Spss statistics 22 for mac

Continuous variables are reported as median (25th-75th percentiles) and categorical variables as numbers (percentages). Where distributional assumptions were violated, measures were either log10 transformed (kynurenic acid, anthranilic acid, 3-hydroxykynurenic acid, xanthurenic acid, 3-hydroxyanthranilic acid, quinolinic acid, pyridoxal 5`-phosphate, riboflavin and N¹-methylnicotinamide) or ranked (CRP) before used in parametrical tests. To assess changes in baseline characteristics between inclusion and one year after surgery Friedman´s test was used. Changes over time between repeated end-point measures were assessed with a random intercept mixed model, adjusted for operation method and multivitamin supplement use. Correlations among continuous variables were assessed by Spearman rank correlation, corrected for age, gender and eGFR, as well as operation method when analysing correlations between changes over time. All tests were two-sided and p-values < 0.05 were considered to be statistically significant. Statistical analyses were performed using SPSS statistics 22 for Mac (IBM Corporation, New York, NY, USA).

IBM SPSS Statistics 22 for Mac (SPSS Inc., Chicago, Ill) was used for all analyses. Pearson's chi-squared test and one-way analyses of variance (ANOVA) with Bonferroni post hoc tests were used to examine differences in demographical, clinical and sleep variables between the four BMI categories “underweight”, “normal weight”, “overweight” and “obese”. Both linear and quadratic terms (weighted) were entered in the ANOVAs to show potential linear and/or U-shaped associations. Multinomial logistic regression analyses were conducted to examine the predictive effect of the sleep variables (independent variables) on BMI-categorization (dependent variable), using “normal range” as the reference category. Both crude and fully adjusted models were examined, the latter adjusting for the following covariates entered in one block: age, gender, parental education and family income, physical activity and depressive symptoms (SMFQ total score). Logistic regressions were also used to examine whether adolescents in the non-compensated group had a greater risk of overweight/obesity than the compensated group. All analyses were conducted on the whole sample as well as stratified by gender.

IBM SPSS Statistics 22 for Mac (SPSS Inc., Chicago, Ill) was used for all regression analyses. Multinomial logistic regression analyses were conducted to examine the predictive effect of the parental work absenteeism (independent variables) on 1) increased symptom complaints and 2) school absence in the adolescent children (dependent variables). Multinomial logistic regression analyses were also used to examine the association between number of weekly symptoms (independent variable) and level of school absence (dependent variable).
In order to assess whether adolescent symptom complaints mediated the association between parental work affiliation and school absence, indirect effects analyses were conducted within a structural equation modelling framework using version 0.5–16 of the Lavaan package [11 (link)] in R for Mac version 3.1.1 [12 ]. Maximum likelihood estimation with bootstrapped (k = 1000) standard errors and confidence intervals was used according to recent recommendations [13 (link)]. Incomplete responses were handled using pairwise deletion, ensuring high data retention in the analyses (92.6% in analyses with mothers and 88.4% in analyses with fathers).

Statistical analyses were performed in SPSS Statistics 22 for Mac (SPSS Inc., Chicago, IL). First, the DASS subscales and ZBI were specified as continuous variables. Then the ZBI was specified as a categorical variable using the cut-offs described above. For each outcome, the first regression model included only the ZBI. The second regression model also adjusted for potential confounding by caregivers' age, sex, education, and years of care; and by age and sex of the index patient, and the index patient's relationship with the caregiver.