Background
There remains a debate about whether symptoms of psychosis lie on a continuum with healthy functioning or exist separately and are taxonic. This issue has important implications for the classification, assessment and treatment of psychosis. Research has highlighted that some symptoms of psychosis, such as paranoia, have a dimensional latent structure but it remains to be seen whether this is true for other symptoms.
Aim
To assess the latent structure of hallucinations in a diverse sample using taxometric methods.
Methods
Three taxometric procedures; MAMBAC, MAXEIG and L-MODE, were applied to a dataset of clinical (n=290) and non-clinical (n=1580) participants who had completed the Launay-Slade Hallucinations Scale- revised (LSHS-R). Analyses were initially conducted with a non-clinical group before a clinical group was included, increasing the likelihood of producing a pseudo-taxon.
Results
Three out of six taxometric analyses found a strong dimensional result (non-clinical sample; MAXEIG and L-Mode analyses. Whole sample; MAXEIG analysis). Two of the other three results were more in favour of a dimension (non-clinical sample; MAMBAC analysis and L-Mode analyses). The final analysis (whole sample; MAMBAC) supported neither a dimension nor a taxon.
Discussion
Despite some ambiguity in the findings, we observe some indications that hallucinations, like paranoia, could be dimensional, especially in the non-clinical sample. Clinical implications of these findings are discussed. Potential issues with the LSHS-R mean that results should be interpreted with some caution. The development of additional scales or assessments for hallucinations, expanding recruitment to more diverse non-clinical and clinical populations, is recommended.

Keywords: Hallucinations, taxometrics, MAMBAC, MAXEIG, L-MODE