What are blind and double-blind trials?

The root of the problem: expectation shapes the result

When people believe a treatment will work, they often do genuinely feel better, even with no active ingredient. This phenomenon, the placebo effect, is not an illusion but a measurable response, and it is especially pronounced in subjective experiences such as pain. The same holds for the side doing the measuring: an assessor who knows which participant is in which group may unwittingly be more encouraging, or score borderline cases in the direction of their expectation. Nobody cheats; bias seeps in through good intentions.

Clever Hans: the horse that read faces, not questions

In early 1900s Germany a horse named Hans became famous for answering arithmetic questions by tapping his hoof, and there was no fraud involved. The puzzle was solved by an elegant series of tests: Hans succeeded when the questioner knew the answer; if the questioner did not know it, or could not be seen by the horse, the success disappeared. The horse was reading not arithmetic but people, catching the involuntary shift in the onlookers' posture as he approached the right number, and stopping there. The episode showed that in experiments anyone who knows what to expect can be a source of involuntary signals, and it became the rationale for blinding.

From blind design to double blind

Single blind: the participant does not know which group they are in, so the expectations of the two groups are equalised and the difference between them can be attributed to the treatment itself. Double blind: those administering the procedure and assessing the outcome are also insulated from group information; the codes are held by a third party until the study ends. This removes the expectations of both participant and researcher. Blinding works together with random allocation to groups, and the pair constitutes the gold standard for testing causal claims.

Double-blind design in astrology tests

The logic of double blinding is not peculiar to drug trials; it can be applied to any kind of claim. In the known experiments testing astrology, astrologers were asked to match the birth charts of people whose identities they did not know with detailed personality profiles; neither the astrologers nor the assessors knew which chart belonged to whom. The design was set up with the consent of the participating astrologers, and the criterion of success was agreed in advance. The results did not exceed chance level. The strength of these experiments is that they accuse nobody in advance: blinding prevents honest people from influencing one another through involuntary signals, and what remains is only the strength of the claim itself. For more detail, see our guide to whether astrology is a science.

Blinding on stage: old violins and blind tastings

The power of blinding has been shown many times outside medicine as well. In the world of classical music the received wisdom for centuries was that the violins of the old Italian masters were beyond comparison with modern instruments. In double-blind comparisons with professional violinists, using arrangements in which the players wore darkened goggles on a dimly lit stage and did not know which instrument they held, the result was surprising: participants could not reliably tell old instruments from new ones, and in many trials preferences fell on the side of the new violins. In the same way, it has often been observed that in blind tastings experienced tasters cannot reproduce the price ranking once the bottle label is covered. These results do not devalue anyone's ear or palate; they show that expectation mixes into perception. Label, price and story genuinely do change the experience, and blinding is the only way to tell how much of that experience comes from the object and how much from the story.

Limits and everyday use

Not every question lends itself to double blinding: in studies of surgical technique, exercise programmes or diets it is usually impossible to blind the participant, and researchers then try at least to blind the assessors. The practical lesson for daily life is this: when you read a report of a study, asking "did the participants and the assessors know the groups?" is one of the quickest ways to weigh the strength of the evidence. Even in your own small trials, simple blinding is possible: you can test whether you really taste a difference by having someone close to you cover the labels.

Frequently asked questions

This guide draws on research methods textbooks, the historical record of the Clever Hans case, and examples of double-blind studies published in peer-reviewed journals.