Link Lab

A look at the use of magazines, videos, blogs, and other nontraditional teaching materials — with some caveats.

Open Syllabus depends on bibliographic catalogs to describe the world of titles that can be found in its syllabus collection. Because we've gotten better at managing these catalogs, the new 2.5 dataset expands the range of distinct titles found from 1.7 million to 4.6 million.

Some of the contributing catalogs, such as the Library of Congress catalog, contain movies, musical compositions, and other non-textual materials. But they don't keep pace with the proliferation of web, video, and periodical resources available to teachers.

Over the summer, we made a catalog of these materials ourselves by analyzing the URLs in syllabi. This method has some clear strengths. It opens up the teaching world of YouTube for the first time. It can identify podcast episodes, blog posts, software tools, and web-based teaching resources.

And it has some limitations. It only picks up instances of titles that include URLs–which is not all of them. It introduces heterogeneity problems: URL conventions vary and not everyone follows the same rules for identifying titles, authors, publishers, or dates. It also balks at media organizations that hinder metadata crawls, including, for now, The Washington Post, Bloomberg News, and The Financial Times.

We used a mix of machine-classification and manual attention to clean these URLs (and their metadata) up. The results don't include YouTube for now and only list titles that appear on at least six syllabi. The resulting data is not yet part of the canonical OS dataset. That's why it's here, for now, in a ‘Lab.'