Andrew Gelman and David Madigan at Columbia recently published an excellent reflection on the similarities between ethical decisions and statistical inferences. (This should be required reading for those of us who are developing content for the Ethics Education Committee’s workshops.) In particular, they point out that when problems (what we often call “dilemmas” in ethics) are either too easy or too hard, it is difficult to learn anything useful from them.
To show this, they begin with a really simple case in statistics, namely, election forecasting. If an election is too close, a statistical inference about the discrete outcome (wether the Democrat or the Republican will win) will not tell us anything about the model we’re testing. Likewise, when the election is a landslide, our statistical model is not really getting tested either. At least not by the discrete outcome. (As Gelman and Madigan point out, the model should be tested, not against the outcome of the election, but against the vote differential.)
They then apply this to ethics cases:
In general … the most informative ethics vignettes are those in which the call is not so close as to seem arbitrary, but not so obvious that the decision can be made without thought. The purpose of discussing the intermediate cases is to explore the way in which careful assessment of goals, motivations, costs, and benefits can help us make better decisions, and also help us understand the decisions of others.
We can take the question of scholarly misconduct, for example. There are really easy cases in which one scholar steals the work of another and publishes it as their own (or the case of student misconduct where a student buys a paper from another and submits it for a grade.) There isn’t really anything to discuss there; it’s wrong and everyone knows it. At the other end of the spectrum there are cases that are too difficult to evaluate: where a dispute over giving proper credit for a limited contribution to a paper arises, and all the facts are a matter of what was said to whom at what time without records being kept. In such cases, a practical settlement may be brokered, but it won’t clarify any principles. The disputing parties are (hopefully) just looking for a way to put it behind them and move forward.
The interesting cases are those where there are judgments to be made and tradeoffs to be considered. In the weeks to come, I’m going to be make an effort to write up some “vignettes”, with the aim of starting a discussion about whether they offer a useful basis for a discussion of ethics. I will be interested both in what is the right way to conduct oneself as a scholar in one’s primary work, and what is the right way to proceed when one encounters a possible ethical violation. As I’ve said before, that second set of problems is as rich an area for ethical reflection as the first.
I should be ready on Tuesday with my first vignette.