Using INFORMS ethics guidelines in the classroom
By David Hunt and Scott Nestler
Work done by members of INFORMS increasingly impacts peoples’ lives as the use of our algorithms and models spreads and as data collection becomes more detailed and more personal. Yet too often we become absorbed by the elegance of our models and research, failing to consider any potential ethical implications. Or, we find ourselves faced with a situation that seems unethical, and are forced to determine the best course of action without previously having established exactly what we consider to be ethical.
INFORMS recently approved a set of ethical guidelines “intended to be aspirational, something members should attempt to follow throughout their career” [1]. The guidelines are organized around our responsibilities to society, to the organization where or for whom we work, and to the profession of operations research. The guidelines include 18 topics, ranging from “accountable” to “rigorous” to “tolerant,” with each topic accompanied by a brief description to provide aspirational guidance without defining an exact course of action.
The INFORMS ethical guidelines can form the basis of a classroom exercise to allow students to develop their own definition of the lines between ethical and unethical behavior. By working through all, or some, of these 18 guidelines, students can create their own code of ethics they can reference, refine and aspire to follow throughout their career.
One way to develop a personal code of ethics is to use the framework contained in Howard and Korver [2]. To illustrate how this works, consider the following from the INFORMS Ethics Guidelines: I aspire to be accurate in my assertions, reports and presentations.
Certainly nobody aspires to be inaccurate, but how much accuracy is sufficient and how can this be expressed in your code of ethics?
Step 1: Draft Standards
The best place to start is by writing a statement that captures your intent, perhaps, “I will strive to be mistake free, and work to be as close to the optimal solution as possible.” With that as a basis, think about the practicality of this statement. Is it too vague? What obstacles might prevent me from achieving this? Are there exceptions when I would willingly violate this statement? These might include:
- How much time and budget do I have to verify the accuracy?
- Does every activity require the same level of accuracy?
- If I spend too much time and budget trying to achieve an unnecessary level of accuracy, is that the same as stealing from my organization or from a client?
- Is checking the accuracy yet again the best use of my time?
The conclusion of a project is perhaps the worst time to objectively consider the right level of accuracy. It is easy to get caught up in trying to close the optimality gap, or to overly wordsmith a report or presentation. Or, even worse, due to a looming deadline or depletion of the budget, shortcuts are taken with regard to accuracy.
It is better to establish the appropriate level of accuracy for a project at the start of the effort. How accurate does this need to be? Is there a standard procedure, such as formal software testing that needs to be followed? Has adequate time been allocated for testing and checking? With that in mind, an improvement to the original accuracy statement would be:
I will apply the appropriate level of accuracy to every project I undertake by:
- Establishing the level of accuracy for my portion of the work at the project start, obtaining agreement with clients, superiors and the project team as necessary.
- Checking that sufficient time and/or budget has been allocated for me to achieve this level of accuracy, and if not, I will point this out to the project manager and/or client.
- Understanding and/or establishing an appropriate testing/verification procedure at the start of the project, and then following it at the end of the project, understanding that some adjustments to the procedure may be necessary provided it does not compromise the agreed upon level of accuracy.
Step 2: Test the Ethics Statement
Once the statement is drafted, it is then time to test it for logic, focus and usefulness. A common test for logic is universality. Would you want this statement to be the universal law for how everyone determines accuracy? Consider a colleague who wrote software for scheduling aircraft maintenance. As you step onto a plane that was maintained using this software, would you be comfortable, or would you be nervous, knowing they had used your statement on accuracy as a guide?
Checking for focus involves making sure the statement is not too unwieldy, thus making it unmanageable. There is a trade-off between a statement that is too brief to provide guidance and one that runs on for pages trying to cover every situation and every exception. Longer statements describing your code of ethics run the risk of creating logical conflicts, where you cannot possibly follow two different exceptions.
Finally, test the code for usefulness. Think of past projects at work or school where you did have accuracy issues, or perhaps, where you spent too much time trying for an exact solution. Would your draft ethical code on accuracy have been helpful in these situations?
Step 3: Refine the Code
Finally, think about any additional refinements necessary to complete your ethical code on accuracy. This should be done as a final step, but it should also be done as your situation changes, such as moving from school into a job, receiving a promotion or switching employers. Minor revision may also be needed when changing from one project to another, as you gain more experience and learn from past issues.
One refinement might involve degrees of separation. How far from the final product do I need to be for me to not feel culpable for an error? If my boss says, “It is good enough, send it out,” does that absolve me of any errors that may be detected? What if I write software code for a project that has software testers responsible for accuracy; is an error the tester’s fault and not mine?
Another refinement might be whether or not you accept, or remain in, a job where your ethical code involving accuracy is not consistent with your organization’s policy or actions. This discussion of accuracy has addressed errors, but what about intentionally “inflating” results? Is it lying if this is done to the benefit of the client? Would you remain with this organization?
Finally, what if your colleague consistently produces results with accuracy problems? Should your code of ethics cover this case? Is it your responsibility to speak to the colleague, or perhaps to a supervisor? Do you wait for someone else to speak up?
Many people, students and experienced professionals alike, do not think about ethics and about how they would respond to a situation until they are faced with an ethical dilemma. Raising the awareness of the types of ethical issues that arise in operations research and analytics should be part of every student’s education. The recently approved set of aspirational ethical guidelines from INFORMS provides a useful framework to use in discussing these issues in the classroom with students.
David Hunt ([email protected]) is an engagement manager at Oliver Wyman, chair of the INFORMS Pro Bono Analytics Committee and chaired the INFORMS Ad Hoc Committee on Ethical Guidelines
Scott Nestler ([email protected]), Ph.D., CAP, PStat, is an associate teaching professor in the Department of Information Technology, Analytics and Operations (ITAO), Mendoza College of Business, at the University of Notre Dame.
References
- INFORMS Policies and Procedures Manual, Section 1.2, “Ethics Guidelines,” https://www.informs.org/content/download/348012/3212878/file/Section%201.pdf
- Howard, Ronald A., Clinton D. Korver, 2008, “Ethics {for the real world},” Harvard Business Review Press, Boston, Chapter 4.