Testing the reliability of sentiment analysis as an indicator of public opinion on United States foreign policy: A study of public responses to President Donald Trump’s Twitter rhetoric on North Korea

This study examines the reliability of sentiment analysis as a method for gauging public opinion on United States foreign policy, particularly in response to former President Donald Trump’s tweets (now known as ‘X’) about North Korea. The research employs two methodologies: automated sentiment analysis using VADER within Python and a manual coding approach using N-Vivo software. The findings reveal that while sentiment analysis provides a generally accurate reflection of public perceptions, manual coding by hand offers greater precision, particularly in identifying nuances such as negative language embedded in ostensibly positive replies. This underscores the importance of complementing sentiment analysis with manual review for more accurate public opinion measurement, particularly in the context of volatile international relations. Applied to United States policy toward North Korea during the Trump presidency, the study highlights that Trump's most negatively received tweets often involved either aggressive rhetoric or premature claims of diplomatic success. Overall the results indicated that the American public looks for the president to navigate a rather narrow pathway on the North Korea problem: steering clear of threats or provocations that could light the fuse for a war on the Korean Peninsula and Northeast Asia while also avoiding concessions to North Korea unless the country’s leaders demonstrate without question a readiness to change.