This paper assesses empirically the role of household, community. individual and regionalrnfactors in determining school en rollment in Ethiopia. It uses logistic regress ion andrndiscriminant analyses methods to analyze the data. The analyses is done in such a way thatrnfirst determinants of primary and secondary school enrollment are studied and then thernoverall school enrollment is studied. Finally appropriate diagnosis is done for each of thernmodels.rnThe results from the analyses showed that age and sex of the person, relation of the person torn'---._.- -rnthe head and spouse (common child or not?), location of res idence (urban/fura!), educat ionrnand occupation of the head and the spouse, the interaction of sex and local ion of residencernand the interaction of sex and distance to primary school, were important determinants ofrnsc hool enrollment. Distance to school. number of school age people in a hou sehold. sex ofrnthe head, proportion of gi rls in the household, the logarithm of school expenditure, return 'srnfrom education and the inte raction of education and occupation of the head were foundrnimportant determ inants of primary schooling in addition to those listed above. Moreover,rnsecondary schooling is found to be a function of education of the head, urban ru ral residencernand size of school age people.rnIn general, the resuhs of data analysis demonstrate the use fu lness of applying multivariaterntechniques to the analysis of inter corre lated variables. The tech niques prove 10 be effectivernto the extent at least 72% in discrm inant analys is and up to 83% in logistic regression of thernstudents (cases) are correctly classified by the computed.mputed fancton.