We have gotten a number of requests over time to make information about article “types” available.  As of last week, for JSTOR journal articles (we are working on Portico journal articles), any newly generated metadata file or JSON-L will now include a docSubType (if we have it, which we usually do.)  You can also filter on this field using our advanced search in the keyword field in the dataset builder.

For example, say you want to research what is going on in frontmatter over the past couple of decades, you could do this search to build a dataset completely of frontmatter articles:


This field is not normalized, however there is a limited set of frequently used types. Below are a list of those which currently have over 500 articles:

research-article: 7,070,992 articles
book-review: 3,802,198 articles
misc: 2,169,329 articles
news: 139,885 articles
editorial: 18,465 articles
frontmatter: 6,409 articles
other: 5,996 articles
review-article: 3,129 articles
backmatter: 3,009 articles
brief-report: 1,280 articles
correction: 1,227 articles
front-matter: 769 articles
Letter: 723 articles
discussion: 561 articles

Please note, if you are working with a previously created dataset, this data is not included.  Just sent us your dataset ID and a request, and we’ll delete the files on the system – when you go back to your dataset in the UI it will rebuild and include this data.

If you use this new feature and data point, let us know!