Yes, research carried out by the Multisensory Perception and Action Group at the Max Planck Institute for Biological Cybernetics in Tübingen confirms the truth of the anecdotal concept.
Scientists in the Multisensory Perception and Action Group at the Max Planck Institute for Biological Cybernetics in Tübingen, led by Jan Souman and Marc Ernst, have now presented the first empirical evidence that people really walk in circles when they do not have reliable cues to their walking direction.
The results showed that participants were only able to keep a straight path when the sun or moon was visible. However, as soon as the sun disappeared behind some clouds, people started to walk in circles without even noticing it.
Speaking about the study, Jan Souman said: "One explanation offered in the past for walking in circles is that most people have one leg longer or stronger than the other, which would produce a systematic bias in one direction. To test this explanation, we instructed people to walk straight while blindfolded, thus removing the effects of vision. Most of the participants in the study walked in circles, sometimes in extremely small ones (diameter less than 20 metres)."
However, it turned out that these circles were rarely in a systematic direction. Instead, the same person sometimes veered to the left, sometimes to the right. Walking in circles is therefore not caused by differences in leg length or strength, but more likely the result of increasing uncertainty about where straight ahead is. "Small random errors in the various sensory signals that provide information about walking direction add up over time, making what a person perceives to be straight ahead drift away from the true straight ahead direction," according to Souman.
All quoted from Walking in Circles - Max Planck Society (Research by Jan Souman and Marc Ernst.)