summary
background A substantial proportion of psychiatric research is based on surveys that rely on self-report scales. Little is known about careless respondents in such surveys.
aim To explore to what extent careless respondents may bias the outcomes of surveys.
method Three surveys were conducted among paid volunteers (undergraduates).
results A small but non-trivial proportion of undergraduates admitted that they often intentionally gave wrong answers in surveys. These respondents also exhibited overendorsement of nonsensical items on the Infrequency Scale. In addition, many of them seemed to show a preference for reporting low prevalence symptoms (e.g. amnesia and derealisation).
conclusion Should we take careless respondents seriously? Our results demonstrate that researchers and readers of their articles are well advised to do so, at least when research reports concern rare symptoms and experiences.