Taking Time Seriously: Time-Series-Cross-Section Analysis with a Binary Dependent Variable

Abstract

Researchers typically analyze time-series-cross-section data with a binary dependent \nvariable (BTSCS) using ordinary logit or probit. However, BTSCS observations are \nlikely to violate the independence assumption of the ordinary logit or probit statistical \nmodel. It is well known that if the observations are temporally related that the results of \nan ordinary logit or probit analysis may be misleading. In this paper, we provide a simple \ndiagnostic for temporal dependence and a simple remedy. Our remedy is based on the \nidea that BTSCS data are identical to grouped duration data. This remedy does not require \nthe BTSCS analyst to acquire any further methodological skills, and it can be easily \nimplemented in any standard statistical software package. While our approach is suitable \nfor any type of BTSCS data, we provide examples and applications from the field of \nInternational Relations, where BTSCS data are frequently used. We use our methodology \nto reassess Oneal and Russett's (1997) findings regarding the relationship between economic \ninterdependence, democracy, and peace. Our analyses show that (1) their finding \nthat economic interdependence is associated with peace is an artifact of their failure to \naccount for temporal dependence yet (2) their finding that democracy inhibits conflict is \nupheld even taking duration dependence into account.