Context
This analysis uncovers COVID-19 themes from data taken out of official websites and social media profiles of large Evangelical denominations, with a focus on Brazilian Evangelical churches during the first semester of 2020.
I collected the data set used for this analysis during the Summer of 2020 for the Preaching Goes Viral project. PGV focused on archiving online religious responses to the pandemic; this data set in particular focused on Brazilian Evangelical churches.
For those interested, to achieve that I used a Latent Dirichlet allocation algorithm, which creates a probabilistic topical model.
Analysis
Topic 3 was the first to stand out to me. It says “psychology”, “patient”, “world”, “faith” and “anxiety”. The topic seems to refer to mental health, with a large preponderance of mentions to “faith” and “anxiety”. We cannot infer with certainty the mechanics of the association between the two words within this topic in the dataset, but we have reasons to believe faith and anxiety were presented in some proximity, either by contrast, association or even causation.
Topic 6 also immediately stands out, with “health”, “problem”, “president”, “country” and “opportunism”. Topic 6 seems to be largely associated to national politics. It is reasonable to sense a sense of dismay in relation to politics, but again the discursive mechanics is not graspable with this method. For example, “opportunism” and “president” could be used in some sort of conflation, in a criticism of Bolsonaro, or of opposition, in a criticism of challenges against him and hope for more national unity.
Topic 8 shows us “temple”, “glory”, “service”, “heart”, “house”. The underlying relationship among the different words seems to be in-person services and social distancing.
Topic 9 mentions the biblical king Jehosaphat and “storm”, which seems to reference the story in which said king won a war through deference to God, fasting and prayer. This topic seems to point to a particular biblical exegesis emphasizing unity and deference to authority. It is uncertain who are the authorities one should be deferential towards though: public health specialists, Bolsonaro or the pastors?
I was not able to make sense of Topic 1, Topic 2, Topic 4, Topic 5 and Topic 7 also uncovered by this model.
Conclusion
This method allowed me to detect themes of: a) faith and anxiety; b) dismay about politics; c) in-person services and social distancing; d) unity and deference to authority. The limitations of this analysis, however, reside in not being able to know the specifics of how these themes were articulated by specific churches in the dataset.
Comments