| Scholars and international nongovernment organizations (INGO) have developed concepts of democratic governance (DG) widely used by domestic and international agencies to evaluate a country's implementation of DG. These broad, global concepts are not always observable, collectable, measurable, or diagnostic in application. The purpose of this study was to investigate whether Bayesian belief networks (BBN) are a viable tool to evaluate a country's implementation of DG. Dahl's and Mainwaring's conceptualization that democracy is not a dichotomous state but a spectrum served as the theoretical framework for the study. The research question focused on whether BBNs could represent this spectrum in a mathematical model via conditional probability tables. The research design consisted of selecting a purposeful sample of 5 countries and then developing, populating, and operating BBNs around the 2 DG variables of "free and fair elections" and "multiple political parties." Study results indicated that BBNs are a viable tool for evaluating specific observables of DG implementation. Hypothesis testing also identified attributes of DG that are neither diagnostic nor observable, and suggested the need for independent observers to review not just elections but also a country's actual implementation of its legal framework for DG. Implications for positive social change stemming from this study include the opportunity for INGOs to use the BBN methodology in program decision-making and to access reliable diagnostic measures of cultural differences and how they may affect the implementation of DG. |
