Validation of a chronotype questionnaire including an amplitude dimension

The aim of this paper was to validate a French version of the Chronotype Questionnaire (Ogińska, 2011, Pers. Individ. Dif. 50:1039-1043), which represents an interesting novelty in the psychometric assessment of chronotype, because it comprises not only an assessment of the morningness-eveningness (ME) dimension, but also a distinctness (DI) dimension (i.e., amplitude), which represents the range of diurnal variation. In study 1, we aimed to confirm the structure of the Chronotype Questionnaire, with two different samples, young adults (n = 338, mean ± SD = 18.70 ± 1.12 yrs, 244 men and 94 women) and old adults (n = 477, mean ± SD = 55.92 ± 11.9 yrs, 168 men and 310 women). The confirmatory factor analysis (CFA) indicated a poor fit in both samples as well as in the whole sample, which could be due to some inconsistencies in the original instrument, above all in the distinctness dimension. We therefore decided to revise the Chronotype Questionnaire, keeping the ME dimension, but refining the DI dimension. In study 2, with a new sample of 197 participants (mean ± SD = 22.71 ± 2.23 yrs, 105 men and 92 women), we examined the factor structure of the revised scale containing 18 items. The resulting questionnaire contained 16 items (i.e., 8 items on each scale), with item factor loadings higher than .45. In study 3, we aimed to confirm the factor structure of the instrument developed in study 2 as well as to examine its convergent validity, with a new sample of 158 participants (mean ± SD = 55.92 ± 11.9 yrs, 97 men and 61 women). Results of the CFA showed that a good fit of the model could be obtained with 16 items in the questionnaire. The new questionnaire derived from the original Chronotype Questionnaire was from now on called the Caen Chronotype Questionnaire (CCQ). Convergence validity was obtained with the Horne and Östberg questionnaire and the ME scale of the CCQ, showing a negative significant relationship (r = -.82). The CCQ showed promising psychometric qualities, and further research should aim to combine it with physiological variables.