ICA San Diego 2017 - The Social Shaping of DTC Genetic Testing: Sentiment Analysis of 23andMe on Twitter

Scholars, educators, regulators, pundits and other observers are advocating for regulation and oversight of ‘direct-to-consumer’ genomic testing (DTC). As a result, the technology has been subject of much public and regulatory controversy. In this article, we explore the sentiment about the DTC company, 23andMe. To describe the nature of the public opinion, we collected Twitter data of over 2000 tweets for a quantitative content analysis and qualitative framing analysis. Our analysis reviews particular frames that seem to underlie social media exchanges related to the company and its technology. Our study shows that people are largely positive towards DTC genomics and 23andMe. We argue that these frames create meaning and play a role in how a technology is interpreted by its users. As social research becomes increasingly driven by large data sets and internet-based research methods, we offer a timely, non-invasive analysis of emerging values associated with DTC genetic testing.

Transcript

1. The social shaping of DTC genetic testing Sentiment analysis of 23andMe on Twitter Alberto Lusoli; Peter Chow-White; Frederik Lesage; Stephan Struve Simon Fraser University. School of Communication. GeNA Lab ICA 2017, May 27th, San Diego, USA
2. Image credits: 23andMe Offic
3. 3 ALBERTO LUSOLI – ICA 2017 SAN DIEGO 0 20 40 60 80 100 120 "23andMe" Search volume Market Agreement Source: Google Trends .
4. Objective Explore the discourses around 23andMe. Understand how this debate about personalized genomics is influencing the development of the technology
5. Theoretical approach Social Construction Of Technology (SCOT) (Bijker, Hughes, & Pinch, 1984)
6. Because it allows unmediated and Non-invasive access to the discourse developing around 23andMe (Cheng, Fleischmann, Wang, Ishita, & Oard, 2010; Fleischmann, Oard, D.W., Cheng, A.S., Wang, & Ishita, 2009; Morris, 1994; Weber, 1990)
7. 7 ALBERTO LUSOLI – ICA 2017 SAN DIEGO RQ1 What are the user sentiments about 23andMe? Are they mainly positive or negative? RQ2 What are the frames that underlie the positive and negative sentiments about 23andMe?
8. 8 ALBERTO LUSOLI – ICA 2017 SAN DIEGO DATA COLLECTION (NCAPTURE+NODExl) DATA CODING (POSITIVE/NEGATIVE) FRAME IDENTIFICATION Unit of analysis The single tweet and linked webpages Data source 2.027 tweets containing the term “23andMe” collected during a 1-week period (Jan.2014). research protocol CONTENT ANALYSIS OF TWEETS (BERG, 2001)
9. 9 ALBERTO LUSOLI – ICA 2017 SAN DIEGO Positive 86% Negative 14% results SENTIMENT ANALYSIS
10. 10 ALBERTO LUSOLI – ICA 2017 SAN DIEGO results FRAME ANALYSIS
11. 11 ALBERTO LUSOLI – ICA 2017 SAN DIEGO results FRAME ANALYSIS Unqualified excitement about the potential of DTC genomic tests Happy to share their results publicly.
12. 12 ALBERTO LUSOLI – ICA 2017 SAN DIEGO results FRAME ANALYSIS Tweets referring to the concepts of family, roots and belonging. Positive attitudes toward the possibility to rediscover the past and reconnect with relatives.
13. 13 ALBERTO LUSOLI – ICA 2017 SAN DIEGO results FRAME ANALYSIS All those tweets which express the possibility to change behaviors on the basis of the test results.
14. 14 ALBERTO LUSOLI – ICA 2017 SAN DIEGO results FRAME ANALYSIS All the tweets which identified as 23andMe as a disruptive innovation, as opposed to the paternalistic and conservative power of the FDA.
15. 15 ALBERTO LUSOLI – ICA 2017 SAN DIEGO results FRAME ANALYSIS Critical frame that questioned the scientific accuracy of the test. Tweets often expressed doubt and brought scientific research as counterarguments to the supposed validity of the 23andMe test.
16. 16 ALBERTO LUSOLI – ICA 2017 SAN DIEGO results FRAME ANALYSIS All the tweets which criticized the actual possibility for customers to extract meaningful information from the data without proper background knowledge.
17. 17 ALBERTO LUSOLI – ICA 2017 SAN DIEGO results FRAME ANALYSIS This frame includes tweets that discuss the negative emotions involved in the unmediated access to health data (e.g. anxiety, stress) (Su, 2013).
18. 18 ALBERTO LUSOLI – ICA 2017 SAN DIEGO 0 20 40 60 80 100 120 Maket Agreement Source: Google Trends . conclusions and next steps
19. A special thanks to everybody who contributed to this project:  Stephan Struve  Frederik Lesage  Peter Chow-White  Nilesh Saraf  Amanda Oldring  Eliot Tran  Lucas Wu THANK YOU For more information: alusoli@sfu.ca twitter.com/albertolusoli genalab.org Photo Credits: All pictures used in this presentation are protected by Creative Common License except where otherwise indicated.
20. Next steps 1. Assess to what extent marketing and promotional strategies on social media can influence the closure process 2. Understand the role that “neutral educational Tweets” have in support of positive/negative sentiments Image credits: Momondo Official Website