Abstract
Background. Attachment has become a central construct in explanatory models of psychosis. The Psychosis Attachment Measure is currently the most widely used measure for attachment in a population with a psychotic disorder. However, concerns have been raised about its psychometric properties. In addition, there are two perspectives to operationalize attachment: dimensional and categorical. The categorical method offers added value for the assessment of disorganized attachment which has been mostly linked with vulnerability to psychosis.
Aim. The aims of this study were to re-evaluate the structural and construct validity of the dimensional and categorical approaches of the Psychosis Attachment Measure in a Dutch sample (N=287).
Results. Confirmatory factor analysis indicated poor fit indices for the two-factor model with low loadings of the reversed questions of the Psychosis Attachment Measure on the avoidance factor. Removing these items improved model fit to conventional threshold values. Latent profile analyses suggested a four-class solution: secure, preoccupied, dismissive-avoidant, and disorganized attachment. The construct validity of both approaches was largely in line with expectations.
Conclusion. Both the dimensional and categorical perspective of attachment in psychosis may be used to gain a better understanding of the complexities in the attachment system.