Online Articles

“Voting in The Modern World: Examining Media Habits to Create a Relevant Guide for the 2016 General Election” by Krystal Lau

Over hundreds of voter guides exist in the United States today; they vary in content style, substance, and medium–including print, podcast, and video (Ballotpedia). The California Secretary of State sends out a Voter Information Guide to every voting household in California. The guide is currently available in ten languages: English, Spanish, Chinese, Hindi, Japanese, Korean, Khmer, Tagalog, Thai, and Vietnamese (California Secretary of State). Similar guides are put together by interest groups, nonparty newspaper organizations, as well as civically engaged individuals (Ballotpedia). However, although voter guides are numerous and varied, their information does not effectively reach a wide proportion of the population. In fact, 60 percent of intermittent voters say they do not know enough about candidates to vote (Pew Research Center).

The dearth and poor accessibility of information about candidates may be contributing to the turnout rates. In the 2014 midterm election, only 36.4 percent of the voting-eligible population cast ballots (Washington Post). Millennial turnout was even lower, at around 21.3 percent for the same year (CIRCLE). As an intern with CALmatters (a nonprofit, nonpartisan media venture that aims to explain California’s politics to voters) I examined the role of the accessibility of voting guides for the November 2016 election.  In order to create a voter guide for the 2016 general election, I sought to closely examine voting behavior and the methods through which voters access and engage with media. This study illustrates the advantages of voter guides, and examine factors that make voter guides an effective, and useful tool for voters. This research also examineselements that different demographics of voters looked for within these guides, Finally, the current study sought to identify variabilities in media habits, and determine how these variabilities influenced civic participation among groups of voters.    

The main goal of the current research is to identify the type of voter guide that maximizes user experience and engagement, and offers the best content style and substance. In order to insure that the findings would be applicable to how individuals behave and act in reality, I conducted user interviews that aimed to determine how individuals access voting information, and why they find certain portals more useful than others. This project utilized a design thinking approach, which applies knowledge about the needs of individuals in order to determine the best way to serve them. Accordingly, one-on-one user interviews were comprehensively examined participants’ preferences and habits. Moreover, a key element of the design thinking approach is a separation between the idea generation stage and the idea evaluation stage. The goal of this separation is to prevent prior bias from influencing the ideation of possible prototypes. In adherence to this approach, a directional hypothesis was not formulated.



Participants in the user interviews were chosen based on their gender, ethnicity, and socioeconomic status. All the participants were civically engaged voters. Many of them were connected to CALmatters as donors or friends of employees. These ten interviewees were comprised of five males and five females, and were of Latino, White, and Asian ethnicity.


This project used a mixed-method research design. The independent variable was type or category of voter guide. This included variations in content (e.g.,whether it is issued by the secretary of state, or by a local newspaper), as well as by medium (e.g., interactive websites, podcasts, or in-person events). User experience, user engagement, and content style and substance of voter guides served as the three dependent variables. Each of these outcome variables is an index.

The user experience index encompassed searchability, scannability, and visual design, rated on a scale of one through five.A score of five indicated a highly functional, usable guide, while a score of one indicated poor, inconvenient, unappealing guide. Searchability was defined as the ease of finding specific information within the guide. Scannability referred to the layout of the guide, and whether it allowed the user to scan the guide and summarize its content. Finally, visual design referred to the appearance and aesthetic of the guide–i.e., whether it was captivating, appealing, and comprehensible.

User engagement, the second index, included sharability, use of images, use of video, and interactive elements. Sharability, was rated on a scale of one through five, with five representing a guide that can easily be shared through multiple outlets (such as Facebook, Twitter, and Email). As the sharability score decreases, sharing options become limited and more difficult to access. A one rating represents a guide that doesn’t offer any sharing options. The other components of the user engagement index were rated on an all-or-nothing scale (i.e., zero to one). For example, a voter guide that contained photos or videos received a rating of one for its use of images or video. Id, and were  Examples of interactive elements included a zip code search bar that identified a voter’s assembly members, a record of user ballot voting preferences, and pop-ups that presented more information about a proposition if a user hovered their cursor over a photo.  

Content style and substance, the third index and dependent measure, included the tone and content of the voter guide. Tone was rated on a scale of one to five, five being a conversational, informal tone, and one being a dry, inaccessible tone. Content was evaluated based on whether the guide included specific analyses of propositions’ purpose. Three types of analyses were examined: an economic impact analysis, a political impact analysis, and a power analysis. All three types of analyses were  Economic impact analysis addressed the fiscal effect of the proposition, if it were to be implemented. Political impact analyses addressed the measure’s potential implications within the sphere of public policy. Finally, power analyses examined the funding behind propositions, and the sources of the funding.  


First, I put together a competitive analysis based on a review of existing voter guides within California. The review rated the voter guides based on the indices stated above. This comprehensive review also included average scores that showed each guide’s overall performance on the different indices.

Afterwards, I interviewed eight individuals who were chosen to participate based on their age, levels of civic engagement, and media habits. Another co-worker in my office also conducted two interviews. Altogether, the project utilized data from ten, one-hour interview sessions. The interviewees belonged to three cohorts; including four young, likely voters, four middle-aged professionals, and two senior-aged political donors, who were highly involved in their communities. Following the design thinking approach, these ten potential voting guide users were asked about their approach to voting and their response to media. There were three categories of questions: personal identity, media habits, and voting behavior. Each interview began with personal identity questions. Participants responded to questions about how long they had lived in California, how long they had been voting, and how they spent the majority of their time. Participants also reported their passions and priorities. Next, participants reported their media habits by walking the interviewer through how they engaged with media on an average day. Participants’ answers were used to identify the types of devices they regularly used to access news, the average amount of time each participant spent reading an article, and the manner in which participants were engaging in media sites (e.g., “What makes you choose to share an article on social media?”) The final set of questions examined participants’ voting behavior. Participants walked the interviewer through the last time they voted, going into detail about the process and referencing on it. Participants who reported that they used a voting guide were asked to identify which voter guides they used, what information they are looking for , and the amount of time they spent going through the guide. ,. Participants also elaborated on why they left parts of the ballot blank, and discussed their incentives to vote.

Forty additional participants were surveyed at the “Millennials and the Future of Work in the Golden State Conference,” sponsored by the UC Berkeley Goldman School of Public Policy. The demographic at this conference matched with one population bloc that this study sought to reach– a new potential audience of civically engaged youth. Survey participants ranged from college students involved in student government to young legislative aides of assembly members.An additional group of participants were surveyed using a panel run on Amazon Mechanical Turk. This online survey contained four questions that offered a broader, more holistic view of voting behavior. Around 320 participants completed the survey. Participants predominantly aged between 26 and 35, and were more civically engaged than the general population.  


Competitive Analysis Results

In the competitive analysis, scores from each index were averaged separately to produce three averages scores for user experience, user engagement, and content style and substance. The three averages were then added to produce an additive sum rating of each of the voter guides (Figure 2: Competitive Analysis of Prevalent Voter Guides).

The voter guides offered by SPUR and Vote Easy received the highest user experience scores. These two sites possess aesthetically appealing, easy-to-use, and highly functional guides. Both guides incorporate eye-catching, well-labeled headings that facilitate navigation. Vote Easy also includes a drop-down menu that allows the user to see where the presidential candidates stand on each issue. Features that focus on improving user experience are especially important for voters who, based on interview responses, take such features into consideration when deciding how and where to seek information about candidates and propositions.

89.3 KPCC and SeePolitical fulfilled every requirement on the user engagement index, and garnered the highest user engagement scores.  Both websites integrate prominent features that allow the user to share content to a wide variety of social media sites. SeePolitical’s guide relied primary on a video format. The site provides short, one-minute animated videos that explain ballot measures. SeePolitical also allows users to save their chosen ballot preferences and  return to them at a later without losing their personalized results. 89.3 KPCC’s voter guide aggregates information from various different sources to produce a well-rounded guide that utilizes text, images, and video. This incorporate to provide, interactive 89.3 KPCC’s guide links to SeePolitical’s videos, supplementing them with original images and articles, 89.3 KPCC’s guide incorporated several interactive elements that promote user engagement, including hyperlinks to relevant websites and polls that surveyed users’ views on ballot measures. 89.3 KPCC  also sources a graphic from MapLight that shows sources of funding for different proposition campaigns. This met the requirements of the power analysis component of the index..

Voter’s Edge (the next generation of Smart Voter) and 89.3 KPCC received the highest content style and substance averages. However, a shortcoming of the content index was the skewed scores within the tone category. Tone was rated on a scale of one to five, where a one rating indicated a dry tone, and a five rating indicated a conversational, informal tone. This made the content style and substance skew towards conversational tones unintentionally. However, the results did not favor conversational voter guides because the guides with conversational tones also tended to have lower economic, political impact, and power analysis scores. That being said, 89.3 KPCC’s high score is mainly a based on the site’s clearly marked headings that partition the proposition articles into sections including background, political analysis, economic analysis, and power analysis. 89.3 KPCC’s articles summarize the significance of the ballot measure, the controversy surrounding it, and the money spent on it. At the end of each article, the website links to other sites that have reported on the same proposition. Once again, this focus on aggregation gives voters access to a wide range of information. Voter’s Edge also utilizes an aggregation approach, taking a less traditional approach to the content. Their articles cover the basics, explaining the measures, and identifying individuals and organizations who endorse the campaign and contribute money to it. However, the website’s guide includes a new, coverage section, in which the website provides links to articles from the LA Times and Capital Public Radio. Voter’s Edge also offers a resources section that presents the pros and cons proposed by nonpartisan organizations such as League of Women Voters, as well as the contact information for advocacy websites. The aggregation approach utilized by 89.3 KPCC and Voter’s Edge proves to be the determining factor for their high content score ratings. Not only does the aggregation approach allow for a higher quantity of information, but it also provides a diverse set of opinions and perspectives. It also makes these websites easier to find by allowing their pages to appear frequently on search engines.  

M-Turk Survey Results

The results of the M-Turk survey indicated that 38 percent of the participants have used a government issued voter guide, 41 percent have not, and 21 percent are unsure. However, a cross tabulation examining responses across different age groups revealed a contrast between young participants and middle-aged participants. While only 28 percent of age 18-25 year olds report having used voter guides, 70 percent of age 36-45 year olds report that they have. Overall, usage of the government issued voter guide appeared to increase with age, peaking at the 36-45 range (Figure 4: Cross Tabulation of Age and Voter Guide Use).    

When asked whom you would trust to give advice on voting, 28 percent of participants indicated that they did not want or need advice on how to vote. 24 percent indicated that they would trust family or friends, and 21 percent of participants reported that they would trust politicians they supported. Advocacy organizations were the least trusted group. This finding aligned with the results from the user interviews, which pointed to a trend towards favoring nonpartisan information provided by third party sources (Figure 5: Who would you most trust to give good advice on how to vote in an election).

Survey Results

The survey results from the Millennials and the Future of Work Conference were inconclusive. The results showed that users were open to trying a wide variety of products, as long as they did not take up too much time (Figure 6: Millennials Conference Survey Results). Overall, the results saw a wide range of preferences, and the only definite trend was a general opinion against audio formats (Figure 7: Millennials Conference Survey Results Continued).

User Interview Results

The findings of the user interview were divided into three sections: who, what, and how. “Who” represents who the users are, as in what their habits are, what preferences they possess, and how they form their identity. “What” stands for the voter guide itself, in what the users are looking for in a voter guide. “How” shows findings related to how the users access the guides– their portals, websites, and other means of media.

In the “Who” section, we saw a theme of the “Ivory Tower.” In the interviews conducted with participants involved in academia, patterns were found in which individuals who discuss, study, and research politics, are not always inclined to participate in it themselves. This led to a second theme, which was especially strong in the females: social shaming. In many instances, voting is seen in relation to one’s image, in that it is an act used to boost one’s public image as a civically engaged citizen. In circles of academia, to admit one does not vote is a shame-inducing subject. This also corresponds to the social aspect of voting. Voting as a social activity was a thread we saw not only in the female participants, but also in the male participants. The pleasure of voting is mainly derived from this social aspect, in that individuals can discuss their preferred candidates and debate over controversial measures with their friends and family. The participants in the older demographic also mentioned participating in election parties in which friends gather to learn about the measures and carry out their research on the upcoming ballot together. In contrast, a theme we saw only in men was the private nature of voting. Certain male participants mentioned their disinclination to share their political preferences with others; they seek to research it on their own. This correlates with another theme we saw predominantly in the males– the theme of independence. Certain males possess a “do it yourself” mentality in that they choose not to rely on voter guides because they believe that “voting is easy” and that they can find the information on their own. They are also disinclined to rely on only one source, as the possibility of bias can enter in such a situation. One male participant stressed his reliance on multiple outlets to make sure he was getting a clear, well-rounded perspective on the measure. Later on, he would go on to share how he is looking for a guide that can aggregate information from different organizations onto one site, thereby simplifying his search process. We saw that brands matter for individuals, regardless of ideology, gender, or race. Newspapers such as the New York Times and the Washington Post have built up trust among our participants, and their long-term relationships with the participants have caused the users to understand the journalists’ ideological leanings and gauge the articles’ biases. Certain identities that we saw as especially significant in the voting booth were ideological identities, Latino identities, and female identities. In addition, we found that most participants set aside around half an hour for researching the ballot measures, yet enter into their research time with preconceived opinions on the majority of the ballot. Finally, several of the participants saw the act of voting as setting an example, whether for their children, their friends, or their community at large.

For the “What” section, we saw that individuals were divided between wanting unbiased, independent third party perspectives, and wanting a voter guide that provided a clear ideological alignment. For the individuals who voted by party line, they showed tendencies to read only articles that aligned with their political ideologies. Almost every participant was most interested in the implications of the proposition, or the result of what would come to pass. Almost all the participants also showed a strong disinclination towards television as a source of news. For these civically engaged users, television is valued primarily for entertainment purposes, while traditional long form journalism is the main source of news. Participants said this is due to a perceived lack of efficiency in watching news TV, in that the amount of time it takes to read an article gives the user more information than nearly half an hour of TV watching. Participants also showed a general trend towards wanting information in small doses. Several saw the government issued voter guide as overwhelming due to its format and large quantity of information. Finally, people tended to find the candidates more interesting than the propositions, due to the personalities of the candidates.

The final “How” section saw that participants use a wide variety of media sources, but tend to rely primarily on the same outlets. Many use Twitter as a gateway to in-depth journalism, and follow political scientists, journalists, or other prominent public figures, rather than friends and family. Facebook, on the other hand, is used as a place not just for receiving and reading articles, but also for sharing them and transmitting information. Aggregators such as Google and Yahoo, as well as sites like Reddit, were found to be preferred for breaking news coverage, while traditional news sites such as the New York Times are read on a daily basis. Overall, we saw a strong predilection for email news reminders.

The competitive analysis was looked through the lens of the focus groups. Because of the findings from the interviews, the content style and substance index was weighted less heavily than the user experience index. The focus groups showed a tendency towards fast, efficient portals to short amounts of information, rather than in-depth, well-researched information with inconvenient portals. Even if a certain guide had a high-rated content substance score, it would not be of relevance to the user, as he or she would not reach that content in the first place.


A shortcoming regarding the competitive analysis was the reliability in which the voter guides were rated. Because the numbers were compiled solely based upon the author’s opinion, the averages were skewed based on individual preferences.

A limitation of the user interviews was the tendency of participants to say what they wanted their habits to be, as opposed to what their habits actually were. Users were also unable to fully know and convey their preferences, and tended to believe they would use most formats of voter guides, regardless of what their past history suggested. We took these limitations into account when creating guide prototypes.  

Finally, sampling bias affected the selection of the participants of the user interviews. These participants were chosen based on their connections to CALmatters, which affected their responses and lifestyle habits.


Becoming an informed voter is integral to the idea of American democracy. If an innovative voter guide can reach a group of potential voters who do not have preset voting habits, this could have far-reaching consequences for our society at large.

Millennials have the power to become one of the largest interest groups politicians must cater to in their policy making and implementation. Millennials have surpassed baby boomers as the nation’s largest living generation, therefore becoming the largest potential voting bloc (U.S. Census Bureau). If this technology-savvy group were to have access to an innovative product that would simplify the voting process, politicians would be more likely to cater their agendas towards the millennial voting bloc.

This research showed that gender, rather than age, plays a large role in deciding what influences voter guide preferences. This could potentially change the ways in which voter guides are marketed and designed, especially in today’s current age as an increasing number of voters go elsewhere beyond the government’s official voter guide to seek information on their candidates and propositions. Although much of the research in this study is based on the preferences of individuals, and cannot holistically account for the greater population of the nation, today’s media world of voter guides and political information is often also based on individual preferences as well. Personal networks connect readers to blogs, and voter guides often stand for niche communities and groups.

Moreover, synthesizing information in ways that will make the public more inclined to read it has important effects on democracy. A main objective of this study was to discern how to best inform readers and make them interested in political material. As a nonprofit organization, CALmatters sought to make voters more informed for the general election– especially regarding the California propositions, which were often lost in mainstream media under the shadow of the presidential candidates. Comprehensive voter guides are sound solutions to ensuring that the voter goes into the polling station with a firm grasp on not just some of the issues on the ballot, but on all of them.   

Works Cited

DelReal, Jose A. “Voter Turnout in 2014 Was the Lowest since WWII.” Washington Post. The Washington Post, 10 Nov. 2014. Web. 29 July 2016.

Fry, Richard. “Millennials Overtake Baby Boomers as America’s Largest Generation.” Pew Research Center. Pew Research Center, 25 Apr. 2016. Web. 04 Aug. 2016.

Heimlich, Russell. “Who Votes, Who Doesn’t, and Why.” Pew Research Center for the People and the Press. Pew Research Center, 18 Oct. 2006. Web. 03 Aug. 2016.

“21.5% Youth Turnout: Two-Day Estimate Comparable to Recent Midterm Years.” CIRCLE. Tufts University Jonathan M. Tisch College of Civic Life, 2010. Web. 29 July 2016.

“Voter Guides.” Ballotpedia. N.p., n.d. Web. 23 Feb. 2017.


Figure 1: A Spectrum of Voters and Non-Voters: How They Differ
Source: Pew Research Center
Notes: 60 percent of intermittent voters say they do not know enough about candidates to vote.

Figure 2: Competitive Analysis of Prevalent Voter Guides
Notes: The three indices are user experience, user engagement, and content style and substance.

Figure 3: Competitive Analysis of Existing Voter Guides
Notes: The sites with the highest scores tended to use aggregation approaches.

Figure 4: Cross Tabulation of Age and Voter Guide Use
Notes: Usage of the government issued voter guide tends to increase with age, peaking at the 36-45 range.

Figure 5: Who would you most trust to give good advice on how to vote in an election?
Notes: When asked whom you would trust to give advice on voting, 28 percent of participants say they do not want or need advice on how to vote. 24 percent would trust family or friends, and 21 percent would trust politicians they support. Advocacy organizations were the least trusted group.

Figure 6: Millennials Conference Survey Results
Notes: The results showed that users were open to trying a wide variety of products, as long as they did not take up too much time.

Figure 7: Millennials Conference Survey Results Continued
Notes: Overall, the results saw a wide range of preferences, and the only definite trend was a general opinion against audio formats.

Figure 8: Profile of Participants
Notes: We presented our research using the design thinking approach– showcasing who the individual was, what we were surprised to learn, and what needs we could satisfy.

We met Khalid*, a Bay area resident for the past five years who still sees himself as new to the area, its candidates, and its concerns.
We were surprised to learn that he doesn’t seem to be interested in a diverse set of perspectives or in being challenged – he’s comfortable with his progressive/liberal identity and values.
It would be game changing if we could give him the depth of trusted information he needs to be the most informed progressive/liberal he can be.

We met Chantal*, an extremely active citizen who is especially passionate about the environment.
We were surprised to learn how much of her political engagement takes place in person.
It would be game changing if we could replicate the intimacy and direct connection she feels from those interactions in online communications, which she generally finds overwhelming.

We met Jose*, a political science undergraduate student whose identity as a Latino is an important factor in his political engagement
We were surprised to learn that he uses so many local news sources to get information related to the election.
It would be game changing if he could have a one-stop shop that provides a variety of perspectives on the propositions.

*Note: All names have been changed to protect the parties’ privacy.