Skip to content
16 February 2010

some thoughts on Facebook’s recent changes, from the perspective of an application designer

There’s a lot to like about the recent changes to Facebook, but, as an application developer, many of the changes are a mixed bag. Changes to navigation and to interaction points between Facebook and applications are problematic, while new application privacy features are a good start but seem incomplete.

Navigation to Apps
Formerly, the application dock made it easy to access an application from anywhere in Facebook. One click to get to a bookmarked application; two clicks (without waiting for page loads) to get to other applications. In the new homepage, this has been removed.

Not so with the new design. If an application doesn’t have a tab on your profile page, the only way to access it is from the home page. From my profile or someone else’s profile, this means: click to the Facebook home page, wait for the page to load, click the app icon (or, if the application is not one of your top three book marks, click more and then click the app icon). Yes, this is only one more click, but it requires waiting for an entire page load, and it’s worse for non-bookmarked apps: one click to the home page, wait for it to load, one click to the application dashboard, wait for it to load, one click for “see all of your applications,” wait for that to load, and finally click on the application.

One possible remedy might be to add an “applications” drop-down next to the new notifications, requests, and messages icons.

Notifications
At the end of the month, Facebook will turn off the ability of applications to send notifications. This is a method I’ve been using to send reminders in Three Good Things, for both automatically generated reminders and user-to-user reminders.

3GT Notifications

3GT Notifications. Left: system generated reminder. Right: user-to-user reminder.

I like that the notifications are less invasive than email reminders. Some 3GT users appreciated their subtlety, though they may have been a little too subtle, at least when they appeared at the bottom of the screen, as many of 3GT users we interviewed never noticed the notifications they received. More importantly, they allowed notifications at the right time. Rather than sending someone a reminder to post — a reminder that might interrupt their other activity or would at least require them to visit the website — the notifications appeared when a 3GT participant was already logged into Facebook, when it was likely convenient for them to post a “good thing” in our application. B.J. Fogg, a champion of persuasive technology, calls this right-time, right-place notification kairos.

I understand that notifications have gotten a lot more intrusive with the addition of push notifications to the iPhone app, and that some app developers have used them more than some Facebook users would prefer. Facebook has also added additional integration points. On the balance, though, I think notifications are going to be an unfortunate integration point to lose.

Application Privacy
Along with some others building health and wellness applications for the Facebook platform, I’ve felt fairly strongly that Facebook needs to give users and developers enhanced privacy controls for applications. At a minimum, this should include the ability to hide one’s use of an application from friends (i.e., not appearing under “friends using this application” in the application’s profile page).

With the recent updates Facebook’s designers and developers appear to have recognized some of these concerns. Application developers are able to set an application as “private,” causing one’s use to not appear in the new application dashboard. This is a good start, but it feels incomplete for a number of reasons. First, users, not developers, should have control of privacy. What’s to stop a developer from later reverting to a more public setting, instantly and completely changing what user activity is revealed? Currently, users do not have any way to remove this information once it appears, either.

Second, this level of privacy does not extend far enough. Friends who use private applications still appear on the application’s profile page under “friends using this application.” Furthermore, the model of application use and content being either private or public is insufficient. In health and wellness applications, for example, participants may benefit from sharing and interacting with other participants in the intervention as well as their friends or family members on Facebook, while also wanting to keep their activity private from coworkers.

This is something that Facebook has already discovered and addressed with the newsfeed (now stream) content, but application content does not enjoy the same privacy controls. To share with only a subset of one’s friends within an application, however, application developers must implement their own social graph features and users must built a second network within the application. Enabling privacy controls similar to those for the newsfeed for application content and use could help people to feel more comfortable using health wellness applications on Facebook while creating more possibilities for designers. A user could only allow an application to be aware of relationships in one or more friend lists or networks, or to select some to exclude from the application while allowing the other connections to remain visible. Assuming a user had created the necessary friend lists or that their privacy preferences mapped to their networks, this would allow someone to filter out their coworkers or to allow only close friends to see their participation.

Update: As of 17 February, Facebook has added a privacy argument to calls for publishing methods (e.g. Stream.publish that gives applications much more control over shared content.

Tags: ,
Comments (0)
28 January 2010

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.

Tags: , , , ,
Comments (0)
24 December 2009

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.

Tags: , , ,
Comments (1)
20 July 2009

three good things

Three Good ThingsThe first of my social software for wellness applications is available on Facebook (info page).

Three Good Things supports a positive psychology exercise in which participants record three good things, and why these things happened. When completed daily – even on the bad days – over time, participants report increased happiness and decreased symptoms of depression. The good things don’t have to be major events – a good meal, a phone call with a friend or a family member, or a relaxing walk are all good examples.

I’m interested in identifying best practices for deploying these interventions on new or existing social websites, where adding social features may make the intervention more or less effective for participants, or may just make some participants more likely to complete the exercise on a regular basis. Anyway, feel free to give the app a try – you’ll be helping my research and you may end up a bit happier.

Tags: , , , , , , ,
Comments (2)
18 July 2009

new opportunities for management experience

résumés in the era of mechanical turk
Update: Per Jude’s suggestion, you can now get stickers and shirts at Café Press.
Tags: , ,
Comments (0)
12 June 2009

privacy on twitter vs. privacy on facebook

In a post describing some teens’ use of Twitter and Facebook (Twitter is for friends; Facebook is everybody; some teens are using private Twitter accounts for communication with friends because Twitter is too public), danah boyd poses the following question:

My guess is that if Twitter does take off among teens and Dylan’s friends feel pressured to let peers and parents and everyone else follow them, the same problem will arise and Twitter will become public in the same sense as Facebook. This of course raises a critical question: will teens continue to be passionate about systems that become “public” (to all that matter) simply because there’s social pressure to connect to “everyone”?

I believe that Twitter may actually be much more resistant to both this pressure and subsequent switch to less “public” platforms than Facebook for two reasons: account norms and Twitter clients.

Account Norms, Privacy, and Collapsed Contexts
On Facebook, everyone pretty much gets one account.1 This leaves me with a choice of collapsed contexts (same profile for everyone) or only friending people from a particular context or set of context. There are many fine-grained privacy controls, but this all adds up to a more-is-less experience, at least for me. There are enough many controls that I don’t particularly remember what I’ve set to be visible to whom. When I comment on something in friend’s profile (or am tagged in one of their photos), I don’t know who can see that.

With Twitter, people can have multiple accounts, and for private accounts, they know exactly who can see their posts: only people who I give permission. This is not to say Twitter is not without some privacy pitfalls – e.g. plenty of private tweets get retweeted or replies on others’ public accounts – but I have a much clearer idea of who can see a status update or reply on Twitter than I do of who can see similar content on Facebook at the time of posting. I suspect that many users of private Twitter accounts do so just to avoid the “what if so-and-so sees this?” question. So it seems reasonable that people could have different accounts for their work, family, friends, etc personas, though there’s a point at which it probably would be too many.

Twitter Clients
Having multiple accounts wouldn’t work well without an appropriate interface, and here Twitter benefits hugely from its API and the many, many Twitter clients available. Using more than one Facebook account, especially simultaneously, is an ordeal – multiple web browsers, no aggregation. With the right client, reading from and posting to multiple Twitter accounts is a breeze.

So while there may eventually be an exit from a more public Twitter, I think there is more room to move within the same service, diversifying accounts, than there might be on Facebook. This will only, work, though if people are willing to set boundaries and accept boundaries – and probably not if mom and dad insist on following the Twitter account their kids use to communicate with friends from school, or if colleagues regularly feel insulted when a coworker-acquaintance declines their request to follow an account they use to communicate with close friends.

1I believe this used to be part of the terms of service, but I don’t see it anymore and can’t be sure that it was ever there.

Tags: , , ,
Comments (0)
26 May 2009

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 Digg.com 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.

Tags: , , , , , , , , ,
Comments (0)
19 May 2009

SI182 Final Projects

A belated congrats to all of the EECS182/SI182 students on finishing the semester. For those not familiar with the course, SI182 is an intro to programming course in the informatics program at UM. Paul Resnick and I taught it this past semester, and arranged the course around pulling data from public feeds, processing this data, and presenting it again, online, in a way that adds value.

Here’s a sampling of the final projects:

Also, a huge thanks to Chuck Severance, who got this course started and gave us early chapters of his book Using Google App Engine, which gave us the confidence to use App Engine in the course and which we were able to rely on for class readings.

Tags: , , , , , , , ,
Comments (0)
2 February 2009

ann arbor craigslist housing ads mapped

I tend to begin my housing search on Craigslist, with one or more general areas where I’d like to live in mind. Because location matters to me, I’ve found HousingMaps.com to be incredibly helpful. Unfortunately, it’s doesn’t include Ann Arbor. Other sites do, but aren’t really compatible with how I search for housing – I tend to search across rooms & shares and apartment & house rentals. I do have a definite price ceiling. I’ll often have a potential housemate or two, but with some flexibility in case preferences ultimately diverge. I also want to be able to limit a display to just new listings since I last checked.

So, after growing impatient with existing tools, put together a map of Ann Arbor Craigslist housing listings (sorry, I don’t include homes for sale. It’s not the right time to buy, anyway). You can:

  • simultaneously display each of the listing types,
  • Filter by price, date posted, and number of bedrooms (all as sliders with min and max), and
  • adjust the price filters to work on price per bedroom.

This was pretty fast to throw together thanks to BeautifulSoup, the Google Maps API, and YUI, but I’m sure that there are some rough edges that will need to get worked out.

Also, I don’t know or care whether it works in IE.

Update, 19 February: More locations now available.

Tags: , , , ,
Comments (2)
9 December 2008

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.

Tags: , , , , , , ,
Comments (1)