Косова, А. Л. Chronotype, sleep length, and school achievement of 11- to 23-year-old students in northern European Russia / Borisenkov M. F., Perminova E. V., Kosova A. L. // Chronobiol. Int. - 2010. - Vol. 27. - P. 1259-1270.

1264 M. F. Borisenkov et al. Statistical Analysis We used Statistica 5.0 software for the statistical analyses. One-way analysis of covariance (ANCOVA) was performed using each factor (age, sex, place of residence, and achievement) as independent variables, whereas the rest of the variables were used as covariates. Sleep length and chronotype were used as dependent variables. Eta-squared (rf) was used as a measure of effect size; effect sizes between 0.01 and 0.05 were considered low, between 0.06 and 0.13 moderate, and those >0.14 high (Randier, 2008a). RESULTS Sleep Length The average (±SD) sleep length of the surveyed subjects was 7.48 ± 1.2 h. During weekends, children and teenagers slept ~2 h more than during the school week (9.16 ± 1.7 h versus 7.17 ± 1.3 h, respectively). Sleep length significantly depended on age; there was a decrease of 1 h and 34 min in sleep length from age 11 to 21 yrs (Table 2). No sex differ­ ences in sleep length were detected. Latitude of residence had a weak influ­ ence on sleep length of both children and teenagers. Type of settlement/ residence significantly influenced sleep length; urban teenagers sleep on average 46 min less than their village counterparts (Tables 2 and 3). Chronotype There was a significant difference in chronotype depending on age and sex of subjects. There was a 1 h and 53 min phase delay of the sleep- wake rhythm between ages 11 and 22 yrs (Tables 2 and 3). The rest- activity rhythm of boys was more phase delayed, by 23 min, compared to girls. The chronotype of teenagers significantly depended on geographi­ cal coordinates and type of settlement. There was a significant phase delay of the sleep-wake rhythm in the East-West and South-North direc­ tions and in urban compared to rural teenagers (Tables 2 and 3). A weak positive, but nonetheless, significant correlation between chronotype and time of sunrise was noted (r = +.39, p < .01). School Achievement The achievement of school children and college students depended on sleep length and chronotype (Tables 2 and 3). In students with low and mean achievement scores, sleep length was 12 min shorter, and in compari­ son to students with high achievements they had a 55-min phase delay of the sleep-wake rhythm. Chronotype exerted an approximately 2-fold stronger