It's just not a binary choice
Public polling takes a wicked turn on vaccine safety perceptions: The CDC's reaction to possible side effects from the Johnson & Johnson Covid-19 vaccine appears to have dealt a non-trivial blow to public confidence in that vaccine. A poll by YouGov found that 52% of US adults considered the vaccine "safe" before the CDC recommended a pause in its use, but only 37% said it was "safe" after.
■ Putting aside for a moment how preposterously low the original 52% figure was, we really need to come to terms with a big problem of framing. There are inherent limitations to polling, to be sure. But asking a binary question like "Is this vaccine safe or unsafe?" is a disservice, because it trains people to examine important questons as though they are simple binaries. For all the good binary math does us in making digital calculations possible, we mostly make analog decisions.
■ Think of it like this: Lying in a bed full of puppies isn't completely "safe" -- maybe one of them will bite you. Dogs commonly carry germs in their saliva that could kill some humans. Yet still, on a scale rating safety with a value from 1 to 10, playing with puppies is probably a 9.95. Does the missing 0.05 make the activity unsafe? Of course not -- unless you're strictly adhering to a mindset that something is either "completely safe" or not.
■ We deal competently with gray areas, odds, and uncertainty all the time. There may be a chance of rain when you're planning to grill a steak medium-well while thinking about your fantasy baseball roster. Many of us profess to know little about math yet turn out to be very good at understanding complex conditional probabiliies when it's time to set up NCAA basketball tournament brackets.
■ We ought to use that comfort to learn not to be quite so innumerate. Public discourse needs to make room for more gray areas, and one step in the right direction would be to reject the framing that something like a vaccine ought to be labeled so simply as "safe" or "unsafe".
■ The side effect of concern with the Johnson & Johnson vaccine appears to be a potential issue with blood clots that has occurred a reported 6 times after nearly 7 million doses. To call that "extremely rare" is extreme understatement. If something bad happens once per million people per year, that's roughly on par with the odds of being struck by lightning. Prudent behavior accounts for those kinds of risks, but doesn't overreact to them: We don't live every moment inside Faraday cages; we just observe the maxim "When thunder roars, go indoors".
■ The same should apply to how we discuss matters like vaccine safety. Asking the public whether they think the Johnson & Johnson vaccine is "safe" is a lot like asking whether a blue sky with sporadic clouds is "sunny". Strictly sunny? No. But meaningfully so? Of course.
■ If we can't acknowledge gray spaces as an ordinary fact of life, then it becomes basically impossible to distinguish when a tiny change really doesn't alter the truth. And in the case of the vaccine, supposing that a 6-in-7-million side effect were to be found consistently, that would represent a baseline change in "safety" of 0.0000882%. Put another way, you could give the vaccine to the entire population of the United States (330,000,000 people) and fewer people would suffer the side effect (around 300) than the 356 people who are currently in the ICU with Covid-19 -- just in Ohio.
■ We need to be able to see that "safe" is not only a vague term, it is wholly unhelpful in most human contexts. And though it may seem paradoxical, the best way to deal with human risks is to use more math in our language and fewer absolutes.