KENNESAW, Ga. | Jul 15, 2022
I’ve written about and given much thought to ethics and its relationship to analytics and data science. I even worked with O’Reilly Media to compile a book on the topic in 2020. However, ethics remains a concern today; one reason is that people in our community are not being trained properly to ensure that everything about their analytics and data science processes is ethical.
This challenge stems in large part because (as I outlined here) what is ethical and what is not is much more subjective and context-dependent than people would like. To address that problem, intentionality is required, which includes intentional focus on effective training and education around the topic of ethics.
Most STEM-based programs focus primarily on teaching specific topics and techniques and then testing students on them. Certainly, that type of format is an important one for learning the foundational concepts. However, there is room to add other formats of instruction to bolster the learning from traditional classes. One alternate format is the case study approach used in many law and business schools. A case study format isn’t as much about learning new concepts as it is about learning how to apply the concepts previously learned.
Case studies are great tools to explore and understand how the theories being learned in a classroom setting have been applied in the real world. As students dive into a case study, they are often required to document the pros and cons of the approach taken. In a legal case study, they might even be asked to develop the strongest arguments for both sides of the dispute. Going through that process helps the students to understand the complexity and ambiguities of the case study and to examine the problem from multiple angles. I know that some people aren’t fans of courses that include a lot of case study work. However, if such exercises are done well, they can greatly enhance the ability of students to be effective in the real world once they graduate and begin applying their knowledge.
Both at the university program and corporate training program levels, I would love to see more emphasis put on case study-based courses to help students and employees better understand how to effectively implement analytics and data science processes that are both legal and ethical. There is no lack of examples to discuss that can challenge people to think through the ethics of a wide range of situations. Whether the Target pregnancy prediction case, the Microsoft Tay case, or Amazon’s case of biased hiring algorithms, robust discussion can be had about what decisions were made in these cases, how they could be defended (or not), and what might be done if a similar situation came up in the future. Those discussions will let participants appreciate how things that seem obvious in retrospect may not have been so clear cut with the information available at the time of the initial decision.
The most important learning that can be gained through a case study curriculum is how to assess novel situations that will arise in the future. The probability of coming across a scenario exactly like any given case study used in a course is very low. However, the probability of coming across a similarly complex and ambiguous situation as the case studies chosen for a course is a certainty. By studying the thought process of those involved in each case and what might have been done differently given the options then available, it is possible to be much more prepared when faced with your own ethical dilemma as you plan and manage your analytics and data science programs.
Help Your Team Stay Ethical with Case Study-Based Training
If you’re serious about keeping your organization on the right side of ethics, it will take ongoing diligence, focus, and debate. If employees are provided the opportunity to explore real-world examples of ethical successes and failures together, they will learn how to assess future situations and to evaluate different perspectives on ethics that exist outside of their own.
Ethics is one area that is full of gray, with little black or white. Instead of making snap decisions and judgements, the gray areas should be explored and debated at length. Once a team is experienced in going through past examples together and reaching consensus on the good and bad of what happened, they’ll be that much more prepared to work together to make the right decision when it is time to take actions that will create a case study for the future.
If you have any ideas about examples and case studies that would be useful and effective in such a course, or if you have strong thoughts for or against what I’ve proposed here, please feel free to send your comments.
By Bill Franks | Jun 15, 2022
Originally published by the International Institute for Analytics