A Psychometric Assessment of the Circadian Amplitude and Phase Scale

Abstract

The Circadian Amplitude and Phase Scale (CAPS) is a new self-report tool that aims to assess amplitude and phase. The CAPS consists of three factors made up of 38 items. Amplitude is posited to be assessed via two of these factors: time awareness (TA) and/or strength of preference (SOP). The remaining factor, phase, is assessed via the existing Preferences Scale (PS). Given there is no published research using this measure, the authors undertook two studies to examine its psychometric properties and construct validity. In a sample of 351 North American students, the authors identified a three-factor 34-item model using principal components analysis. The components explained 39% of the variance, and scale reliability ranged from 0.73 (SOP) to 0.88 (PS). The correlations between the components were small, suggesting they are each assessing different constructs. Thus, it is unlikely TA and SOP may both be indicators of amplitude. The distributions for each scale were divided into two groups (≤25th and ≥75th percentiles), and these groups were used to assess construct validity using alertness ratings by time-of-day. The results from a multivariate general linear model indicated a significant difference (p < .001) in sleepiness ratings by time-of-day for the PS scale only. In the second study, the authors applied confirmatory factor analysis in an attempt to replicate the factor structure identified in the student sample in a working sample (n = 388). The results suggested the model fit was not optimal. Subsequent analysis suggested a 26-item model was a better fit, but it also was not optimal. The intercorrelations between the factors were larger than in the student sample but again small, indicating the constructs are independent measures. The inability to replicate the factor structure is most likely explained by the differences in the sample characteristics of age, sex, and perhaps difficulty in interpreting the items. The limitations of these studies are discussed and recommendations for future studies are made. (Author correspondence: v.dimilia@cqu.edu.au)