Google “ethics and statistics” (318 million) or “ethics and data science” (14 million) and you will find millions of results. The Institute of Electrical and Electronics Engineers (IEEE) published a 2012 description of the computer science undergraduate curriculum; the American Statistical Association (ASA) published 2014 guidelines on undergraduate statistics curricula and non-statistics majors’ statistics courses; and the National Academy of Sciences, Engineering, and Medicine published a 2018 report outlining curriculum features for the undergraduate major in “data science.” All these recommendations feature “ethics.” None discuss what it means, looks like, or should result in if a program or instructor does feature or teach ethics.
It is clearly an important topic, but it is also hard to conceptualize. That makes it challenging to teach and even harder to test, assess, or grade. The millions of search results might make the problem seem overwhelming.
One approach is to use “responsible conduct of research” materials, which are plentiful. There are clearinghouses and books about this topic in the US because it is the title of two major funders’ (National Science Foundation and National Institutes of Health) perspectives on “ethics” for science. However, not everyone who will learn or use statistics and data science is engaged in science or research, for example, business and government applications.
Others have created checklists intended to provoke conversations about whether ethical decisions were made during a project. Unfortunately, these checklists do not offer guidance about how to ensure everything up to completing the checklist was ethical, nor is there information about how to identify if something was not ethical, nor what to do if unethical decisions are discovered.
Apart from these checklists, which are discipline specific but not informative, a challenge for domains such as statistics, mathematics, computer science, and engineering is that a great deal of material and instructional orientation used to teach “ethics” or “responsible conduct of research” (RCR) is actually derived from the more clinical or biomedical domains. As such, in institutions with a single “RCR training” course for all incoming graduate students (for example), a great deal of time is focused on how to secure consent for/safeguard, handle, or work with human and animal research participants.
Instead, a different paradigm can be pursued than what is typical for STEM ethics education: the development of ethical reasoning throughout a program or career.
Ethical Reasoning (ER) comprises the following six knowledge, skills, and abilities (KSAs) that can be applied across the most simple cases:
- Identify and ‘quantify’ your prerequisite knowledge. Professional practice standards, local laws, and stakeholder analysis are important knowledge needed for reasoning ethically. This is a necessary KSA for ethical reasoning with any source document—ethical practice standard, federal data guidelines, or workplace policy.
- Identify decision-making frameworks. “What would the ethical practitioner do in this case?” “How can benefits be maximized while harms are minimized?” Although these represent the virtue and utilitarian perspectives, respectively, in-depth understanding of ethical decision-making is not as important in the ER process as recognizing which of these two perspectives is most relevant to the case, situation, or task at hand.
- Identify or recognize the ethical issue. Does something about a case, situation, or task represent an undue imbalance in the harms and benefits identified in the stakeholder analysis? Or, if such guidance is available, is there something about a case, situation, or task that seems inconsistent with that guidance? Note that consideration of harms/benefits (i.e., stakeholder analysis) works as well as ethical practice standards to teach and give practice with ER.
- Identify and evaluate alternative actions (on the ethical issue). Just like real life, every ethics case analysis requires a decision be made. To ensure viable options are considered, at least two plausible alternative options (decisions, actions) are always available to be identified and evaluated. These actions always include, “do nothing/ignore the issue identified” and “do something about the issue identified.” Even without a formal process for dealing with identified ethical problems, learners can evaluate these alternatives, either in the stakeholder analysis (i.e., comparing harms to benefits in terms of their severity and/or to whom they accrue) or by appeal to ethical practice standards.
- Make and justify a decision. What to do in the face of the ethical challenge that was identified—including at least some discussion of how stakeholder effects were considered. An essential part of the ER normalization process, notifying practitioners.
- Reflect on the decision. What makes this case “hard”? What additional information would be/would have been helpful? How can you get better at these challenging features of ethical reasoning? How does a case/analysis like this one help create the culture that promotes fluency in ethical reasoning and/or a more ethical workplace?
Ethical challenges can arise in new and wholly unexpected situations throughout a career. It is not plausible to assume scientists, and those who use science or its results at work, will somehow prepare themselves for specific unknowable eventualities. However, learning how to reason ethically—which includes identifying ethical challenges given then-current specific ethical practice standards or policies—can serve to prepare anyone to identify and appropriately respond to currently unknowable ethical challenges.
ER is content independent, but can be deployed with ethical practice standards (e.g., Association of Computing Machinery (2018) Code of Ethics and American Statistical Association (2022) Ethical Guidelines for Statistical Practice) or workplace policy (e.g., Principles and Practices for a Federal Statistical Agency, 2021).
The figure shows the foundational role ethical reasoning plays in the development and support of professionalism.
