You may be aware of the placebo effect, where an inert pill has an effect because of what the patient thinks it does. You may even be aware of the nocebo effect, where an inert pill causes ‘side-effects’. But a fascinating 1970 study reported evidence for the anti-placebo effect, where an inert pill has the opposite effect of what it is expected to do.
Storms and Nisbett were two psychologists interested in attribution, the process of how we explain the causes of events and the impact this has on how we feel.
We know that attributions have a big impact on our level of physical and emotional health. For example, your heart is racing when you’re about to give a talk. If you attribute it to a weak heart, you may start worrying whether you might pass out and become incredibly stressed, but if you attribute it to the situation, you might just think its a natural reaction for the event and feel primed and ready.
In anxiety disorders, we know that people often attribute natural bodily reactions to frightening causes, which makes people feel more on edge, and hence, their body kicks into an even higher gear, and so on. The cycle continues, to fever pitch. In essence, it’s anxiety-fuelled anxiety.
Insomnia has an element of this. People can be worried that they’re not sleeping, and so get anxious thoughts when they go to bed, and so feel on edge, ad nocturnum, until the early hours.
So rather than getting people to fill in questionnaires about causes of insomnia, a typical method in attribution research, Storms and Nisbett wanted to test these ideas in the real world.
They recruited a group of patients with insomnia and told them they were doing a four-night study on dreaming and asked them to rate their difficulty in falling asleep each night.
The first two nights were exactly that, a sleeping and rating exercise, but on the third night the participants were given pills. One group was told that the pill would make them feel more aroused, like a shot of caffeine, while the others were told that the pill would make them feel more relaxed, like a sleeping pill.
On the fourth night, the group were given the ‘opposite’ pill, but in reality, all the pills were identical and completely inert, containing nothing more than sugar.
Now here’s the thing. The insomnia patients taking the ‘relaxation’ pills slept really badly, and the patients taking the ‘arousal’ pills slept much better.
What seemed to be happening was that patients taking ‘uppers’, normally trapped in a cycle of anxious self-monitoring, could attribute any arousal they had to the pill. Any sign of feeling wired wasn’t them, it was the pill, so they could relax and fell asleep easily.
In contrast, those who had taken the ‘downers’ thought that any arousal must be their insomnia causing them problems, and it must be really bad, because it was getting to them despite the supposed sleeping pill they’d taken. In other words, they were freaking out because they couldn’t sleep despite the ‘medication’.
It turns out that this simple experiment wasn’t easily replicated but the problem was solved in 1983 when it was realised that this effect only held for people with insomnia who obsessively self-monitored.
But what these experiments tell us is that the effects of medication, the symptoms of illness and even the process of ‘being sick’ is partly dependent on our own ideas about what’s happening.
Link to PubMed entry for original Storms and Nisbet study.
Link to 1983 replication.