I wouldn't say "believe" is the most accurate way to describe it. They're just very skeptical of the insights from statistical analysis of works in the mediums

I wouldn't say "believe" is the most accurate way to describe it. They're just very skeptical of the insights from statistical analysis of works in the mediums because there usually aren't very good large data sources. Cinemetrics.lv was amazing before the domain was stolen and had a large community of film academics using it. In literature, a big problem has been the way that publishers do not allow access to any of their data, some public access programs are starting to try to overcome that though. In general, though, most humanities academics aren't interested in statistical analysis or its insights, not because they don't believe in them, but because the insights generally aren't that useful or likely so general as to be not that useful for very specific research programs.