Multivariate analysis was conducted using SPSS complex samples logistic regression analysis stratifying samples by district/archipelago and including as main effects all variables which reached an α level of at least 0.25.
Secondly, the influence of the duration of pain was studied using results of the total sample in regression analysis.
See Figure 1 for a sample logistic regression analysis of this data set using the statistical freeware package "R".
Descriptive statistics, ANOVA, and Pearson's correlation (r) were used to describe the sample and regression analysis to investigate the association between the predictive psychosocial factors.
To determine the scaling factor, for each sample, linear regression analysis was performed comparing the values for the reference gene (x) versus the average values for the reference genes across the sample set (y).
As with our ADHD GWAS, to allow for population stratification within the sample, the regression analysis is performed conditional on two covariates (the first two principal components estimated from our GWAS data using EIGENSTRAT software designed for this purpose).
To determine the contributions of epithelial tumor cells and of stroma cells in the gene expression profiles of macrodissected samples, linear regression analysis was performed on the twice-amplified macrodissected samples with their corresponding microdissected tumor and stroma samples (Table 3).
Statistics: Spearman correlation, Wilcoxon-test for unpaired samples; multiple regression analysis; IBM SPSS 22. Age 68 ± 12 years, 19 female, 23 male, APACHE-II 19 ± 8, SOFA 8±± 4. IAP was increased (≥12 mmHg) in 162/399 (40.4%) measurements, and patients were under mechanical ventilation in 262/399 (65.7%) of measurements.
In the study of all samples, logistic regression analysis showed that the presence of diabetes, as other studies found [19], was the most important factor in predicting inadequate myocardial reperfusion after angioplasty.
Concerning possible influence of protein degradation in the samples, linear regression analysis between time of biopsy after death and expression levels of ABCA1, ABCG8, NPC1L1 and FAT/CD36 did not show any association, whereas even a positive relation was found for ABCG5 (R2 = 0.25; P = 0.0001).
Complex samples logistic regression analysis was used to investigate associations between socio-demographic characteristics and the likelihood of suspecting cancer.
