Assessing non-dysfunctional attitudes toward sleep: psychometric properties of the Charlotte Attitudes Toward Sleep scale in Portuguese samples

Abstract

To adapt the Charlotte Attitudes Toward Sleep (CATS) scale, the only self-assessment instrument measuring non-dysfunctional attitudes toward sleep, into Portuguese and to study its psychometric properties in a Portuguese sample. A sample of 1858 participants, recruited through non-probabilistic methods, was randomly split in two subsamples; one was used to develop the CATS model, the other for testing model invariance. We used structural equation models to assess factorial validity, measurement invariance, and relationships with other variables (e.g., the Karolinska Sleep Questionnaire) through confirmatory factorial analysis and causal models using a robust maximum likelihood method with Satorra-Bentler correction. The CATS factorial model showed excellent evidence of factorial validity (robust CFI = 0.987, TLI = 0.979, SRMR = 0.026, RMSEA = 0.043), good reliability indicators (α and ώ1 > 0.75), and strict invariance of measurement (|∆CFI|< 0.01). CATS factors were weak-to-moderate predictors of sleep behaviors (β < 0.4). The scale showed evidence of divergent validity with the Karolinska Sleep Questionnaire. Some items had significant sensitivity problems and/or did not have adequate factorial weights and had to be dropped from the model. The CATS is a new and promising scale with solid evidence of validity in terms of internal structure, but with sensitivity issues at item level. Further work should be carried out at item level to increase sensitivity and predictive validity, and further research with other samples, such as clinical sleep medicine patients, should be conducted.