PSYCHOMETRIC DATA FOR ENLARGED VERSION OF THE SLEEP-WAKE PATTERN ASSESSMENT QUESTIONNAIRE (SWPAQ)

Mike A., Onischenko

Psychology Faculty, Novosibirsk State University, Novosibirsk, Russia

Objectives. The inheritance of human psychosocial traits was shown in several researches, i.e. in the study of Ginsburg et al. (1987) on the example of the Cattell's personality questionnaire. The main purpose of my study is to measure the inheritance of individual traits related to everyday sleep-wake habits. The instrument for multi-dimensional assessment of sleep-wake pattern was developed by Putilov (1990, 1993, 2000). His questionnaire exists in two - original and enlarged - versions. The former consists of 10 tetrads (the groups of 4 closely related statements), the latter includes 3 additional tetrads. The 5-factor solution of the original version, accounting for 39% of variance explained, separated groups of items referring to night sleep (1st), morning lateness (2nd), evening lateness (3rd), wakeful ability (4th) and sleep ability (5th factor). Consequently, the questionnaire tetrads were divided into 5 groups labeled as S-, M- and E-scales (the tetrads s1, s2, s3, m1, m2, m3, e1, and e2, respectively), and w- and f-sub-scales (the tetrads w1 and f1, respectively). Thus, unlike other published chronobiological questionnaires the SPAQ allows measurement of 5 individual traits of the sleep-wake behavior instead of one-three traits. However, it was found (i.e. Putilov, 1993) that although Cronbach's alpha was high for the scales of SWPAQ (0.78-0.81), it was rather low for sub-scales of the original questionnaire (0.54, 0.57). To improve the internal consistency for the sub-scales, two new tetrads (w2 and f2) were added in the enlarged version. Besides, one more tetrad was added to E-scale (e3) to separate further the respondents' preferences for morning or evening lateness. On the present stage of my research I analyzed psychometric data on the enlarged version of SWPAQ to compare them with data on the original version.

Method. The Russian language questionnaire was filled by 321 residents of Siberia (these were mostly the students of the Novosibirsk State University and members of their families). Conventional psychometric analysis was applied to compare the original and enlarged versions of the questionnaire on their factor structure and on internal consistency of the main scales and sub-scales.

Results. The Chronbach's alphas for the scales S, M and E were similar high for both versions. For the enlarged version, the subscales' alphas were higher than for the original version and similar to that for the scales (see Appendix). The original and enlarged versions were similar in several respects. First of all, the factor analysis revealed three factors corresponding to S, M and E scales. Moreover, two w-tetrads were similarly loaded on two factors (M and E), and f-tetrads were partly loaded on S-factor. Nevertheless, there were some differences between two versions. Some of m-tetrads of the enlarged version instead of loading on only one factor (M) were loaded on two or three factors (M and E for m3 and M, E and S for m1). Besides, f-tetrads of the enlarged version were loaded not only on S-factor, but also on E-factor.

Discussion. By adding new items, the psychometric features of the enlarged version were improved considerably in terms of internal consistency of the sub-scales. Besides, the factor analysis of the enlarged questionnaire showed that, in general, the factor structure of this version remains similar to the factor structure of the original version. However, the factor structure of the enlarged version does not look like being as clear as the structure of the original version. An analysis of a bigger sample may help to clarify whether this difference is caused by selection biases or it is a stable feature of the enlarged version (i.e. due to adding new e-tetrad).

Conclusion. The enlarged version of the SWPAQ might be recommended as a questionnaire tool to study inheritance of several individual traits of the sleep-wake behavior.