Research that blends social science and computational approaches is a diverse and growing discipline. The rapid evolution of technological advances has made data more pervasive and accessible than ever before. With this expansion, many tensions have arisen within the field.
Fairness, transparency, and accountability as it relates to the collection, manipulation, and assessment of the data are ongoing discussions, but often pose challenges for established mechanisms for ethical review. For example, in the United States, Institutional Review Boards (IRBs) regulate human subjects research--which the collection and analysis of publicly available data (e.g., tweets) is often not considered to be [7,9]. Therefore, computational research or even social science research that examines trace data (like much of the research published at ICWSM) does not have the clear-cut ethics guidelines (e.g., around consent) as some other types of research. However, trace data like tweets are still created by people, and represent both individuals and communities.
Often at the core of these tensions is protecting the people or communities in the data from identification or harm [6]. Recent scholarship in big data and research ethics has highlighted the diversity of domains where these tensions have increasing visibility. For example, the tension between individual privacy and usage of public data highlight complexities like the often ambiguous boundary between public data [8], reverse identification of data [1], identification of people [2], and data labeling (e.g., identifying the sexuality or mental health of a poster) [5,8]. Underlying this tension is the potential misalignment between the perception of the people who create the data and those that use it for secondary analyses [2]. Additionally, the tension between privacy protections and open science is growing as the nature of the data collected for research is increasingly more personal while fields are simultaneously promoting and expecting datasets to be published in open repositories [3].
Based on these considerations, when is it appropriate to contemplate these or other tensions or challenges in social media research (e.g., apriori to the research or post hoc case studies)? Are there topical domains where these considerations are more prevalent (e.g., for stigmatized issues like mental health)? How are researchers handling these tensions? This workshop aims to convene a dynamic and diverse subset of the ICWSM community to discuss their data collection, analysis, assessment, and reporting practices related to these tensions.
REFERENCES
[1] Ayers, J.W., Caputi, T.L., Nebeker, C. et al. Don’t quote me: reverse identification of research participants in social media studies. npj Digital Med 1, 30 (2018). https://doi.org/10.1038/s41746-018-0036-2
[2] Casey Fiesler and Nicholas Proferes. 2018. “Participant” Perceptions of Twitter Research Ethics. Soc. Media Soc. 4, 1 (January 2018), 2056305118763366.
[2] Dennis, S., Garrett, P., Yim, H. et al. Privacy versus open science. Behav Res 51, 1839–1848 (2019). https://doi.org/10.3758/s13428-019-01259-5
[3] Hauge, MV, Stevenson, MD, Rossmo, DK (2016) Tagging Banksy: Using geographic profiling to investigate a modern art mystery. Journal of Spatial Science 61(6): 185–190.
[5] Sam Levin. 2017. New AI can work out whether you’re gay or straight from a photograph. The Guardian.
[6] Jacob Metcalf and Kate Crawford. 2016. Where are human subjects in big data research? The emerging ethics divide. Big Data Soc. 3, 1 (2016), 1–14.
[7] Jacob Metcalf. 2017. “The study has been approved by the IRB”: Gayface AI, research hype and the pervasive data ethics gap.
[8] Anja Thieme, Danielle Belgrave, Aane Sano, Gavin Doherty. 2019. Reflections on mental health assessment and ethics for machine learning applications. ACM Interactions.
[9] Jessica Vitak, Nicholas Proferes, Katie Shilton, and Zahra Ashktorab. 2017. Ethics Regulation in Social Computing Research: Examining the Role of Institutional Review Boards. J. Empir. Res. Hum. Res. Ethics (August 2017), 1556264617725200.
[10] Michael Zimmer. OkCupid Study Reveals the Perils of Big-Data Science. WIRED. Retrieved April 23, 2018