Earlier this year, I have published my first paper on semisupervised methods (Newsmap and seeded LDA) in Social Science Computer Review. My second paper on semisupervised method (Latent Semantic Scaling) has appeared in Communication Methods and Measures a few days ago. I wrote these research articles and developed software packages as part of my effort since 2018 to expand the scope of text analysis to new domains and languages.
Social scientists must respond rapidly to political and economic upheaval in the ongoing COVID19 crisis to understand its impact on many aspects of our lives. For example, conspiracy theories became even more prominent online during the pandemic to the extent that it hampers western governments’ responses to the COVID19 pandemic. For example, conspiratory beliefs would make people not only disobedient to behavioral rules to contain the spread of the virus but reluctant to take vaccination even when it comes widely available sometime next year.
I believed quantitative text analysis is useful in studies of unexpected social changes because it can applied to the unstructured data (e.g. newspapers and social media posts) that had been produced continuously. In fact, I have developed the semisupervised methodology trough a study of Russia’s anti-west propaganda by online news sites using conspiracy theories. I am currently applying LSS in my projects to measure people’s perceived threats of refugees and foreign countries in newspapers and extreme political ideology and mental health on social media. I hope that other social scientists to employ semisupervised methods to analyze the stream of unstructured data focusing on specific aspects of texts.
