Operational Code Analysis: Assessing Leadership Behavior in International Relations

Lecture Center Concourse

In 2013, Donald Trump mused whether Vladimir Putin would become his best friend. Since 2016, this question has gained important implications for US foreign policy. However, how could we have predicted whether the two leaders would become best friends? Operational code analysis is one approach to understanding and predicting the behavior of foreign leaders, including Vladimir Putin. Early applications of operational code analysis (OCA), relied on human interpretive capabilities, subject to influence by author bias. In the 1990s the Verbs in Context System (VICS) for coding texts was developed for OCA to reduce author bias and produce more systematic analysis. More recently, computer/automated coding systems have been developed to enable more systematic and reproducible analyses. Automated coding quickly provides non-biased, and reliable data to compare and analyze documents--a distinct advantage over laborious human coding. For example, analysis of over one million words spoken by Russian President Vladimir Putin (Dyson, 2017) shows Putin’s leadership style is more of an opportunist than a strategist, and that he reacts hyper-aggressively towards terrorism. However, work is needed to increase the accuracy of the automated VICS coding scheme compared to ‘gold standard’, hand-coded phrases. The goal is to raise accuracy from the currently unknown level to the accepted level of above 80%. Thus far, the accuracy of an improved and extended automated VICS coding scheme has been improved from 6% to 59% and an initial analysis of international leaders illustrates the use of OCA.