herring. 2009. Beyond Microblogging: Conversation and Collaboration via Twitter. Proceedings of the 42nd Hawaii International Conference on Systems Science, 2009. HICSS. Bloomington, IN. January 5-8, 2009. Pages 1-10. Horrigan, J. B. (2001). Cities online: Urban development and the Internet. Washington, DC: Pew Internet & American Life Project. Horrigan, J. B., Rainie, L., & Fox, S. (2001). Online communities: Networks that nurture long-distance relationships and local ties. Washington, DC: Pew Internet & American Life Project. Huberman, B.A., Romero, D.M., and F.Wu. 2008. Social networks that matter: Twitter under themicroscope. First Monday. http://arxiv.org/PS_cache/arxiv/pdf/0812/0812.1045v1.pdf Jansen, HJ, & Koop, R. (2005). Pundits, ideologues, and ranters: The British Columbia election online. Canadian Journal of Communication, 30, 613-632. Jacobs, J. (1992). The Death and Life of Great American Cities. Vintage Books, New York. Java, A., Song, X., Finin, T., and Tseng, B. Why We Twitter: Understanding Microblogging Usage and Communities. Proc WebKDD/SNA-KDD07, 56-65. Jansen, B., Zhang, M., Sobel, K., and A. Chowdury. 2009. Twitter Power: Tweets as Electronic Word of Mouth. Journal of the American Society for Information Science and Technology . 60:2169-2188. Kavanaugh, A. and Patterson, S. (2001). The impact of community computer networks on social capital and community involvement. American Behavior Scientist, 45(3), 496-509. Koop, R. and H. J. Jansen. 2009. Partisan Blogs and Blogrolls in Canada: Forums for Democratic Deliberation? Social Science Computer Review 27(2): 155-173. Kwak, H. Lee, C., Park, H. and S. Moon. 2010. What is Twitter, a Social Network or a News Media? Proceedings of the 19th International Conference on World Wide Web. Accessed December 18, 2010: http://portal.acm.org/citation.cfm?id=1772690.1772751 Lakhani, K.R. and Panetta, J.A. (2007) The Principles of Distributed Innovation, Innovations: Technology, Governance, Globalization 2(3): 97112. Lenhart, A. and S. Fox. 2008. Twitter and status updating. Piew Internet and American Life Project. Accessed December 18, 2010: http://fortysouth.com/wp-content/uploads/2009/05/Twitter-and-statusupdating.pdf

Pennebaker, J., Booth, R. and E. Francis. 2007. Linguistic Inquiry and Word Count [computer software]. LIWC, Inc.. Lietsala, K. and A. Joutsen (2007) Hang-a-Rounds and True Believers: A Case Analysis of the Roles and Motivational Factors of the Star Wreck Fans, in A. Lugmayr, K. Lietsala and Lowry, M. (2010). Online public deliberation for a regional transportation improvement decision. Transportation, 37(1):39-58. Maloney-Krichmar, D., Abras, C., & Preece, J. (2002, June 6-8). Revitalizing an online community.Paper presented at the International Symposium on Technology and Society (ISTAS) Social Implications of Information and Communication Technology, Raleigh, North Carolina. Naaman, M., Boase, J., and Lai, C.H. 2010. Is it Really About Me? Message Content in Social Awareness Streams. Proceedings of the CSCW10, 189-192. Nayer, P.K. 2004. Virtual world. London: Sage. Noveck, B. S. (2009). Wiki Government: How Technology Can Make Government Better, Democracy Stronger, and Citizens More Powerful. Washington DC: Brookings Institution Press. Pang, B., & Lee, L. (2008). Opinion mining and sentiment analysis. Foundations and Trends in Information Retrieval, 2, 1-135. Parks, M.R. and K.Floyd. 1996. Meanings for Closeness and Intimacy in Friendship. Journal of Social and Personal Relationships. 13(1): 85-107. Patterson, S. J., & Kavanaugh, A. L. (2001). Building a sustainable community network: An application of critical mass theory. The Electronic Journal of Communication, 11(2). Pearanalytics (2009). Twitter study. Retrieved December 15, 2009, from http://www.pearanalytics.com/ wp-content/uploads/2009/08/Twitter-Study-August-2009.pdf Pinkett, R. D. (2003). Community technology and community building: Early results from the creating community connections project. The Information Society, 19(5), 365-379. Putnam, R. D. (2000). Bowling Alone. New York: Simon & Schuster. Rheingold, H. (2000). The virtual community: Homesteading on the electronic frontier . Cambridge, MA: The MIT Press. Rheingold, H. (2002). Smart mobs: The next social revolution. Cambridge, MA: Perseus. Smith, A. and L. Rainie. 2010. 8% of Americans use Twitter. Pew Internet and American Life Project. Accessed January 8, 2011 at: http://pewinternet.org/Reports/2010/Twitter-Update-2010/Findings.aspx

Suh, B., Hong, L., Pirolli, P., and Chi, E. H. (2010). Want to be Retweeted? Large Scale Analytics on Factors Impacting Retweet in Twitter Network. In IEEE SocialCom 2010, pages 177-184. Institute of Electrical and Electronics Engineers. Suenami, A., & Yutaka, M. (2010). Voter behavior analysis and election campaign strategies using information on the web (in Japanese). In Proceedings of the JSAI SIG-KBS, No. 87. 16, Tokyo. Tang, K. and Waters, N. (2005). The internet, GIS and public participation in transportation planning. Progress in Planning, 64(1):7-62. Thelwall, M. (2008). Homophily in MySpace. Journal of the American Society for Information Science and Technology. 60(2): 219-231. Tausczik, Y. R., & Pennebaker, J. W. (2010). The psychological meaning of words: LIWC and computerized text analysis methods. Journal of Language and Social Psychology , 29, 24-54. Townsend, A., Maguire, R., Liebhold, M., and Crawford, M. (2010). A Planet of Civic Laboratories: The Future of Cities, Information, and Inclusion. Technical report, Institute for the Future, Palo Alto, CA. Tumasjan, A., Sprenger, T.O., Sandner, P.G., and I.M. Welpe. 2010. Election Forecasts with Twitter: How 140 Characters Reflect the Political Landscape. Social Science Computer Review. 1-17. http://ssc.sagepub.com/content/early/2010/09/24/0894439310386557.full.pdf+html Tweetminster (2010). Is word-of-mouth correlated to General Election results? The results are in. Retrieved June 25, 2010, from http://www.scribd.com/doc/31208748/Tweetminster-Predicts-Findings. Watters, E. (2003). Urban tribes: Are friends the new family? London: Bloomsbury. Wellman, B. (2001). Physical place and cyberplace: The rise of personalized networking. International Journal of Urban and Regional Research, 25(2), 227-252. Wellman, B. (2002). Little boxes, globalization, and networked individualism. In M. Tanabe, P. van den Besselaar, & T. Ishida (Eds.), Digital cities II: Second Kyoto Workshop on Digital Cities, LNCS Vol. 2362 (pp. 10-25). Heidelberg, Germany: Springer. Wellman, B., & Haythornthwaite, C. A. (Eds.). (2002). The Internet in everyday life. Oxford: Blackwell. Wellman, B., Quan-Haase, A., Boase, J., Chen, W., Hampton, K. N., Daz de Isla Gmez, I., et al. (2003). The social affordances of the Internet for networked individualism. Journal of Computer-Mediated Communication, 8(3). Wellman, B., Salaff, J., Dimitrova, D., Garton, L., Gulia, M., and Haythornthwaite, C. (1996). Computer networks as social networks: Collaborative work, telework, and virtual community. Annual Review of Sociology, 22, 213238. Wright, A. 2009. Mining the Web for Feelings, Not Facts. New York Times, 2009-08-23.

Yu, B., Kaufmann, S., & Diermeier, D. (2008). Exploring the characteristics of opinion expressions for political opinion classification. In Proceedings of the 2008 International Conference on Digital Government Research, (pp. 8291). Montreal. Zarella, D. 2009. Science of Retweets. Accessed December 19, 2010: http://danzarrella.com/the-scienceof-retweets-report.html Zhao, D and Rosson, M.B. 2009. How and Why People Twitter: The Role That Micro-blogging Plays in Informal Communication at Work. Proc GROUP09, 243-252. Table 1. Summary of Time Investment and Effectiveness by Technique Participation Technique Public Meeting Survey administered on-line or paper Medium time/ opportunity cost Medium Medium Low Demographic questions allow respondent stratification Facebook Post Twitter Comment

Citizen Investment

High time/opportunity cost Low

Low time/opportunity cost Low Low Medium Difficult to collect comments

Lowest time/ opportunity cost Medium Low Low Non-linear responses can make agency action seem opaque

Effectiveness of Participation Purposes (APA 2006) Notes:

Involving Citizens

Voice in High Planning Sense of Low-medium Community Direct access to decision makers; likely includes multiple agenda items; board chair may re-order items

Table 2. Common Social Networking Terminology URL Hashtag Presence of a website address; Universal Resource Locator Represented by #; used to indicate a common grouping of tweets. For example, #APA2009 might represent the American Planning Association Conference in 2009. Mention Follower Represented by @; the number of usernames specified in a tweet A user who is following the tweets of an author

Followee A friend who the author is following Status A representation of a users current status, including mood, current news, or other information the user wishes to share Retweet Push Represented by RT; The sharing of a tweet with other users A message being pushed out by a social networking user. This is typically news or an announcement intended to be shared.

Table 3. Format of Communication Users User Group One-time (1) Light (2-5) Medium (6-20) Heavy (21-79) Very heavy (80+) Total Total 3,690 684 49 14 2 4,439 Share % 83.1% 15.4% 1.1% 0.3% 0.0% 100% Total 3,690 1,760 465 537 178 6,576 Messages Share 56.1% 25.9% 7.1% 8.2% 2.7% 100%

Note: This includes only general microblogs.

Figure 1. Ten Most Discussed Microblog Topics


Motorcycle/Scooter Planning

Categories of Comment

Other Walking Passenger Rail Cars Bus Bicycling Traffic 0% 5% 10% 15% 20% 25% 30% 35% 40% 45%

Percent of Total (n=8,308)

Chi-Square results: Hypothesis that topics would be discussed equally. Chi-square is 8430.191 with 9 degrees of freedom and p-value of 0. When comparing the two most frequently discussed topics (Traffic and Bicycling) the Chi-square is 967.185 with one degree of freedom and a p-value of 0. These results indicate that traffic clearly dominated the microblogs.

Figure 2. Positive/Negative Sentiment for Topics with five or more microblogs Note when this converts to pdf in the JPER manuscript center I cannot get it to appear properly. The line is 50%. Light gray is positive and dark is negative sentiment.

Drunk Driving Road Rot Rally Train Cap Metro Parking Highway Roads Bus Car Sidewalk Service Rail Transit Electric Car MetroRail Toll Road Light Rail Urban Rail Motorcycle Ped/Walk SNAPPatx Prop 1 Bike Street Names Taxi Bike Lanes Car2Go
0
% Negative

50
% Positive

100

Figure 3. Sentiment Profile by Transportation Mode


Space Present-oriented Time Affect Social Positive emotion Leisure Perception Past-oriented Work Negative Emotion Tentativeness Achievement Anger Certainty Future-oriented Money Sadness Anxiety 0 5 Bicycle 10 Bus 15 Auto 20 Rail 25 30 35 40 45

Pedestrian

Figure 4. Sentiment Profile by Transportation Issue


Mobility Plan Bond Election

Achievement Work

Time

Future-oriented 8 Time Present-oriented 7 6 Past-oriented 5 4 Social 3 2 1 Positive emotion 0

Space Perception

Negative Emotion Anxiety

Certainty
Tentativeness

Anger

You might also like