How Much Does Facebook Really Know About You? Predicting Motives For Facebook Use From Logged Data

People consume media for many different reasons and, indeed, individuals use a wide range of channels to achieve diverse ends: entertainment, edutainment, information seeking, communication, socializing, and surveillance to name but a few. Uses and Gratifications (U&G) is a theoretical framework that facilitates studying these motives and outcomes – casting light on the how and why of media consumption. U&G has been applied to formats as diverse as tabloids, reality TV, mobile phones and, lately, social network sites. A U&G study typically involves three elements: motives for media use, social and psychological antecedents, and cognitive, attitudinal, or behavioral outcomes. The value of U&G lies in combining these factors: the motives people have for using a service are seen in the context of specific kinds of users (e.g. via antecedents such as demographics) and the behaviors they engage in (e.g. outcomes such as purchasing content).

The combination of the three elements is at the heart of our approach.
The combination of the three elements is at the heart of our approach.

Most U&G studies collect this information from surveys or questionnaires. While this technique is good at eliciting motives – asking why – it is weaker at capturing antecedents and outcomes. Indeed, data for these categories is typically limited to a few basic demographic and usage questions: asking about age, gender, and time or frequency of use. Beyond being coarse-grained, such information can also be hard to recall accurately. To address these problems, we studied Facebook motivations with a data-driven approach. We surveyed motives and simultaneously collected a rich dataset from the Facebook API. This included eleven usage metrics (e.g. number of photographs, “likes” given, events attended) and eight personal network metrics (e.g. network size, density, connected components) that we used as novel outcomes and antecedents. This data provided novel insights into Facebook use.

– In line with prior work, we identified seven motives for Facebook use: social connection, shared identities, photographs, content, social investigation, social network surfing and newsfeed.

– We showed that data gathered from the Facebook API predicts these motives. For instance, we can determine if Facebook users are primarily interested in connecting with others, looking at photographs, or stalking, based directly on their usage patterns and network structure.

– We gained a deeper understanding of motives for using Facebook. For example, the Social Investigation motive was linked to longer times spent on site but negatively linked with posting status updates. This profile highlights the “lurkers” on Facebook: the silent majority who observe but don’t post.

For more, see our full paper, Understanding Motivations for Facebook Use: Usage Metrics, Network Structure, and Privacy.

Tasos Spiliotopoulos, Madeira Interactive Technologies Institute, University of Madeira
Ian Oakley, Ulsan National Institute of Science and Technology

Warping Time for More Effective Real-Time Crowdsourcing

REAL-TIME TRANSCRIPTION is a vital accommodation for deaf and heard of hearing people in their daily lives. Captioning is typically expensive due to the years of training that is required.

LEGION:SCRIBE introduced a method that used multiple non-experts to caption audio with high quality and low latency, at far lower costs. We recently developed TimeWarp to help make the task easier for individual workers without hurting collective performance. TimeWarp makes each captionist’s job easier by selectively slowing down and speeding up the playback speed of the audio.

Warping time with 4 workers
An example of time warping using 4 workers. The workers each individually hear their segment (matched by color) in slow motion. However, the crowd can collectively still keep up in real-time.

OFFLINE CAPTIONISTS OFTEN SLOW DOWN AUDIO to make it easier to caption. However, this necessarily puts the worker behind real-time. That’s fine for offline captioning, but means it can’t be used by one person and still keep up with real-time speech.

TimeWarp relies on:

  • People’s ability to hear faster than they can type
  • Scribe’s need for workers to only caption a small part of what they hear

For the parts of the audio workers as asked to type, the audio is played slower. In order to catch up with real-time, the audio is played slightly faster during parts in between where the worker listens for context.


  • 12.6% mean improvement in accuracy
  • 11.4% mean improvement in coverage
  • 19.1% mean improvement in latency

The surprising improvement in latency is due to workers being able to keep up with each word as it was said, instead of memorizing it and then typing it later.


For more, see our full paper, Warping Time for More Effective Real-Time Crowdsourcing.
Walter S. Lasecki, University of Rochester
Christopher D. Miller, University of Rochester
Jeffrey P. Bigham, University of Rochester

Don’t Hide in the Crowd! Increasing Social Transparency Between Peer Workers Improves Crowdsourcing Outcomes

Although crowdsourcing is a useful social computing technique, its unreliability has greatly undermined its utility. In this study, we found that carefully manipulating the social transparency and various peer-dependent reward schemes can successfully motivate crowds to generate high-quality work.

Previous research has shown that social transparency can make people more accountable for their own actions in online collaborative work. Nevertheless, it is not easy to utilize social transparency in crowdsourcing since crowd workers usually work individually. Our previous study on peer consistency evaluation demonstrated that simply making the rewards of crowd workers depend on each other can create social effects between workers, motivating them to perform significantly better. However, not all peer-dependent reward schemes create positive social effects for collaborative work. The possible social effects of peer-dependent reward schemes from previous literature are summarized below:

  • Altruistic Motives: crowds may work harder to benefit their colleagues (Bandiera et al. Quarterly Journal of Economics ‘05)
  • Social Loafing: crowds may feel that they can hide in the crowds because personal effort is hard to evaluate (Karau et al. Journal of Personality and Social Psychology ‘93)
  • Social Facilitation: crowds may perform better because they think their work can be used as a point of reference for others in the group (Harkins, Journal of Experimental Social Psychology ‘87)

We conducted a 3X2 experiment by setting two levels of social transparency (low: anonymous, high: demographic information revealed) and three different peer-dependent reward schemes (individual, teamwork, competition). The main findings of our experiment are as follows:

1. Social transparency successfully motivated crowds to generate high-quality outcomes when their rewards were codependent.


When the workers worked individually, there was no significant difference between the workers who were anonymous and those who shared their demographic information. However, when making the rewards of the workers codependent, the difference between the performances of the two groups became significant. This shows that connecting the crowds is the key for us to utilize social transparency to enhance the reliability of crowdsourcing outcomes.

2. Social loafing harms the performance of crowds when only the collective outcomes are evaluated.

In team environments, the rewards of the workers were decided by the average of their performance and their teammates. We found that when a crowd worker was paired with a teammate that had good performance, they performed significantly worse. This result shows that social loafing really creates a negative effect on crowd work when the personal effort is difficult to evaluate.


3.  Social facilitation was effective only when there was social transparency between the crowds.

In competitions, the rewards of crowds were decided by the positive difference between their performances and their opponent’s. We found that, when workers shared their demographic information, the workers were motivated to outperform their opponents. However, this effect did not exist when the participants worked anonymously, which indicates that social transparency makes social facilitation more effective.


For more, see our full paper, Don’t Hide in the Crowd! Increasing Social Transparency Between Peer Workers Improves Crowdsourcing Outcomes

Shih-Wen Huang, University of Illinois at Urbana-Champaign

Wai-Tat Fu, University of Illinois at Urbana-Champaign

CommunityCompare: Visually Comparing Communities for Online Community Leaders in the Enterprise

We introduce CommunityCompare, a visual analytic system to enable online community leaders to make sense of their community’s activity with comparisons.CommunityCompare

Emma, an online community leader, wants to learn how she might increase member visits (views) by looking at example communities that are doing well on this metric. Each line in the main chart represents a community; Emma’s community is highlighted in orange. She wants to find example communities that have similar size and age compared to her community, and at the same time have more views than hers. To achieve this, she creates two visual filters on the community size and community age axes around her community, and creates another filter on the views axis above her community. She now sees 8 communities shown in the main chart and listed in the Select Communities widget. She wants to know what has been posted in these communities, so she clicks on the name of each community in the list to view their Most Valuable Posts below. She sees several examples of posts that give her new ideas for discussion topics in her community. She then saves the list of communities and decides to contact the leaders to learn more about their activities.

We motivate and inform the system design with formative interviews of community leaders. From additional interviews, a field deployment, and surveys of leaders, we show how the system enabled leaders to assess community performance in the context of other comparable communities, learn about community dynamics through data exploration, and identify examples of top performing communities from which to learn.

For more, see our full paper, CommunityCompare: Visually Comparing Communities for Online Community Leaders in the Enterprise.
Anbang Xu, University of Illinois at Urbana-Champaign
Jilin Chen, IBM Research – Almaden
Tara Matthews, IBM Research – Almaden
Michael Muller, IBM Research – Cambridge
Hernan Badenes , IBM Research – Almaden

Web Tutorials as a Gathering Place for Community Contribution

Web-based tutorials play an important role in how people learn and use complex software. Unfortunately, tutorials aren’t always as helpful as they could be. The quality of the instruction may be poor (or just not matched to the user’s level of knowledge), and a tutorial may not exist for the exact task the user is trying to do, forcing them to try and adapt tutorials for similar tasks.

Our paper investigates community enhanced tutorials, a new kind of web tutorial system with the potential to enable tutorials that improve as they are used by a community of users. This is achieved by embedding a fully functioning application into the tutorial, turning it into a hub for both learning and actually performing the tutorial task. As the tutorial is used to complete tasks, it can record users’ efforts and generate alternate demonstrations of each tutorial step.

System diagram for Community Enhanced Tutorials
Community Enhanced Tutorials improve over time as they are used by a community of users.

From a crowdsourcing perspective, community enhanced tutorials have two main advantages. First, they create a concrete gathering place for users interested in a particular task in the application. We looked at how this could be used to collect demonstrations from users, but this gathering of users could be mobilized in other ways as well. Small contributions could be solicited from these users to improve the tutorial content, or to assist other users directly.

Second, because community enhanced tutorials exist on the web, they are compatible with all the incentives that drive users to create traditional tutorials, such as earning recognition, ad revenue, or membership fees.

Our paper presents FollowUs, a prototype community enhanced tutorial system that we created to test these ideas. We included a range of features for browsing and working with multiple video demonstrations in a tutorial, which you can see in our video.

We also conducted a study to answer a key question that underlies this approach: Can providing additional demonstrations make a tutorial more robust? We found that users do make use of additional demonstrations when they are available, and our results suggest that multiple demonstrations can improve a tutorial’s quality and make it more widely applicable.

For more, see our full paper, Community Enhanced Tutorials: Improving Tutorials with Multiple Demonstrations.
Ben Lafreniere, University of Waterloo
Tovi Grossman, Autodesk Research
George Fitzmaurice, Autodesk Research

Limiting, Leaving, and (re)Lapsing: An Exploration of Facebook Non-Use Practices and Experiences

With over one billion active accounts world-wide, it can be easy to forget that some people don’t use Facebook. While previous work has compared users and non-users of social networking sites, this study is one of the first to give a sense for the prevalence of non-use.

Results suggest that non-use is fairly common. Of 410 questionnaire respondents, almost one third of respondents had left Facebook by deactivating or deleting their account. Deactivation hides everything a user has done on Facebook but retains the data and allows a user to reactivate at any time. Deletion permanently removes all the user’s data from Facebook’s servers.

Those who left reported being happy with their decision to leave. Of those who deactivated, two thirds reported being “somewhat happy” or “very happy” with their decision. Those who deleted felt even better about the decision, with over 90% reporting being “somewhat happy” or “very happy” about leaving FB.

Respondents reported being happy with their decision to deactivate (left) or delete (right) their account.
Respondents reported being happy with their decision to deactivate (left) or delete (right) their account.

However, over half of deactivators subsequently returned to Facebook, many of whom had intentionally deactivating for short periods of time. Not so with deleters, 85% of whom did not return.

Proportions of participants who showed various types of use and non-use.
Proportions of participants who showed various types of use and non-use.

Analyzing stories that participants told about their own or others’ experiences around leaving or limiting use of Facebook, the authors identified six motivations for various forms of non-use.

  • Privacy: concerns about which other users have access to personal information.
  • Data use and misuse: concerns about what Facebook as a corporation was doing with personal data, for example, monetizing or profiting from it.
  • Banality: the Facebook interactions are at times described as shallow or meaningless, especially in contrast to face-to-face or other media.
  • Productivity: use of Facebook negatively impacting productivity, either at work or at school.
  • Addiction: feelings of withdrawal, idolizing non-users, and other experiences consistent with descriptions of addiction.
  • External pressures: being friended by a boss, a student, or an ex romantic partner; leaving because of being, for example, a military officer or a parolee.

From Luddites smashing looms to Amish horse-drawn buggies, history provides numerous examples of technology resisters. Studying these historical moments helps understand the unfolding of larger conversations and negotiations about the role of technology in our daily lives.

For more details, please see the full paper.

Eric P. S. Baumer, Cornell University
Phil Adams, Cornell University
Vera Khovanskaya, Cornell University
Tony Liao, Cornell University
Madeline Smith, Northwestern University
Victoria Schwanda Sosik, Cornell University
Kaiton Williams, Cornell University

All the news that’s fit to read: which social annotations matter

Despite the ubiquity of social annotations online, little is known about their effects on readers, and relative effectiveness.
Despite the ubiquity of social annotations online, little is known about their effects on readers, and relative effectiveness.

Every day, we are influenced by people around us. Increasingly, in addition to receiving news via broadcast media, our friends tell us about news items both on social media and through word of mouth.  The provenance of news items is important.  We use these provenance to determine how interesting a news item is, how much attention we should pay, and how much we should trust it.  As online news reading becomes more social, news websites and applications today show articles recommended by friends, strangers, and computer algorithms: How do people respond to these annotations differently?

To answer this question, we conducted two experiments. The first experiment simulates a user’s logged-out experience with annotations from strangers, a computer agent, and a branded company.  This experiment is intended to shed light on how users respond to annotations when the service provider do not know anything about the user (such as when no social network information is available)

560 participants from Mechanical Turk looked at real news items with the above three types of annotations.  The participants then decided which articles to read based on the headline and these annotations.  Comparing with a control condition with no annotations allowed us to understand the influence of these annotations.  Unsurprisingly, participants clicked just as many news headlines with strangers’ annotations than those with no annotations. Yet, participants clicked on those headlines with annotations from unknown companies (we made up the name of the company) more frequently than the baseline. This suggests strangers’ annotations aren’t persuasive, but those by companies are.

The second experiment simulates an experience with personalized content and annotations from friends (we show articles that a user’s friends have publicly +1’d on Google).  Compared to the baseline (no annotation),  friend annotations are persuasive. Participants clicked more news articles with such annotations, and also rated them as more interesting (than similar articles with no annotation). In contrast, articles that strangers annotated were rated lower than similar articles with no annotation.

Post-experiment interviews indicated possible explanations– friends above a threshold of closeness had known expertise and shared context, for instance. Strangers, on the other hand, lacked these qualities.

Who influences what we read online? Not strangers, but existing friends.  Not only do friends persuade us to read articles that they recommend, they also convince us that these articles are interesting. Surprisingly, branded companies also have a persuasive effect.

For more, see our full paper, All the News that’s Fit to Read: A Study of Social Annotations for News Reading, at CHI 2013. (If you’re attending CHI, this is in the Social Tagging session, Wed 11am).

Chinmay Kulkarni, Stanford University
Ed Chi, Google, Inc.

Instagram at the museum: Communicating the museum experience through social photo sharing

How do people use mobile phones and social media to document and share their museum experience? That question is the foundation of our paper which will be presented at SIGCHI on Tuesday.

nomethod story
An example of a visitor-created collection of instagrams referred to as by this instagram user (@nomethod / @petergaudiano) as a “chamber of horrors”

The notion of social photo sharing is not new. For decades, people have been taking photographs and sharing them with each other when having friends over to watch slides from their latest vacation, at family gatherings, in school yards, etc. In 1987 Richard Chalfen coined the term ‘Kodak culture’ when describing how people take, organise and show their photographs to family and friends. As modern mobile technologies such as smartphones with high quality built-in cameras and access to social media networks have offered new possibilities for people to engage, participate, document and share their experiences, new photography practices have emerged.

In our paper, we focus on one such ongoing creation and sharing practice where one particular photo sharing application, Instagram, is used to communicate visitors’ experiences while visiting a museum of natural history. Using Instagram, visitors have the opportunity to communicate their experiences through both choice of photo subject and ways they choose to manipulate and present them.

Based on an analysis of 222 instagrams created at the museum, as well as interviews with 16 of the visitors who created them, we examine the compositional resources and concerns that shape the creation of the visitor-driven content, organisation and sharing of multi-layered aesthetic presentations of images and texts. This analysis is focused on the process of creating and sharing instagrams including photographic choices of various kinds, the formulation of captions and hashtags, and the role of the social media audience.

We found several interesting examples of ways that visitors re-categorise and document the objects found museums to share their experiences and explore them to unpack the ways instagrammers draw upon features within the environment, the mobile application, and the social network in their work to construct their own narratives from museum exhibits.

For more information, please visit the project website.

Alexandra Weilenmann, University of Gothenburg

Thomas Hillman, University of Gothenburg

Beata Jungselius, University of Gothenburg

Collaboration in the Remixing and Reuse of Open Educational Resources

Education systems around the world are under immense pressures and undergoing rapid change. Open approaches could be an important means to increase access and reduce the costs of high quality learning experiences.

But how do those working in education utilise openness to collaborate across organisations? How are the motivations, practices, and support needs different from those seen in open source software or Wikipedia?

Remixing and Reusing Open Educational Resourses
Remixing and Reusing Open Educational Resources

This project studied the Bridge to Success initiative, which aimed to harness openness to improve US college achievement. Open Educational Resources (OER) were developed and made available for any organisation or learner to use as they wish. These are free online course materials in study skills and mathematics, combining tutorials, activities, quizzes and multimedia objects.

Rather than creating wholly new materials, or just sharing existing ones, the initiative went a step further by Remixing OER for a new target audience. Proven materials from courses in the UK were converted for those entering, or struggling with, college in the US. This involved extensive collaboration across organisations and continents.

Less than a year later, these resources have been used by thousands of learners and over 16 US-based organisations. Through studying the initiative, we identify and explore four emerging practices in interactions with OER:

Creative Remixing: Remixers build on the existing OER, creating additional material where they feel it is required for their audience. They are inspired by the values of the original text but also create entirely new additions where they feel this is useful.

Adaptive Remixing: This lower cost approach involves only the effort required to make the materials appropriate, such as editing language and replacing culturally-specific examples.

Contextualisation: Organisations interested in using an OER explore how it might best fit into their specific context. This might include introducing students to the materials to fill the gap between signing up for a course and the beginning of term, or using data to target resources towards struggling students.

Wrapping: Individuals then do the work to integrate resources into their teaching. This may involve selecting relevant parts of the OER, integrating with other resources, providing specific instructions and working across face to face and online learning situations.

Our paper explores a range of issues where current views of ‘open’ in should be expanded through understanding these practices and their support needs. OER empowered individuals to share and creatively tackle issues faced in their work. Through this, new forms of ownership become possible, and tools and processes are shared outwards from the organisations involved.

For more, see our full paper, Building Open Bridges: Collaborative Remixing and Reuse of Open Educational Resources across Organisations.

Tim Coughlan, University of Nottingham
Rebecca Pitt, The Open University UK
Patrick McAndrew, The Open University UK

Close the Door: Crowdsourcing and Community Activism

“Close the door, were you born in a barn?!”  The exhortation to keep doors shut has been passed down from generation to generation, and serves the practical purpose of keeping dirt out and heat in.  Yet there is one place where this adage tends to be ignored: the British High Street.  Shops often leave their doors open regardless of the weather to tempt customers over the threshold, wasting up to 10 tonnes of CO2 a year and creating uncomfortable working conditions for staff.

Close the Door Decal

Enter the Close the Door campaign, a grassroots community activism group.  We developed an iPhone app to assist the campaign in monitoring the behaviour of local shops, and also tested different motivational factors:

  • Control: Users simply recorded whether a shop door was open or closed; intrinsic motivation (i.e. support for an environmental campaign) would be at play.
  • “Gamification”: Points, badges, and a leaderboard were used as extrinsic, virtual motivators.
  • Financial Incentives: Same as above, but participants were also paid based on the number of badges earned and the points gained.


The Financial group collected more data than either of the other groups, to a statistically significant level.  However, a subsequent survey showed levels of intrinsic motivation decreased: only 75% of those in the Financial group would be willing to use the app without payment, compared to 100% of both the Control and Gamified group.


Shops Added















* Points/badges earned were not visible to members of the Control group.

Interviews with participants highlighted the following aspects that need to be taken into account when designing community-based crowdsourcing applications:

  • Be aware of different threshold and engagement motivators: What motivates someone to “step over the threshold” and download a crowdsourcing application can vary greatly from what motivates them to remain engaged with the app.  In this instance, a pro-environmental attitude was not necessarily enough to encourage pro-environmental behaviour (i.e. using the app).
  • Users must be enabled to use the app: Being motivated was only half the battle: users also had to have time to use the app.  Those who were time rich, e.g. not in regular employment, tended to use the app far more.
  • Competition is not created equal: Similar trends in both the Gamified and Financial group showed that close competition occurred among the top 25% of users in each group, with these users spurring each other on.  Distant competition, however, occurred when users were hundreds or thousands of points behind; this served to demotivate participants.  Other participants reported “self gamification”, i.e. competing with him or herself, which deserves further research as a motivational factor.
  • Aim to be average: Some participants mentioned wanting to be sure they were on the leaderboard, not to be the best, but rather to show they “did their bit”.  The benefits offered by “normification”—harnessing the power of social norms—are also worth pursuing further.


At the end of the two-week trial, it was possible for the Close the Door campaign to tell at a glance which shops had been added to the master database and which were keeping their doors closed (green) or not (red).

For more, see our full paper, Using Crowdsourcing to Support Pro-Environmental Community Activism.
Elaine Massung, University of Bristol
David Coyle, University of Bristol
Kirsten Cater, University of Bristol
Marc Jay, University of Bristol
Chris Preist, University of Bristol