Abstract Although several studies have examined the reliability of diagnosing delusions there is no comprehensive review of the literature. Therefore, the reliability of diagnosing ‘delusions in general’ and the subcategory of ‘bizarre delusions’ was reviewed, including both structured interview and standardized instrument methods. The literature suggests that delusions in general can be diagnosed reliably with both structured interview and standardized instruments. However, bizarre delusions are not reliably diagnosed by either, suggesting that this concept may have little clinical validity. Nevertheless, many of the studies reviewed are poorly designed or subject to significant confounds. Criteria are suggested for adequate future studies.