The Television News Archive (TVNA) includes over 62,000 hours of broadcast television news. Over 90% of the archive footage is described and cataloged in an expansive database with details the contents of the media down to the news segment and commercial break level. Researchers can access the database through the TVNA search page.
Basic Search
When the Advanced Search features are collapsed, users enter a basic search in the single field. Users can use. Multi-word terms match as an exact phrase (e.g. “Anne Thompson”). A single word matches as a prefix (e.g. “Czech” matches Czechoslovakia, Czechia, and Czech Republic), and commas separate terms that must all appear in search results. Basic searches will check all titles and descriptions for matches, without any limits on date, network, or type of content. Basic searches do not include transcripts.
Advanced Search
The advanced search works similarly to basic search, but with more options for narrowing the scope of the search, as well as enabling to transcript search. Users also have boolean search options with OR and NOT fields to build more sophisticated search combinations.
Filters and Facets Explained
Date Range: Limit your search to a specified date range. TVNA does not hold any content prior to August 5, 1968. New content is added monthly with a minimum of one-month delay for extensive content coverage.
Record ID: This search field overrides all other selections and returns the segment or broadcast that matches the unique ID provided. Users can search the new or legacy ID indexes.
Title & Description: These two fields contain text written or generated by TVNA, and are best used to find content that most likely about the subject, event, or person in question.
Transcripts: The full text of all speech included in the segment or broadcasts. Transcripts are generated using speech-to-text technology and can include spelling mistakes. Due to copyright restrictions, transcripts cannot be displayed in full online, so at this time a user cannot see the context of the search term within the transcript.
Broadcast Types: Users can include or exclude the various types of broadcasts. See the Data Dictionary for a description of the different broadcast types.
Segment Types: As of 2026, only evening news broadcasts are split into segments, so these facets only apply to evening news. The most common use of this filter is to exclude anything but news or commercials in the search. For most searches, we suggest excluding previews, teases, and good nights especially in transcript searches because these segments might mention a search term, but the content will not feature the subject.
Network: This facet can be used to only include specified news networks in the search: ABC (since 1968), CBS (1968), NBC (1968), CNN (1994), Fox News (2004), MSNBC (2024), and NewsMax (2024). If other networks or stations are listed, then content exists for that network, however coverage will be limited to specials.
Contributors, Locations, and Topics: These additional subject and vocabularies are not currently fully populated across the entire collection and thus not reliable for actual filtering yet. TVNA is creating workflows for adding these labels for new content and creating projects for building the dataset for the backlog. This work is scheduled for completion by Fall 2027.
Versioning and Updates
The TVNA database should be considered an ever-improving data set. Search counts, topic frequency, or failed searches can vary over time as we routinely correct spelling mistakes and apply new data enhancements to increase precision and recall. We also are actively digitizing historical video tapes, and processing new series of content. At this time, we are not maintaining a version library of the data, to recreate past search counts, etc. If your research relies on trends, mentions, or word counts, contact the McGee Applied Research Center for Narrative Studies for support and consultation on computational research.