In a recent article in Scientific American, authors Nathan Ballantyne, Jared Celniker, and Peter Ditto examine the question as to whether images of the impact of COVID-19, such as ill people in hospital beds, can help bypass the “persuasion fatigue” that is setting in as people tire of arguing over whether the disease is sufficiently serious to justify the series of lockdowns and other restrictions imposed by governments around the world. If sharing arguments and statistics hasn’t settled the issue to date, then perhaps graphic photographic evidence will do the trick by shocking doubters into agreement with those who preach the seriousness of the disease and the appropriateness of government action.
As the authors point out, the efficacy of even explicit photographs is minimal. The problem, they argue, is that photographs have a powerful emotional impact, such as shock or disgust, primarily for those already convinced of the terrible nature of the situation represented in the picture. If, on the other hand, “you entered our study doubting the threat, the images didn’t shock you and, accordingly, didn’t move your thinking… one person’s shock is another person’s shrug”, write the authors.
The limitation of photographic evidence to persuade, they conclude is a result of an “empathy gap” whereby we mistakenly assume that what moves us is what moves others, and in similar ways. This is a kind of emotional question-begging, whereby “when using images to persuade, we may take for granted precisely what we’re trying to prove” (italics in original).
I think this is all basically correct, but I think that the failure of “empathy” in these cases is in fact a disagreement over risk.
To see this consider the economic concept of expected value. When considering whether to undertake some action, A, one generally considers two factors: first, what is likely to happen should A be undertaken; secondly, what is the impact of those possible happenings. For example, in deliberating whether or not to attend law school, one may try to estimate the probability of getting a legal job upon completion with all the benefits that entails, such as easily paying off student loans. This will be balanced against the probability of not getting such a job and having to pay off student loans in those circumstances. This combination – probability (legal job) x benefit (legal job) + probability (no legal job) x cost (no legal job) – is the expected value of the decision to attend law school.
Notice that even if the cost of failing to secure a legal job is very high – suppose the student loans will take a lifetime to pay off in that case – one may reasonably decide to ignore that outcome if the probability of it happening is sufficiently low. If, for example, there is a 99.9% chance of getting a job as a lawyer upon completion of the law degree, then it may make perfect sense to take on very large student loans to complete the degree assuming a lawyer’s salary is high enough. If, on the other hand, the salary of a lawyer is not very high, then it may make no sense to pursue a law degree even if there is a 100% chance of getting a job at the end. Similarly, if one considers an outcome to be sufficiently bad, then even a very low probability of occurrence might be compelling; for example, one may refuse to skydive because one takes the very low probability of death to outweigh the enjoyment of the very high probability safe landing. What matters, in other words, is not just the goodness or badness of the outcome but its probability of occurrence, and a decision can only be assessed if both dimensions are taken into account.
Consider, for example, something very familiar: getting in a car and driving to work. We all know that a serious car accident, such as a highway collision, is a very bad outcome: death or serious injury usually result. Still, one will likely be willing to get on the highway and drive to work every day so long as the probability of such a bad outcome is sufficiently low, which in fact it is. In such a case, one will take the chance knowing how bad the outcome is. Even if one is shown a gruesome image of the aftermath of a highway collision, then one will likely continue to drive to work because the image addresses the cost of an accident not its probability. This is a general feature of a photograph: because it shows a single incident, it cannot tell us very much at all about the probability of such an incident occurring. Similarly, an image of a victim of a violent mugging in another city will not cause one to stay home at night unless one is convinced that such a beating is sufficiently likely to occur to oneself when out on the street. If not, then the picture will have minimal impact.
Something important should be made explicit, however: there is a subjective element to all of this, namely, at what probability does a bad outcome become too risky? There is no agreed upon answer to this. We each have our own set of risk tolerances.
So, the logic of the “empathy gap” is in fact a disagreement over what counts as a reasonable risk. Some believe that COVID-19 poses a probability of suffering or death sufficiently low to be worth risking exposure. For them, images of such suffering or death simply miss the point: they may make vivid how badly things may go but not how likely it is that they will go that way. These are like people who continue to drive to work knowing that there is a chance they die in the process. For those, on the other hand, sufficiently moved by the badness of the disease to think that even a very low probability of death or suffering is not worth the risk, then a graphic image will just reinforce what they already believe: that the disease is too horrible to take a chance on. They are like people who refuse to skydive despite the safety record or parachutes.
In other words, for one side, the low probability outweighs the magnitude of the outcome while, for the other, the magnitude of the outcome outweighs the low probability. Showing a photograph is not so much a matter of taking for granted what one is trying to prove – “the outcome of COVID-19 is horrible” – as it is an irrelevant gesture in this debate. Since a photograph doesn’t tell one how probable what it depicts is, it cannot address those who think the risk is low enough to be worth taking.
What of those who think the disease is no worse than the flu or who think it is a hoax? The former would be those who dispute the consensus of the medical community: how likely are they to see one or two photographs as convincing? The latter are those who dismiss the evidence and opinions of the medical community as bogus: how likely are they to see some photographs as veridical, rather than fakes? In both cases we have a similar problem of assessing probabilities. The first group, for whatever reason, considers the medical evidence to date to give a low probability to the proposition that COVID-19 is much worse than the flu, so is likely to see a few bad photographed cases as outliers, the kind of thing that happens in any flu season. The second group assigns no significant probability to anything the medical community has to say on the matter, so will assign a higher probability to the photograph being a fake than anything else. In short, for those who have lost trust in the medical establishment, some gruesome pictures will be dismissed as irrelevant or fakes. For those who have not lost their trust, a photograph cannot tell them the probability of what is depicted therein, so cannot impact their assessment of the risk involved in locking down or not.
We see, then, why there is no way to close an empathy gap with any photographic intervention: the problem is a logical one of probability plus outcome underdetermining risk tolerance, which is subjective. Photographs simply don't address that issue.
No comments:
Post a Comment