Survey Experiments: Testing Beliefs by Asking Others

Test social predictions by asking a small sample of others what they actually think — instead of assuming.

Why it works

Many anxiety-maintaining beliefs are social predictions ("everyone notices my blushing," "people think I’m incompetent when I make a mistake"). These predictions are not testable by individual behavioral experiments alone because they are about others’ mental states. Survey experiments collect real data from real people, directly confronting the socially- inferential assumption with evidence. Because the mind is notoriously poor at mental-state attribution under anxiety (assuming others attend to and negatively judge self-focused concerns), external data is often powerfully disconfirming.

How to do it

1. Identify a social belief that is about what others notice or think: "I believe people notice when I am anxious."
2. Design a survey: ask three to five people whether they notice [specific thing] in social situations.
3. Approach the survey neutrally — do not hint at what you hope to hear.
4. Record responses and compare to your prediction.
5. Formulate a revised belief based on what the data shows.

Evidence

Survey experiments are a standard CBT technique for testing social predictions, particularly in social anxiety disorder. They directly test the mind-reading errors (assuming others notice or judge specific behaviors) that are well-documented cognitive distortions in social anxiety. (clinical)

Survey experiments are clinically established; large-scale controlled evidence for this specific format is embedded in social anxiety CBT trials rather than isolated.

Sources

• Bennett-Levy et al. (2004), "Oxford Guide to Behavioural Experiments in Cognitive Therapy" — extensive survey-experiment clinical guidance

Common mistake

Selecting people for the survey who are known to be exceptionally kind or encouraging — the sample must be representative, or the data confirms what you already wanted to believe.