부산대학교 도서관

To reliable prediction of dependent variable, the temporal relationship, that is, whether the independent variable (cause) precede the onset of the dependent variable (effect) is necessary and without information on the temporal relationship it is hard to determine whether a factor is an independent variable or a dependent variable [[2]]. Although cross-sectional studies can be very useful for public health planning [[4]], they usually do not show the temporal relationship between the independent and dependent variables because these studies measure the associations between variables at the same time [[3]]. While, generally, to identify the most important independent variables, firstly, univariable logistic regression analysis should be performed for each variable and then, any variable with a I p i value <.25, be entered into the multivariate model [[6]]. [Extracted from the article]