Autistic children who create imaginary companions: Evidence of social benefits.

The study relied entirely on parent report and did not test the children directly. Because parents reported retrospectively, forgotten or misremembered imaginary companions could affect the results. The design compared groups at one time and cannot show that imaginary companions caused higher understanding-of-minds or social-skills scores. Participants were recruited through Facebook and Prolific, so the sample may not represent all autistic children. The online questionnaire was shortened, so some subscales were omitted and details such as imaginary-companion type and autism diagnosis were not collected. The perception subscale had poor reliability, so the lack of group difference for perception should be interpreted cautiously. Some parents completed the questionnaire quickly, which could mean they did not read questions carefully.

One author declared financial interests in the M-CHAT autism screening tool and an advisory role at Quadrant Biosciences. Participant recruitment was supported by internal quality-related funding from York St John University.

Do not conclude that imaginary companions improve autistic children's social skills or understanding of minds. The study only compared parent-reported groups and did not establish cause and effect.