Topics and Attitudes in COVID-19 Online News Comments about Italy and Sweden. Appraisal Framework applied to Corpus-Assisted Discourse Analysis

Abstract: This study conducts a mixed-methods, comparative study of online news comments. The subject of the articles were the two countries of Sweden and Italy during the COVID-19 pandemic. Articles about Italy and Sweden during the pandemic were identified, and the comments posted to those articles were collected and divided into two corpora based on country. The two corpora in total consisted of over 500 articles, and over 4.5 million words. Several computer programs were written to identify that the articles selected were either predominantly about Italy or Sweden during the COVID-19 pandemic, and subsequently to collect the comments posted to those articles. Corpus-assisted discourse analysis was then used to conduct the analysis in two steps. First, the corpus-assisted part allowed for the identification of prominent topics of discussion in the two corpora. Secondly, the discourse analysis part took a sample of comments related to a selection of the prominent topics of discussion and used the appraisal framework developed by Martin and White (2005) to observe frequencies of attitude and polarity. The results showed among other things that topics such as herd immunity, self-isolation, vaccines, and lockdown were key topics present in the corpora and therefore represent topics of interest and importance to those writing the comments. The discourse analysis was able to show among other things that in the Italy corpus sample Judgement was the most common attitude, while in the Sweden corpus sample it was Appreciation. The frequency of Affect was the same in both samples, as was the overwhelming frequency of negative polarity present in both.