Measuring stigma

King et al (2007) frequently state that their stigma scale is measuring ‘the stigma of mental illness’ but, when closely scrutinised, it measures nothing other than stigmatisation perceived by users in out-patient, in-patient and crisis settings. There is no evidence that this is an objective assessment of stigmatisation. Users’ perception of stigma is affected by their mental state, depression, persecutory delusions or hallucinations. These symptoms can help to exaggerate the estimate of social stigmatisation (including rejection and discrimination) and hence the assessment is by no means an accurate measure. Measurements of more objective perceptions of stigmatisation can only be obtained from users in remission.

The reported negative correlation between self-esteem and perceived stigma can be confounded by high rates of both low self-esteem (e.g. Axford & Jerrom, 1986; Barrowclough et al, 2003; Blairy et al, 2004) and persecutory ideation and depressive cognition, including ‘self-stigmatisation’ in people with mental illness. Indeed, low self-esteem is a common symptom in psychiatric conditions such as depressive disorders, in which people can perceive more rejection and discrimination than warranted. Overemphasis on this correlation can divert attention from the fact that the correlation has to do more with people's mental state than objective level of social stigmatisation.

An instrument can only be called ‘standardised’ if it is shown to be both reliable and valid. This instrument is not validated and so cannot be called standardised, on the basis of mere test–retest reliability. The correlation between the stigma scale and self-esteem scale is not an indication of validity of the instrument and although King et al admit this, they end up referring to their instrument as ‘standardised’ and to the correlation as ‘concurrent validity’.

A wide range of people with diverging diagnoses and mental states were recruited by King et al but there was no randomisation and no exclusion criteria. Even the ‘perceived stigmatisation’ cannot be attributed to a particular category of patients with a given diagnosis, or at least to psychiatric users in general, owing to lack of randomisation and inclusion of arbitrary proportions of participants with different diagnoses. This is likely to cause problems in comparative studies. Also, stigma by definition excludes ‘positive aspects of mental illness’. This is why the authors decided to reverse the scores of the ‘positive aspects of mental illness’ factor. For this reason, they should have also called the factor ‘negative aspects of mental illness’, as a high score on this new factor then represents stigmatisation and its negative influence on the person.

In brief, a scale which partly measures people's mental state and partly objective social reality is neither valid nor standardisable because it cannot measure what it is supposed to measure (i.e. it cannot satisfy the fundamental condition of validity).