Although a wealth of literature exists studying the effect of sponsor characteristics on self-reports of mental health, little work assesses a related but potentially powerful effect: a context comprehension effect, that is, a change in the respondent's interpretation of a survey question, given the concept elicited by the interviewer. Further, most studies of sponsor effects assess mean differences in responses according to sponsor type; no work has studied how these effects can affect the reliability of a psychometric instrument, which are of general importance to sociological studies of health. Here, using a study of 169 recently discharged hospital patients, we find that both effects are likely to have affected responses and, together, negated the reliability of a psychometric instrument. Future studies should consider the effect of not only sponsor characteristics upon responses but also consider how questions might be interpreted given in an interview context.