Methodology
DATA COLLECTION
The Preaching Goes Viral project began in February of 2020 by Professor Dr. Hillel Gray at Miami University in his Religion 101 and 286 classes. Originally, students, staff, and volunteers submitted materials, including the metadata and audio/visual media if applicable, one at a time through a public Google Form. That data was stored on a Google Drive and uploaded in batches to the online software Dedoose where it was then transcribed, coded for themes, and analyzed. Early transcribing work, coding, and data analysis was done in large part by Miami University students. Early data collection was done through Google searches such as “‘covid-19’ ‘sermon,’” as well as combining through newspaper articles, social media, and religious databases. The intention of the project was to collect data from a wide variety of sources, but our focus began with Primitive Baptist and Hasidic communities, which we were already familiar with. The project then expanded as we explored other religions and denominations, and tried to alleviate gaps in denomination, religion, and geographic location in our work. Since a search for religious and COVID-19 related sources was fairly specific to begin with, researchers found success in casting a large net and allowing sources to naturally lead them to similar material.
Once submitting responses one at a time became too exhaustive, web scraping was used to collect large batches of material from databases and Sermon agrigators. In conjunction with the ongoing submission of individual items through the Google Form, PGV Staff used a web scraping tool called Parsehub to collect large numbers of posts, articles, sermons, and updates. The websites chosen to be web scraped came from a list generated by staff that centered on sermon aggregators, large churches, international churches and denominational websites, for example, with the Episcopal church (https://episcopalchurch.org/). Using Google Sheets to track the web-scraped data, researchers were able to identify the religious communities and denominations that were underrepresented, and keep track of the websites that had already been collected from.
The Summer 2020 interns, who began their involvement with the project in June, also contributed large amounts of data. Often areas of focus were much more specific and largely aimed at collecting from communities underrepresented in our database, such as LGBTQ+ religious communities, or African immigrants in Maine. Interns synthesized and analyzed the data they had collected and created digital exhibits on Omeka and Dovetail to present their findings.
DEVELOPMENT OF THEMES
Developing and identifying common “themes” was the central component of the project’s qualitative data analysis. To do this for audio-visual materials, transcriptions were essential - initially done by hand in Dedoose, and later, transcriptions were able to be automatically populated in Dovetail. In reading through materials, students and researchers were able to standardize a list of recurring themes in the data, tag materials with the applicable theme or themes, and perform data analysis. Coding the data with themes such as “the pandemic as punishment from God” or “collective responsibility,” illuminates thematic patterns in the data for specific religious sub-sects and across denominations. The research platform, Dovetail, gave researchers the capability to individually code sentences and paragraphs within one single given source, and sort excerpts by their tags.