Discussion

My initial belief was that in 1924, the database exhibited a higher usage of shellshock over “shell-shock," since I seemed to observe usage of “shellshock” over ‘shell-shock”  peaking in 1930. Similarly, there seemed to be a greater usage of “shellshock” over “shell-shock” in 1942, 1943, 1944, and 1946. When clicking on the node for 1946, usage of “shell-shock” seems to be collating both “shell shock” and “shell-shock” as the same search term.

However, this form of searching may be providing a model of the database itself, showing where it may have more textual data to measure. To test this, I tested the Arclight search through a variety of terms, including "war" and "Hollywood" alongside "shellshock. Instead of producing the same results, the Media History Digital Library exhibits a high amount of hits for "war" around 1915 (around the time of the First World War) and spikes even further during World War Two. We see a somewhat unpredicable model for the term "Hollywood," but it does seem to rise and then fall across the corpora.