The idea of being more— rather than less—sensitive seems intuitively appealing. Take COVID tests for instance: humanity has put enormous amounts of effort into making sure the disease is brought under control. So, more sensitive tests sound like a great idea. As I tried to express in my previous post, however, there are costs involved. And some people being treated with suspicion is one of them.
To get to the bottom of the problem, I find it necessary to ask why people care so much about having more and more sensitive tests in the first place. When it comes to COVID, it seems that we want to avoid health risks and, ultimately, death. To some extent, I believe that wherever people make emotionally charged claims about the value a test or some other detection activity brings, this very frequently comes down to safety. Or rather it comes down to people feeling safe.
A recent discussion at my current work place centered around skin cancer detection. In theory, it sounds great if pretty much all skin lesions that could turn into a melanoma—or some other form of cancer—were detected to prevent this from happening. If one were to pursue this course, however, it would dramatically increase the incidence rate. And with that would come messages of, “we found a spot on your skin that could turn into cancer, and we will remove it. And then we will put you on monitoring for the rest of your life.” Congratulations! Do you feel safer already?
Another instance I am thinking of is how—after terrorists used explosives in their shoes to attempt detonating a plane—screening procedures at airports to this day may include having to remove one’s shoes. In that case, the cost is mainly inconvenience. And given that pretty much everyone has to comply with this regulation equally, without discrimination, at least it doesn’t come with a kind of blanket accusation leveled at one group.
Quite differently, in the case of crime prevention and detection, statistically significant differences in perpetration rates are used by police and security personnel to target people who fit a profile. And similarly, it seems that in the political struggle between ideologically more extreme positions, certain superficial signs are taken as a litmus test of moral character. In short, some people feel scared, and use a marker of some sort to detect “badness” in others.
How does this matter?
Signal detection of any kind is generally imperfect. The task of finding and securing the sub-class of dangerous items or individuals out of a larger line-up of candidates always comes with two inversely related properties: sensitivity and specificity. If more true-positive cases are identified, more true-negative ones are classified as false-positives, and vice versa.
What I find highly problematic is a tendency I see everywhere: uncritically accepting that higher sensitivity is better, not bothering about the cost. This makes it incredibly difficult to have the necessary and likely painful conversation about what a potentially more optimal trade-off between sensitivity and specificity might be in a given situation.
If I look at crime detection as an example, the cost of attempting to detect crimes before they happen clearly requires putting more people under suspicion—which is true of petty crimes as well as thought crimes. Doing so will likely increase the pressure any group of people being profiled experiences. And this increased pressure is likely to translate into all sorts of outcomes that are difficult to measure.
If this were openly discussed, that might undermine efforts in service of safety and security. Which is why I believe it is such a difficult conversation. In this case, different people reference the high-sensitivity end of the trade-off as either “law and order” versus “systemic racism” or “political correctness” versus “free speech limitations”. And neither absolute end of this spectrum is a desirable place to exist in. The increasing opposition to the “defund the police” slogan and cancel culture demonstrate that.
Unfortunately, many people have come to expect that safety can be increased at little or no cost. And in some areas in which self-imposed passive protection exists—like seatbelts being introduced as a standard feature in cars—this may be (almost) true. In other areas, the costs of increasing safety far beyond any optimum of trading off sensitivity and specificity are enormous. Or at least they are enormously unknown.
In the political domain, our desire to spot enemy partisans, who disagree with our cherished beliefs, has increased polarization to the point where normal discourse is often replaced by ideological shouting matches. We no longer seem able to step back and think about those consequences. Instead, we have become certain that without fighting those elements of reality we perceive as dangerous, the world will surely go to hell. And I believe that it is, ironically, this attitude of certainty about our own preferences in some trade-off domains that contributes to the going-to-hell part.
For me, personally, that means I want to learn how to remain open and curious. If I meet someone who has a very different preference compared to my own—especially in a domain where signal detection and prevention of any kind plays a role—I believe it is vitally important to appreciate that each preference point on this continuum has merit. That is the nature of a trade-off. Different people experience the world differently.
Some may feel so insecure about petty criminals that they wish for perfect surveillance of those who most closely fit the profile of previous crime perpetrators. Others may feel so frightened of airborne diseases that they wish masks to be a permanent feature of our culture in almost all circumstances. Yet others feel that certain political ideologies are so dangerous that any potential adherents must be fought with the utmost vehemence. In each case, there are overlooked costs incurred. And without realizing that one’s own preference is merely one point along a continuum for which no objective optimum exists, it is easy to get stuck believing that others must adopt one’s own stance on a given issue.
The second thing for me to learn is that feeling safe is an inner experience. And demanding additional, externally implemented safety measures is but one way to achieve it. In fact, I would go so far as to say that ultimately the feeling of safety cannot be achieved by any such measure. The more we fall prey to the belief that “feeling safe” can be achieved through external measures, the more unsafe we feel. We are losing our agency—in order to feel safe—to aspects we ultimately cannot control. Instead, I believe it is possible to learn how to best navigate the real risks that life brings. And in doing so, I can begin to feel safe because I know where the real dangers are.