Can virtue be measured? That was the question before a conference held at Oriel College, Oxford in January sponsored by the Jubilee Centre for Character and Values of the University of Birmingham. I didn’t attend, but the nearly 40 papers delivered at the conference are available online at the Centre’s website. Reading through some of the papers and the paper abstracts I was struck by two things. First, while some of the presenters raised basic philosophical questions about the very idea of measuring virtue, many of the presenters, after a caveat or two about complexity, confidently said “yes,” virtue or character can be scientifically measured.
Second, I was struck by how little they agreed on how to do it, which one might have expected to undermine their confidence. Granted, the various speakers were not all addressing the same aspect of the elephant—some were concerned with measuring only a single virtue, some were asking about measuring character education programs, some offered instruments to measure how individuals view moral concepts, and so on—yet everyone seemed to have his or her favorite method or construct, and virtually no two were the same. The empirical assessment of virtue/ethics/character seems to a thriving industry. So far, though, it has produced no strong consensus of thought.
Perhaps the most interesting thing about the conference was the very question of measuring virtue. Who is asking this question and why now? Certainly one demand for measurement is coming from the new science of morality that is going great guns in psychology, neuroscience, and other fields. More generally, though, and feeding the science, are the imperatives of the “audit society” and the multiplication, over the past few decades, of formal “checking up” systems that were created in the name of accountability. These systems have spread far and wide, from government to education to medicine and beyond, and they require measurable indicators that are verifiable.
Michael Power, author of The Audit Society (1997), defines in an earlier paper a key aspect of verifiability: “that attribute of information which allows qualified individuals working independently of one another to develop essentially similar measures or conclusions from an examination of the same evidence, data or records….” This normally quantitative attribute is one thing in the context of a financial audit, where money is the medium, or quality assurance of, say, light bulbs, where readily visible criteria of success or failure exist and preexist the audit process. In these contexts, auditing and certifying are a secondary monitoring of compliance or performance.
However, it is quite another matter in contexts where the performance of doctors or teachers or professors is at issue. In these complex contexts—and this has implications for something as qualitative, relational, and holistic as virtue—activities that are not easy to quantify must first be “made auditable” by finding some feature of them to measure or rate. While this need not be a bad thing—it might, for instance, force professionals to think seriously about their goals—it can take on a life of its own.
The very act of creating measures and benchmarks and rating scales can badly distort the nature of the thing being audited, throwing off all sorts of unintended consequences. Far from a merely derived and neutral activity, auditing and performance measurement can construct a system of knowledge and then re-shape the organizational environment to make that system successful. More germane to virtue is the distinct possibility that because the disposition itself is not readily amenable to verifiable, non-subjective measurement, what will be quantified is simply some aspect that is easy to count, often a crude and not very meaningful aspect at that. This aspect, because verifiable and thus more tractable and “real,” then gets confused with the thing itself. Virtue becomes, as one of the speakers at the Oxford conference argued, “what virtue tests test.”
I recently heard a social scientist argue that when it comes to measuring morality any measure is better than none, an at-least-we’re-counting-something view which Power also observes in his research. But surely, in light of the dynamics of real-world assessment practices, such a facile view is deeply mistaken. Only a very good measure is better than none.