Multilevel Modeling: Application for Life-Span Research

Multilevel modeling (MLM) is a type of analysis similar to multiple regression but more directly applicable to life-span research. In multiple regression, differences on some measure of interest are explained by several other measures (i.e., predictors) using a mathematical equation with (a) an overall intercept (when all of the predictors are at zero) and (b) slopes for each of the predictors. MLM also uses an equation with intercepts and slopes; however, MLM analyses have the distinct advantage over multiple regression in that they account for when scores cannot be assumed to be independent (an underlying assumption behind many other analytic approaches). For example, when measuring people over time, scores at one time point directly influence later scores. In another example, blood pressure in the afternoon ...