The topic of research ethics is enormous to examine. There are books focused solely on studying and discussing research ethics (even before the era of big data, network analysis, simulation models and large-scale online experiments). This suggests that the topic is problematic and challenging for academics but also important. Our scholarly aims are to study how societies and individuals behave and describe these in some manner (and, for some of us, criticise these behaviours and propose alternatives). We are also concerned that we do not cause (extensive) harm to the humans we are studying. That is, we do not want our research or its results to lead to negative consequences for our participants or the wider society. These are not new concerns. These principles were established long before computational social sciences in any of its forms was used as a research methods.
The research ethics discussion is not separate from research practice. Rather, it needs to follow what scholars are doing. Sometimes the discussion takes place after ethics violations, showing a reflection of the practices. For example, medical research for subjects without consent was banned in the Nuremberg Code, following the World War II trials and acknowledgement of bad practices for medical research. Similarly, in social psychology, famous studies like the Stanford Prison Experiment (Zimbardo et al., 1971) and Obedience to Authority (Milgram, 1963) raised the question of research ethics in a non-social-science context, suggesting that informed consent might not be sufficient to ensure the ethical conduct of research.
The most recent debate related to an emotional contagion study, where it was shown that users become sad when they see negative valence social media posts.9.1(However, companies like Facebook run these kinds of studies on a regular basis to improve their services. The difference was that the academic publication was more open than experiments conducted in corporate settings. This observation was highlighted by, for example, Meyer, 2014, .) Scholarly and media debates highlighted many ethical challenges, such as potential for harm for experimental participants (Shaw, 2016), institutional arrangements giving researchers access to these kinds of data (Kahn et al., 2014; Flick, 2016), differences between industry research and academic practices regarding ethics (Kahn et al., 2014; Flick, 2016), a participant pool potentially including minors or vulnerable persons (Shaw, 2016) and a lack of an explicit-enough consent from study participants (Flick, 2016; Kahn et al., 2014; Shaw, 2016), including the question of who the participants are if the experiment is run on user-generated content (Selinger and Hartzog, 2016). Boyd (2016) moves this discussion forward by suggesting that public upset was not related to research ethics as such but rather on the practices and business models of big data. She suggests that
scholars and the public latched onto this study to channel their broader anger [- as they] begin to understand the manipulations possible through `big data.' [- -] Because of this, any effort to address the ethical issues introduced by this study requires moving beyond critiquing Facebookâs data science research or questioning the particulars of this study and reconsidering how to hold public companies accountable for the decisions they choose to make on behalf of their consumers. What is at stake is the underlying dynamic of how Facebook and other major social media sites run their businesses, operates their systems, and make decisions that have nothing to do with how its users want those companies to operate.Whether or not we seek to expand the case to discuss corporate ethics, it suggests that computational social scientists with their novel methods and data sets must consider research ethics as well - wherever they end up working.
To engage in research ethics, we must understand what ethics are. Ethics relate to making a normative judgement about what is good and bad about potential decisions. There are three different schools of normative ethics, focusing on different aspects of making these evaluations. Briefly elaborated, consequentialist ethics focus on the outcomes of oneâs actions, urging people to make decisions that lead to a better world. Duty-based ethics or deontology suggest that each of us have duties and rights and that we make the right choice based on balancing these duties and rights. Virtue ethics highlight that there are ideals (or virtues) that tell us how to behave in a situation and suggest we ought to follow these examples, no matter the outcomes or our duties and rights in a situation. Naturally, we use and mix these approaches when we make our decisions in life or research. Therefore, one rarely considers which school of normative ethics are used. However, as suggested by (Salganik, 2017), reflecting on cases using these schools may be helpful and give insights into different positions in computational social sciences. Approaching an ethical question from different schools may lead to different final verdicts, thus breaking the case and argumentation. This can help in some questions or debates on research ethics as well.
I have not yet addressed the question: `Why should I follow ethical principles?' Extreme violations of research ethics may lead to institutional punishments, such as loss of funding or even employment. However, this is not my main motivation to follow ethical principles in my research. I think that I am a good person (suggesting duty-based ethics) and believe that ideal researchers should ensure that their research is standing on solid ethical ground (indicating I also have some virtue ethics). Furthermore, if my research actions are published on the front page of the New York Times or Guardian (unlikely, but it has a non-zero probability), I would need to explain myself to my family and friends. In this discussion, I really want to ensure that they do not raise their eyebrows, as this could harm my relationships with them (indicating a consequentialist ethics approach). Even more important motivation for me on research ethics is that we are all in this together. For computational social sciences to thrive, like all academics, we need resources such as data and funding. Public support for research may be helpful, as is support from other academics. It is helpful if our work is not seen as harmful to the participants or wider society. I also believe that these concerns are real - not imaginary. For example, in 2018 Facebook closed many of its application programming interfaces that allowed academics to gather data. This seemed to be in part a reaction to the Cambridge Analytica scandal, where data were pulled from Facebook to target voters. This malpractice also limited scholarly access to an important source of social media data (for further discussion, see Puschmann, 2019; Bruns, 2019). If one screws up in research ethics, the larger community might bear the costs of it.