Simultaneous Feedback Models with Macro-Comparative Cross-Sectional Data

Abstract

Social scientists often work with theories of reciprocal causality. Sometimes theories suggest that reciprocal causes work simultaneously, or work on a time-scale small enough to make them appear simultaneous. Researchers may employ simultaneous feedback models
to investigate such theories, although the practice is rare in cross-sectional survey research. This paper discusses the certain conditions that make these models possible if not desirable using such data. This methodological excursus covers the construction of simultaneous
feedback models using a structural equation modeling perspective. This allows the researcher to test if a simultaneous feedback theory fits survey data, test competing hypotheses and engage in macro-comparisons. This paper presents methods in a manner and language amenable to the practicing social scientist who is not a statistician or matrix mathematician. It demonstrates how to run models using three popular software programs (MPlus, Stata and R), and an empirical example using International Social Survey Program data.