ABSTRACT
Educational researchers frequently apply statistical modeling to answer research questions in education. A model may be described as 'a formal representation of a theory', or as 'a set of assumptions together with implications drawn from them by mathematical reasoning' or as, 'an idealized representation of reality'. This chapter describes statistical modeling in educational research, focusing particularly on modeling strategies that take into consideration the sampling procedure and classification of variables. Additional considerations are also discussed. The process of choosing participants from a population is called 'sampling'. In general, there are two ways of classifying variables: level of measurement and descriptive orientation. The choice of statistical modeling techniques depends on the number of independent variables (IVs) and dependent variables (DVs), as well as numerical characteristics of the variables; that is, whether they are continuous, discrete, categorical, or dichotomous. With a single continuous DV, general linear modeling (GLM) is used; with a discrete DV, categorical or dichotomous, generalized linear modeling is used.
