Explanation of Mixed Results about the Reasons behind
North Korea’s Provocations:
Classifications and Applications of Multilevel Models
Byung-Deuk Woo1
Received: 15 September 2022 / Accepted: 11 March 2023 /
Published online: 1 June 2023 Ⓒ Korean Social Science Research Council 2023
Abstract After the breakdown of the 2019 Vietnam summit between the U.S. and North
Korea and the improved nuclear capabilities of North Korea, North Korea’s provocations have
been noticed by scholars again. Based on previous qualitative studies, researchers have
continuously conducted quantitative studies to examine North Korea’s provocations. The
increased number of available data makes it possible to apply sophisticated methods to unveil
the reasons behind North Korea’s provocations. However, results from recent studies are
inconsistent and even conflicting. Thus, this article aims to address possible reasons for those
mixed results. Moreover, this article suggests a direct and indirect classification and an
application of multilevel models to explain the conflicting results and enrich empirical studies
about the provocations. By applying multilevel Poisson and negative binomial models to
provocation data, this article finds that internal and external factors heterogeneously influence
the number of total, direct, and indirect provocations.
Keywords North Korea․provocations․direct and indirect classification․multilevel poisson
․negative binomial models