Exposure Therapy: When Probabilities Fail
By Bryan Caplan
In chapter 4 of Selfish Reasons to Have More Kids, I show that – objectively speaking – kids today are safer than ever. And I’m far from the first social scientist to point out the public’s systematically biased beliefs about risk – try Aaron Wildavsky, Kip Viscusi, Cass Sunstein, or Levitt & Dubner. One fact that we quants tend to overlook, unfortunately, is that many people still feel extremely afraid even after they accept the facts. This doesn’t mean that correcting misconceptions is useless – many people’s fear does respond to facts. But it does leave a tough question: What can we do for everyone else?
I’m pleased to report that an effective supplement to objective risk analysis exists. It’s called exposure therapy. Psychologists use it to help people with severe anxiety problems – people who have nothing to fear but fear itself. The best source I’ve found is the Handbook of Exposure Therapies. Intuitively, the idea is to get people to “face their fears.” From the introductory chapter: “Exposure therapy… involves deliberate and planned exposure to a feared stimulus, or representation of the stimulus, until the intensity of the person’s distress recedes…” The basic steps:
1. Choose to vividly experience a disturbing but tolerable fear. Depending on the fear, this means either first-hand experience (e.g. spiders), or detailed imagination (e.g. the kidnapping of your child). “Regardless of the way exposure therapy is conducted, clients are encouraged to confront their fears in the present-tense. For example, a client recounting a traumatic event in exposure therapy is encouraged to describe the event as if it were being relived.”
2. Continue doing #1 until the experience or imagination has become less frightening.
3. Move on to more intense anxieties and repeat.
The Handbook also reviews clinical evidence on exposure therapy vs. other talk therapies vs. drugs vs. nothing vs. combinations of the above. The contributors to this volume are obviously sympathetic to exposure therapy, but their survey of the evidence is still impressive. In almost every case, they conclude that exposure therapy plus X is no better than – and often worse than – exposure therapy alone. The zero or negative marginal benefit of drugs is awfully Hansonian:
With respect to short-run efficacy, a number of studies suggest that [some drugs] may enhance the effects of exposure-based CBT [cognitive-behavioral therapy]. However, an approximately equal number of clinical trials provide no support for this conclusion, and a meta-analysis of this literature indicates that combined treatment is no more effective than CBT alone… On the other hand, clinical trials have consistently failed to support an advantage of combined treatment when long-term outcomes are considered. In fact, the two largest and most well-designed trials of combined treatments provide unambiguous evidence that pharmacotherapy… interferes with the durability of exposure-based CBT.
Does this mean that teaching probabilities is a waste of time? Not at all. Many people’s anxiety does respond to objective information. But even if your anxiety is more stubborn, it’s still good to learn about probabilities. If the sky really is falling, you should run for cover. But if the sky isn’t falling, the best way to feel better is to sit down under the open sky and face your fear.
P.S. How can a Szaszian take this stuff seriously? Easily. Szasz would deny that anxiety disorders are a “disease.” But he’d surely admit that extreme fear of minor risks is a “problem in living” – and there’s nothing anti-Szaszian about carefully comparing different strategies for overcoming such problems.