Modeling Sustained Educational Change With Panel Data: The Case for Dynamic Multiplier Analysis

Abstract

This article considers the problem of modeling sustained educational change via the use of dynamic multipliers applied to panel data. Dynamic multipliers arise from the incorporation of lagged endogenous variables in linear models. Three types of dynamic multipliers can be defined: (a) the impact multiplier, (b) interim multipliers, and (c) the long-run equilibrium multiplier. An impact multiplier gives the effect of a unit increase in an exogenous variable on an endogenous variable in the particular sample period. An interim multiplier gives the effect of a unit increase in an exogenous variable on an endogenous variable when that effect is sustained for a specified amount of time. A long-run equilibrium multiplier gives the effect of a unit increase in an exogenous variable on an endogenous variable when sustained into the indefinite future. This article seeks to develop and advocate dynamic multiplier analysis for education research. Extensions to multivariate dynamic linear models and multilevel linear models are provided. Three examples are presented to illustrate the methodology. The article closes with a discussion of the implications of dynamic multiplier analysis for education policy analysis.