Andreas Weigend | Social Data Revolution | Fall 2014
School of Information | University of California at Berkeley | INFO 290A-03

Homework 2: Facebook Video

Assigned: Tue Oct 7, 2014
Due: Mon Oct 20, 2014 at noon
Peer comments due: Fri Oct 24, 2014 at noon

Recently, Facebook has been receiving some criticisms over revelations that it has been secretly conducting various experiments on its users. These experiments deliberately attempted to influence the emotions of users positively or negatively by showing more positive or negative news feed items to them. The experiment results suggested that the user’s mood was affected by the chosen newsfeed items. Some people see the experiment as being manipulative as it was conducted without the users’ prior knowledge.

However, many other companies conduct A/B testing. Google, the company that popularized the technique, is sometimes ridiculed for its extensive A/B testing. Infamously, it tested different shades of blue to find the perfect color for its advertising links.

The line between a company doing A/B testing to improve a product and presenting different information to a user to see if it provokes user reaction is a thin one. Facebook COO Sheryl Sandberg said in an interview that the controversial study was "part of ongoing research companies do to test different products" and was "poorly communicated," suggesting that the problem with the experiment was not so much its nature but the way it was portrayed to the public. Indeed, Facebook has conducted other media-grabbing experiments on voting, organ donation, and the diversity of information seen via the platform.

The general public also seems to have a limited understanding of the product that Facebook provides, according to research conducted by Karrie Karahalios of the University of Illinois and her colleagues. A staggering 62.5% of the subjects of this study reported they were unaware or uncertain that Facebook filtered newsfeed information using algorithms. Many said they were surprised to see what information was included in their feed and what was not, with some expressing interest in having more direct control over what they see. To conduct the study, the researchers developed an auditing tool they called FeedVis, which allowed subjects to compare their Facebook feed to the posts created by all of their friends, making Facebook's product more visible to the users. Karahalios has argued for greater transparency around data products like Facebook, possibly through a similar auditing tool.

Has the public, or at least the publicized, perception of Facebook become overly critical? Were Facebook’s experiments really substantially different from A/B tests performed by other tech companies? To what extent do tech companies need to communicate how their services work to users? Do Facebook's users understand what they give and what they get by participating on the platform? Should there be some form of algorithmic auditing for data products? What categories of data products might be audited? In order to think about these questions, we are interested in first finding out more about the public perception of Facebook’s actions and its services in general.

Relevant Readings


Task


Your task for this homework is to partner with another classmate to make a short video consisting of interviews with friends/family/strangers about their understanding of how Facebook makes recommendations of information to its users. The video should be between three and ten minutes long and primarily consist of footage with multiple interviewees.

The goal of the interviews is to find out how people are using Facebook and how they perceive the service. For interviewees who are familiar with the Facebook experiments being discussed, some time should be spent discussing their thoughts on the topic.

Here is a list of some possible questions to ask. These are just initial suggestions meant to trigger your own questions.
  1. What do people expect to see on their Facebook newsfeed?
  2. What do people want Facebook to show them? Should Facebook filter content for the user?
  3. What do they think Facebook knows about them? What do they allow Facebook to know? What do they like Facebook to know? What do they not like Facebook to know?
  4. How valuable to they think that the information they provide to Facebook is?
  5. When do they feel they get an appropriate value in exchange for the information they provide?
  6. What types of experiments would they like to see Facebook conduct? Should any types of experiments be off limits?

Video Requirements


The video should be at least 3 minutes long and consist of selected interview answers that you think provide insight into how people understand Facebook as a service, and how Facebook could better communicate its services to users.

The video should be horizontally oriented and have at least 1280x720 resolution.

Please be aware that your videos may be incorporated into public material for the class and posted to sites such as YouTube. You should inform your interviewees of this. Also, please provide basic biographical information for each interviewee (e.g. what’s their profession/major?) to give some context to the answers.

You can use simple and free video editors such as Windows Movie Maker or iMovie that came with your computers to edit the video.
You should ensure that your videos are of reasonable quality. In particular, the audio must be clear.

Grading Scheme


Entertainment value
35%
Level of insight
35%
Technical quality
20%
Peer comment
10%
Technical quality is determined mostly by audio quality. Please ensure that the interviews are clear and audible. There is also a 10% component that comes from commenting on videos submitted by other groups.

The remaining 70% of the grade comes from the content of your video. This will be judged partially based on the comments left by your fellow students.

Submission


Please post your videos to the Social Data Revolution page on Facebook at
https://www.facebook.com/socialdatarevolution

At least one student per pair should have a Facebook account. You are encouraged to comment on and like the videos submitted by other teams as a portion of your grade will come from the comments you leave on the Facebook page.



Direct any questions to Raven Jiang (raven at cs dot stanford dot edu)