Search results

Methods to automatically analyze media content are advancing significantly. Among others, it has become increasingly popular to analyze the framing of news articles by means of statistical procedures. In this article, we investigate the conceptual validity of news frames that are inferred by a combination of k-means cluster analysis and automatic sentiment analysis. Furthermore, we test a way of improving statistical frame analysis such that revealed clusters of articles reflect the framing concept more closely. We do so by only using words from an article's title and lead and by excluding named entities and words with a certain part of speech from the analysis. To validate revealed frames, we manually analyze samples of articles from the extracted clusters. Findings of our tests indicate that when following the proposed feature selection approach, the resulting clusters more accurately discriminate between articles with a different framing. We discuss the methodological and theoretical implications of our findings.

Content analysis of political communication usually covers large amounts of material and makes the study of dynamics in issue salience a costly enterprise. In this article, we present a supervised machine learning approach for the automatic coding of policy issues, which we apply to news articles and parliamentary questions. Comparing computer-based annotations with human annotations shows that our method approaches the performance of human coders. Furthermore, we investigate the capability of an automatic coding tool, which is based on supervised machine learning, to generalize across contexts. We conclude by highlighting implications for methodological advances and empirical theory testing.