Analysis and Discussion

In general, our results mirror the results of surveys like the Wiley Open Research Survey 2019 \cite{yudm4h}, where researchers in the Life Sciences most often report that they share data, with Medicine closely behind.
The source data associated and shared with Table \ref{133180} includes two files with domains and repositories for: higher-level WOL Level 1 categories; and for more detailed WOL Level 2 categories.  This data can be used to identify commonly-used repositories, as inspiration and direction for journal teams and researchers. Dynamic Figures \ref{692056} to \ref{144977}, discussed below, plot the source data for WOL Level 1 categories and for a selection of WOL Level 2 categories. Figures \ref{687558}, \ref{629953}, and \ref{500750} offer further detail about DOIs that resolve to the nih.gov domain. 
Figure \ref{692056} shows the large number of unresolvable doi.org links in submitted DASs. Our sample for this analysis only included DASs that indentified datasets (excluding DASs that report, for example, that data are available on request from the authors and similar). It is reasonable to expect submitted manuscripts to contain errors, including in links provided by authors. These errors may be the reason why so many links that appear in our analysis to be doi.org links do not resolve. Publishers and research authors correct errors and oversights through the publishing process. Providing clear information when it is needed, to support better data citation practices by research authors, is important \cite{mrqa8p}. The large proportion of unresolvable doi.org links in our data set may have implications for the observations we make below, which may change if we were able to resolve those unresolvable doi.org links.