Successfully reported this slideshow.
We use your LinkedIn profile and activity data to personalize ads and to show you more relevant ads. You can change your ad preferences anytime.

Sampling and recruiting on Facebook

307 views

Published on

With the rise of new media and social media, a new era of big data has emerged, which has brought about various methodological and theoretical challenges for conducting social research. With over a billion daily active users, Facebook is widely recognized as the leading social media platform in the world. Beyond the use of Facebook to connect people from around the world, Facebook affords various opportunities for academics to conduct research. In this presentation, we will discuss our approach to integrate Facebook data as part of an online survey to study people’s privacy concerns, with a particular focus on methodological challenges associated with sampling and recruiting participants on Facebook.

Sampling: Considering, there is no easy searchable directory of Facebook Groups or Pages, how do researchers identify and sample Facebook Groups or Pages? Problematically, it is difficult to systematically sample across Facebook to get a “representative” sample of Facebook users. Facebook Group Directory, algorithmically-filtered search, and custom-curated directories can be used to sample; however, each approach introduces biases and challenges.

Recruitment: After the selection of the Group/Page of study, how can researchers invite people to participate in the study? Facebook’s Terms of Service does not allow contacting users directly unless you have conducted “business” with them. We outline the various options for recruitment, including buying an ad, posting directly to the group/page, and contacting the moderator.

Ethics: As more Canadians are joining and contributing to Facebook, their automatically recorded data are rapidly becoming available to third parties to mine for both commercial and academic purposes. Ethical questions need to be considered throughout the entire research process. This is particularly true of social media research, which presents unique ethical and personal considerations. In this part of the presentation, we will outline the Social API Terms of Service online guide created by the Social Media Lab at Ryerson University that Internet researchers can use to learn what they can or cannot do with social media collected from sites like Facebook.

The presentation will conclude with the discussion of our plans to develop a unique Facebook app that will allow any Facebook user to review their publicly available social media data and participate in our survey on Social Media Privacy Concerns.

Published in: Social Media
  • Be the first to like this

Sampling and recruiting on Facebook

  1. 1. Canadian Communication Association Annual Conference May 30, 2017 Sampling and Recruiting on Facebook Anatoliy Gruzd @gruzd Jenna Jacobson @jacobsonjenna
  2. 2. Outline Study Objectives Ad Creation Pilot Results 3.1. 4. Sampling & Recruiting 2.
  3. 3. data consumers Survey Facebook users about privacy concerns using their own data 1. Study Objectives Project website: http://socialmediadata.org/
  4. 4. 2. Sampling Strategies Source Description Facebook Public Directory Direct list of Facebook Pages Facebook User-Based List Personalized recommendations on Groups and Pages by Facebook Netvizz/API Extracts data from Groups and Pages for research purposes Socialbaker Social media marketing /analytics company that compiles list of top Pages, grouped by industry Official Lists Official external lists based on topic (e.g. size, demographics, politics) Curated/Expert Lists Unofficial external lists (e.g. by media outlets or blogs) Researcher-Created List Ethnographic approach to collect all groups that fit topic of interest
  5. 5. • Facebook’s advertising policies 3. Ad Creation https://www.facebook.com/business/
  6. 6. Design an Ad Banner Properties to consider  Attention grabbing & informative
  7. 7. Design an Ad Banner Properties to consider  Diversity of gender, race, and age
  8. 8. Design an Ad Banner Properties to consider  Generic vs. Group/Page specific
  9. 9. 4. Pilot Ads Ad #1: Corporate Page Super 8 Motels in Canada Ad #2: Community Group “Pokémon GO” in Winnipeg • 2 days (Oct 8-9 ‘16) • $46 spent • Reached ~4K people • 67 clicks (~$0.69 per click) • No completed responses • 3 days (Aug 28-31 ’16) • $50 spent • Reached ~5K people • 169 clicks (~$0.30 per click) • Only 3 partial survey responses Most were mobile users: - Survey was too long (20min) - Some didn’t remember their sign-on info to login via a mobile browser (required by our survey tool)
  10. 10. 4. Pilot Ads - Demographics
  11. 11. Recommendations 1. Strategize design of ads (informative, attention grabbing, intended audience) 2. Remember most will be mobile users • Keep survey as short as possible 3. Recognize survey participants are likely lurkers, rather than active participants 4. Assess your budget (clicks ≠ completion) 5. Offer a prize or another incentive to complete survey 6. Ask group moderator(s) permission to post ads
  12. 12. #Congress2017: Ryerson Social Media LabTour At the Social Media Lab, we study: how social media is changing the ways in which people communicate, disseminate information, conduct business and form communities, and how these changes impact the social, economic and political structures of modern society. http://socialmedialab.ca
  13. 13. Canadian Communication Association Annual Conference May 30, 2017 Sampling and Recruiting on Facebook Anatoliy Gruzd @gruzd Jenna Jacobson @jacobsonjenna

×