1.1.4: An Aside on Method

In the process of writing this dissertation, I collected a series of around 700 books, pamphlets, reports, and journals and developed a digital corpus for sake of analysis (x.1.1). While doing the data entry for this material, I isolated and tracked each image in the corpus: each image was given an identifier ([the book identifier]_[pagenumber][a, b, c, d… if there were multiple images on a page]) and were defined using a set of criteria. Images were tagged based on media type—photograph, x-ray, illustration, and graphic¹—as well as being tagged with at least one descriptive note—anatomical, pathological, organ, equipment, child, professional and so on.

Not unlike content analysis, I wanted to be able to describe in broad terms some of the general trends in the larger corpus. I also wanted to develop a regimented way where I spent some time (even if only a second or two) with each image,² My interest however has less to do with some of the bigger data or social scientific uses of content analysis, as I did not follow such a rigorous, detailed methodology, nor do I expect that the material can be used to make predictive hypotheses about other kinds of material. As part of this digital dissertation, I have made the data and metadata from the corpus available, should more computationally interested scholars find use for it.

I lay this out because most of this chapter is a result of finding unexpected trends while working with the tuberculosis corpus, and because I wish to fortify my claims about the broader discourse.