Skip to content

{ Tag Archives } politics

Sunlight Labs’ Inbox Influence: Sunlight or Sunburn?

Last week, Sunlight Labs released Inbox Influence, a set of browser extensions (Chrome, Firefox) and bookmarklets that annotate senders and entities in the body of emails with who has contributed to them and to whom they have contributed.

I really like the idea of using browser plugins to annotate information people encounter in their regular online interactions. This is something we’re doing on a variety of projects here, including AffectCheck, BALANCE, and Rumors. I think that tools that combine personal data, in-situ, with more depth can teach people more about with whom and with what they are interacting, and this just in time presentation of information is an excellent opportunity to persuade and possibly to prompt reflection. Technically, it’s also a pretty nice implementation.

There are some reasons why this tool may not be so great, however. With Daniel Avrahami, Sunny Consolvo, James Fogarty, Batya Friedman, and Ian Smith, I recently published a paper about people’s attitudes toward the online availability of US public records, including campaign contribution records such as the ones on which Inbox Influence draws. Many respondents to our survey (mailed, no compensation, likely biased toward people who care more about this issue) expressed discomfort with these records being so easily accessible, and less than half (as of 2008) even knew that campaign contribution records were available online before they received the survey. Nearly half said that they wanted some sort of change, and a third said that this availability would alter their future behavior, i.e., they’d contribute less (take this with a grain of salt, since it is about hypothetical future behavior).

Unless awareness and attitudes have changed quite a bit from 2008, tools such as Inbox Influence create privacy violations. The data is being used and presented in ways that people did not anticipate at the time when they made the decision to donate, and at least some people are “horrified” or at least uncomfortable with this information being so easily accessible. Perhaps we just need to do better at educating potential donors about in what ways campaign contribution data may be used (and anticipate future mashups), though it is also possible that tools like this do not need to be made, or could benefit from being a bit more nuanced in when and about whom they load information.

Speaking personally, I’m not sure how I feel. On the one hand, I think that campaign contributions and other other actions should be open to scrutiny and should have consequences. If you take the money you earn from your business and donate it to support Prop 8, I want the opportunity to boycott your business. If you support a politician who wants to eviscerate the NSF, I might want to engage you in conversation about that. On the other hand, I don’t like the idea that my campaign contribution history (anything above the reporting limit) might be loaded automatically when I email a professional colleague or a student. That’s just not relevant—or even appropriate—to the context. And there are some friendships among political diverse individuals that may survive, in part, because those differences are not always made salient. So it also seems like Inbox Influence or tools that let you load, with a click, your Facebook friends’ contribution history, could sometimes cause harm.

CHI Highlights: Diversity, Discussion, and Information Consumption

For my second post on highlights from CHI this year, I’m focusing on papers related to opinion diversity and discourse quality.

  • Normative influences on thoughtful online participation
    Abhay Sukumaran, Stephanie Vezich, Melanie McHugh, Clifford Nass

    Two lab experiments on whether it is possible to foster more thoughtful commenting and participation by participants in online discussion forums by priming thoughtful norms. The first tested the effects of the behavior of other participants in the forum. The dependent variables were comment length, time taken to write the comments, and number of issue-relevant thoughts. Not surprisingly, being exposed to other thoughtful comments led people to make more thoughtful comments themselves. One of the audience members asked the question about whether this would break down with just one negative or less thoughtful comment (such as how merely one piece of litter seems to break down antilittering norms).

    The second study tested effects of visual, textual, and interaction design features on the same dependent variables. The manipulations included a more subdued vs. more playful visual design, differing CAPTCHAs (words positively correlated with thoughtfulness in the thoughtful condition and words negatively correlated with thoughtfulness in the unthoughtful condition), and different labels for the comment box. The design intended to provoke thoughtfulness did correspond to more thoughtful comments, suggesting that it is possible, at least in the lab, to design sites to prompt more thoughtful comments. For this second study in particular, I’m curious if these measures only work in the short term or if they would work in the long term and about the effects of each of the specific design features.

  • I will do it, but I don’t like it: User Reactions to Preference-Inconsistent Recommendations
    Christina Schwind, Jürgen Buder, Friedrich W. Hesse

    This paper actually appeared a in health session, but I found that it spoke much more to the issues my colleagues and I are confronting in the BALANCE project. The authors begin with the observation that most recommender systems are intended to produce content that their users will like, but that this can be problematic. In the health and wellness domain, people sometimes need to hear information that might disagree with their perspective or currently held beliefs, and so it can be valuable to recommend disagreeable information. In this Mechanical Turk-based study, subjects were equally likely to follow preference-consistent and preference-inconsistent recommendations. Following preference-inconsistent recommendations did reduce confirmation bias, but people were happier to see preference-consistent recommendations. This raises the important question: subjects may have followed the recommendation the first time, but now that they know this system gives recommendations they might not like, will they follow the recommendations less often in the future, or switch to another system altogether?

  • ConsiderIt: improving structured public deliberation
    Travis Kriplean, Jonathan T. Morgan, Deen Freelon, Alan Borning, Lance Bennett

    I really like the work Travis is doing with Reflect and ConsiderIt (which powers the Living Voters Guide) to promote more thoughtful listening and discussion online, so I was happy to see this WiP and am looking forward to seeing more!

  • Computing political preference among twitter followers
    Jennifer Golbeck, Derek L. Hansen

    This work uses follow data for Congresspeople and others on Twitter to assess the “political preference” (see comment below) of Twitter users and the media sources they follow. This approach and the information it yields has some similarities to Daniel Zhou’s upcoming ICWSM paper and, to a lesser extent, Souneil Park’s paper from CSCW this year.

    One critique: despite ample selective exposure research, I’m not quite comfortable with this paper’s assumption that political preference maps so neatly to political information preference, partly because I think this may be an interesting research question: do people who lean slightly one way or the other prefer information sources that may be more biased than they are? (or something along those lines)

In addition to these papers, Ethan Zuckerman’s closing plenary, Desperately Seeking Serendipity, touched on the topics of serendipity and homophily extensively. Zuckerman starts by suggesting the reason that people like to move to cities – even at times when cities were really quite horrible places – is, yes, for more opportunities and choices, but also “to encounter the people you couldn’t encounter in your rural, disconnected lifestyle… to become a cosmopolitan, a citizen of the world.” He goes on, “if you wanted to encounter a set of ideas that were radically different than your own, your best bet in an era before telecommunications was to move to a city.” There reasons to question this idea of cities as a “serendipity engine,” though: even people in urban environments have extremely predictable routines and just don’t go all that many places. Encounters with diverse others may not as common as idealized.

He then shifts gears to discuss what people encounter online. He walks through the argument that the idea of a Freshman Fishwrap or Daily Me is possibly quite harmful as it allows people to filter to the news that they want. Adding in social filters or getting news through our friends can make this worse. While Sunstein is concerned about this leading to polarization within the US, Zuckerman is more concerned that it leads people to see only news about where they are and less news about other places or from outside perspectives. This trend might lead people to miss important stories.

I tend to agree with the argument that surfacing coincidences or manufacturing serendipity is an incredibly powerful capability of current technology. Many of the approaches that the design community has taken to achieve this are probably not the kind of serendipity Zuckerman is looking for. I love Dopplr’s notifications that friends are also in town, but the time I spend with them or being shown around by them is time that I’m less likely to have a chance encounter with someone local or a traveler from elsewhere. The ability to filter reviews by friends may make for more accurate recommendations, but I’m also less likely to end up somewhere a bit different. Even serendipity has been repurposed to support homophily

Now, it might be that the definition of serendipity that some of the design community hasn’t quite been right. As Zuckerman notes, serendipity usually means “happy accident” now – it’s become a synonym for coincidence – and that the sagacity part of the definition has been lost. Zuckerman returns to the city metaphor, arguing for a pedestrian-level view. Rather than building tools for only efficiency and convenience, build tools and spaces that maximize the chances to interact and mix. Don’t make filters hidden. Make favorites of other communities visible, not just the user’s friends. Zuckerman elegantly compares this last feature to the traces in a city: one does not see traces left just by one’s friends, no, but traces left by other users of the space, and this gives people a chance to wander from the path they were already on. One might also overlay a game on a city, to encourage people to explore more deeply or venture to new areas.

While I like these ideas, I’m a little concerned that they will lead to somewhat superficial exposure to the other. People see different others on YouTube, Flickr, or in the news, and yes, some stop and reflect, others leave comments that make fun of them, and many others just move on to the next one. A location-based game might get people to go to new places, but are they thinking about what it is like to be there, or are they thinking about the points they are earning? This superficiality is something I worry about in my own work to expose people to more diverse political news – they may see it, but are they really considering the perspective or gawking at the other side’s insanity? Serendipity may be necessary, but I question whether it is sufficient. We also need empathy: technology that can help people listen and see others’ perspectives and situations. Maybe empathy is part of the lost idea of sagacity that Zuckerman discusses — a sort of emotional sagacity — but whatever it is, I need to better know how to design for it.

For SI and UM students who really engage with this discussion and the interweaving of cities, technology, and flows, I strongly, strongly recommend ARCH 531 (Networked Cities).

Using Mechanical Turk for experiments

In my upcoming CHI paper, “Presenting Diverse Political Opinions: How and How Much,” we used Amazon’s Mechanical Turk (AMT) to recruit subjects and to administer the study. I’ll talk a bit more about the research questions and results in a future post, but I’ve had enough questions about using Mechanical Turk that I think a blog post may be helpful.

In this study, Paul Resnick and I explored individuals’ preferences for diversity of political opinion in online news aggregators and evaluated whether some very basic presentation techniques might affect satisfaction with the range of opinions represented in a collection of articles.

To address these questions, we needed subjects with known political preferences, from the United States, and with at least some very basic political knowledge, and we wanted to collect some demographic information about each subject. Each approved subject was then assigned to either a manipulation check group or to the experimental group. Subjects in the manipulation check group viewed individual articles and indicated their agreement or disagreement with each; subjects in the experimental group viewed entire collections and answered questions about the collection. The subjects in the experimental group were also assigned to a particular treatment (how the list would appear to them). Once approved, subjects could view a list up to once per day.

Screening. To screen subjects, we used a Qualification test in AMT. When unqualified subjects viewed at task (HIT – human intelligence task, in mTurk parlance), they were informed that they needed to complete a qualification. The qualification test asked subjects two questions about their political preferences, three multiple choice questions about US politics, and a number of demographic questions. Responses were automatically downloaded and evaluated to complete screening and assignment.

To limit our subjects to US residents, we also used the automatic locale qualification.

Assignment. We handled subject assignment in two ways. To distinguish between the treatment group and the manipulation check group, we created to additional qualifications that were automatically assigned; an approved subject would be granted only one of these qualifications, and could thus could only complete the associated task type.

Tasks (HITs). The task implementation was straightforward. We hosted tasks on our own server using the external question HIT type. When a subject loaded a task, AMT passed us the subject’s worker ID. We verified that the subject was qualified for the task and loaded the appropriate presentation for that subject. Each day, we uploaded one task of each type, with many assignments; assignments are the number of turkers that can complete each task.

Because we needed real-time access to the manipulation check data, the responses to this task were stored in our own database after a subject submitted the form; the subject could then return to AMT. This was not necessary for the experimental data, and so the responses were sent directly to AMT for later retrieval.

Quality control. Careless clicking or hurrying through the task is a potential problem on mTurk. Using multiple raters does not work when asking subjects about their opinions. Kittur, Chi, and Suh recommend asking Turkers verifiable questions as a way to deal with the problem1. We did not, however, ask verifiable questions about any of the articles or the list, because that might have changed how turkers read the list and responded to our other questions. Instead, we randomly repeated a demographic question from the qualification test. 5 subjects changed their answer substantially (e.g. aging more than one year or in reverse or shifting on either of the political spectrum questions by 2 points or more). Though there are many possible explanations for these shifts – such as shared accounts within a household, careless clicking, easily shifting political opinions, deliberate deception, or lack of effort – all of these explanations are not desirable for study subjects, and so they were excluded. We also examined how long each subject took to complete tasks (looking for implausibly fast responses); this did not lead to the exclusion of any additional subjects or responses.

Some reflection. We had to pay turkers a bit more than we expected (~$12/hr) and we recruited fewer subjects than we anticipated. The unpaid qualification task may be a bit of a barrier, especially because potential subjects could only complete one of our paid tasks per day (and only one was listed at a time). Instead, we might have implemented the qualification as a paid task, but that might result in paying for subjects who would never return to complete an actual task.

Further resources

1. Kittur, A., Chi, E. H., and Suh, B. (2009). “Crowdsourcing User Studies With Mechanical Turk,” Proc. CHI 2009: 453-456. (ACM | PARC)
2. Mason, W. and Watts, D. J. (2009). “Financial incentives and the ‘performance of crowds,’” SIGKDD Workshop on Human Computation: 77-85. (ACM | Yahoo)

This study is part of the BALANCE project and was funded by NSF award #IIS-0916099.

updated viz of political blogs’ link similarity

I’ve been meaning to post a simple update to my previous visualization of political blogs’ link similarities. In the previous post, I used GEM for layout, which was not, in hindsight, the best choice.

In the visualization in this post, the edges between blogs (the nodes, colorized as liberal, independent, and conservative) are weighted as the Jaccard similarity between any two blogs. The visualization is then laid out in GUESS using multidimensional scaling (MDS) based on the Jaccard similarities.

Sidelines at ICWSM

Last week I presented our first Sidelines paper (with Daniel Zhou and Paul Resnick) at ICWSM in San Jose. Slides (hosted on slideshare) are embedded below, or you can watch a video of most of the talk on VideoLectures.

Opinion and topic diversity in the output sets can provide individual and societal benefits. If news aggregators relying on votes and links to select and subsets of the large quantity of news and opinion items generated each day simply select the most popular items may not yield as much diversity as is present in the overall pool of votes and links.

To help measure how well any given approach does at achieving these goals, we developed three diversity metrics that address different dimensions of diversity: inclusion/exclusion, nonalienation, and proportional representation (based on KL divergence).

To increase diversity in result sets chosen based on user votes (or things like votes), we developed the sidelines algorithm. This algorithm temporarily suppresses a voter’s preferences after a preferred item has been selected. In comparison to collections of the most popular items, from user votes on and links from a panel of political blogs, the Sidelines algorithm increased inclusion while decreasing alienation. For the blog links, a set with known political preferences, we also found that Sidelines improved proportional representation.

Our approach differs and is complementary to work that selects for diversity or identifies bias based on classifying content (e.g. Park et al, NewsCube; ) or by classifying referring blogs or voters (e.g. Gamon et al, BLEWS). While Sidelines requires votes (or something like votes), it doesn’t require any information about content, voters, or long term voting histories. This is particularly useful for emerging topics and opinion groups, as well as for non-textual items.

visualizing political blogs’ linking

There are a number of visualizations of political bloggers’ linking behavior, notably Adamic and Glance’s 2005 work that found political bloggers of one bias tend to link to others of the same bias. Also check out Linkfluence’s Presidential Watch 08 map, which indicates similar behavior.

These visualizations are based on graphs of when one blog links to another. I was curious to what extent this two-community behavior occurs if you include all of the links from these blogs (such as links to news items, etc). Since I have link data for about 500 blogs from the news aggregator work, it was straightforward to visualize a projection of the bipartite blog->item graph. To classify each blog as liberal, conservative, or independent, I used a combination of the coding from Presidential Watch, Wonkosphere, and my own reading.

Projection of links from political blogs to items (Oct - Nov 2008)

Projection of links from political blogs to items (Oct - Nov 2008). Layout using GEM algorithm in GUESS.

The visualization shows blogs as nodes. Edges represent shared links (at least 6 items must be shared before drawing an edge) and are sized based on their weight. Blue edges run between liberal blogs, red edges between conservative blogs, maroon between conservative and independent, violet blue between liberal and independent, purple between independent blogs, and orange between liberal and conservative blogs. Nodes are sized as a log of their total degree. This visualization is formatted to appear similar to the Adamic and Glance graph, though there are some important differences, principally because this graph is undirected and because I have included independent blogs in the sample.

This is just a quick look, but we can see that the overall linking behavior still produces two fairly distinct communities, though a bit more connected than just the graph of blog to blog links. It’d be fun to remove the linked blog posts from this data (leaving mostly linked news items) to see if that changes the picture much. Are some media sources setting the agenda for bloggers of both parties, or are the conservative bloggers reading and reacting to one set news items and liberal bloggers reading and reacting to another? I.e., is the homophily primarily in links to opinion articles, or does it also extend to the linked news items?

I’m out of time at this point in the semester, though, so that will have to wait.

bias mining in political bloggers’ link patterns

I was pretty excited by the work that Andy Baio and Joshua Schachter did to identify and show the political leanings in the link behavior of blogs that are monitored by Memeorandum. They used singular value decomposition [1] on an adjacency matrix between sources and items based on link data from 360 snapshots of Memeorandum’s front page.

For the political news aggregator project, we’ve been gathering link data from about 500 blogs. Our list of sources is less than half of theirs (I only include blogs that make full posts available in their feeds), but we do have full link data rather than snapshots, so I was curious if we would get similar results.

The first 10 columns of two different U matrices are below. They are both based on link data from 3 October to 7 November; the first includes items that had an in-degree of at least 4 (5934 items), the second includes items with an in-degree of at least 3 (9722 items). In the first, the second column (v2) seems to correspond fairly well to the political leaning of the blog; in the second, the second column (v3) is better.

I’ll be the first to say that I haven’t had much time look at these results in any detail, and, as some of the commenters on Andy’s post noted, there are probably better approaches for identifying bias than SVD. If you’d like to play too, you can download a csv file with the sources and all links with an in-degree >= 2 (21517 items, 481 sources). Each row consists of the source title, source url, and then a list of the items the source linked to from 3 October to 7 November. Some sources were added part way though this window, and I didn’t collect link data from before they were added.

[1] One of the more helpful singular value decomposition tutorials I found was written by Kirk Baker and is available in PDF.

US political news and opinion aggregation

Working with Paul Resnick and Xiaodan Zhou, I’ve started a project to build political news aggregators that better reflect diversity and represent their users, even when there is an unknown political bias in the inputs. We’ll have more on this to say later, but for now we’re making available a Google gadget based on a prototype aggregator’s results.

The list of links is generated from link data from about 500 blogs and refreshed every 30 minutes. Some of the results will be news stories, some will be op-ed columns from major media services, others will be blog posts, and there are also some other assorted links.

At this early point in our work, the results tend to be more politically diverse than an aggregator such as Digg, but suffer from problems with redundancy (we aren’t clustering links about the same story yet). As our results get better, the set of links the gadget shows should improve.

Update 15 December: I twittered last week that I’ve added bias highlighting to the widget, but I should expand a bit on that here.

Inspired by Baio and Schachter’s coloring of political bias on Memeorandum, I’ve added a similar feature to the news aggregator widget. Links are colored according the average bias of the blogs linking to them. This is not always a good predictor of the item’s bias or whether it better supports a liberal or conservative view. Sometimes a conservative blogger writes a post to which more liberal bloggers than conservative bloggers, and in that case, the link will be colored blue.

If you don’t like the highlighting, you can turn it off in the settings.


A note came across the anthrodesign list earlier this week about, a website on which people who disagree about a certain issue are paired up, given some guidelines for productive conversation, and then have a facilitated discussion in reaction to a provided scenario.

One of my concerns about virtual community and online discourse has been the ease with which people can choose to associate primarily with people who agree with them. This is particularly true in the blogosphere and is why I stopped political blogging after the 2004 election. Lately I’ve been looking more towards online interactions that are not overtly about politics as a possible source for building mutual respect and understanding as a foundation for constructive civic engagement. I suppose that this is a more subtle approach.

RedBlue, in contrast, is anything but subtle. I’m very interested in seeing where this goes. They’re currently recruiting testers, and if anyone happens to give it a try, I’d love to hear your thoughts.

California sues automakers

While reading the news the other day, I noticed that California’s attorney general, Bill Lockyer, filed suit against six automakers. Interesting. Lockyer is “demanding that they pay for environmental damage caused by the emissions of their vehicles.”

Interesting, but irrational (as far as I can tell).

To the best of my knowledge, the automakers have remained in compliance with state and federal laws regulating emissions. Turning and filing suit against companies that have been compliant feels a bit like playing “gotcha.” If the emissions laws were insufficient, he should be complaining about the federal government and the state of California. It he’s looking for the real culprit, though, he would be suing the people of California who just keep on using their cars so much despite the known environmental consequences. That probably wouldn’t work so well for Mr. Lockyer, though, as he currently wants those same constituents to elect him to state treasurer in the fall.