MULTIVARIATE LOGISTIC REGRESSION ANALYSIS Clause Samples
MULTIVARIATE LOGISTIC REGRESSION ANALYSIS. A weighted multivariate logistic regression analysis was employed to study the relationship between social support during pregnancy and postpartum depressive symptomology. The full model incorporates all sociodemographic, maternal, and pregnancy-related characteristics of interest from the previous literature, as well as survey year to control for potential differences in sampled characteristics between 2012-2013 and 2014. Additionally, based on the preliminary assessment, correlates were chosen as potential effect measure modifiers if there was evidence that the estimated effect between low social support and PPD differed between levels of the variable, indicated by a jointly significant interaction term at a p-value < 0.20. The results of the interaction assessment can be found in Table 4 of Appendix II. The variables included as potential effect measure modifiers were maternal education, marital status, abuse before or during pregnancy, and cigarette use during pregnancy. Therefore, the full model included the exposure, 13 correlates, and 4 interaction terms:
MULTIVARIATE LOGISTIC REGRESSION ANALYSIS. An analysis of the multivariate logistic regression showed that demographic factors that aggravated women’s sleep quality were chronic illness and having below average financial status, as mentioned in the literature review above. Being a female in above-average financial group was protective for them against being a bad sleeper. These two results related to wealth indicated that females in KSA might have been worried about financial issues, which could have affected them positively or negatively. For males, having a chronic disease or being young are aggravating factors to be a bad sleeper, while being a rural resident is a protective factor. Being a young male in KSA (25-44) may have stressors like taking care of old parents and small kids, as well as fulfilling duties for other members of the extended family. Also, since men are the only ones who drive, many young men are required to do the shopping and take relatives places in addition to their work. I think all of these factors combine to make them bad sleepers. However, being a rural resident may mean having less work and less stress, leading to better sleep quality. An analysis of sleep factors for both males and females showed similar results. The more time it took to fall asleep, the greater the chance of being a bad sleeper. The longer the person slept, the better the chance of being a good sleeper. Interestingly, time to go to bed and time to wake up were insignificant (i.e., were not related to sleep quality). As mentioned, the number of sleep hours has the most effect on sleep quality, so it may not matter when people went to bed or woke up if they had enough sleep hours.