Big, Small, and In-Between: Control Models In Regression Analysis

• In Event: 546 - Section on Sociology of Population Roundtable Session (one-hour).
  In Refereed Roundtable: Table 08. Demographic Methods and Population Estimation

Abstract

One of the fundamental problems that researchers address using regression analysis is to what degree is an effect that is associated with a particular independent variable due to the relationship between that independent variable and other independent variables. In regression analysis, we call these other independent variables control variables. While using control variables in regression analysis is common, there is little in terms of written explanation of the procedure. This paper fills a gap in the regression analysis literature by providing a thorough explanation of control modeling. The paper first explains how controls in regression work using the concepts of elaboration and demographic standardization. The paper then discusses a variety of approaches to control modeling including big and small models, the one-at-a-time approach, the step approach, and hybrid approaches.

Author

• Roger A. Wojtkiewicz, Ball State University