Our Methodology

The web-based platform behind www.propwatch.org, is an online database that can isolate and catalogue video segments of propaganda from embedded video and be queried to produce examples for viewing analysis. Our underlying treatment methodology is inspired by the messaging strategy of "inoculation," a social psychological/communication theory which has been shown to be effective at protecting individual attitudes and beliefs against devices of mass manipulation.

Inoculation messaging is analogous to the medical inoculation process, whereby a subject is exposed to a threat in order to build up an immune response, which will be triggered again when a future threat is encountered. This not only bolsters the original recipient's resistance to immediate propagandist messaging, but also builds resistance to future challenges.

For inoculation be effective, the message receiver must first be able to identify the threat as real, which is why we showcase authentic content; since it's one thing to read the definition of a propaganda technique, it's another to see it being used in a national debate or cable news interview.

In order to measure the project's effectiveness, we've developed a series of assessments to determine a viewer's ability to detect propaganda. Based on our current controlled study data, as of July 2021, a participant's ability to identify propaganda improved on average 44.17% after 90 minutes of viewing selected content on the propwatch.org website. See more about our latest research here.

Transparency of Methodology

Our methodology for detecting propaganda in this type of media is unique. Once a video clip is analyzed and an instance of propaganda identified, the instance is recorded in the form of a "play-by-play" comment. All videos clips are analyzed by two analysts working independently on the same clip. After both analysts have completed their review, we then compare the results for instances where the same technique was identified at the same timecode by both analysts. The video clip is then reviewed by an editor. If the editor confirms a strong match to a glossary technique, the comment is added to the database. If not, the comment is deleted from the database.

For example, when identifying a statement that uses minimization- characterizing something that's difficult to deny as trivial or insignificant, before shifting the focus to "more important" things.
, there should be present a clear pattern of trivializing something that is being scrutinized, followed by a deflection to something the speaker promotes as being more important. When a match is detected, the comment that follows must clearly spell out the matter being trivialized, followed by identifying who or what the focus is being shifted to.

When identifying a statement as an appeal to compromised authority- insisting something is true because an expert or authority on the issue says it's true, when that authority is compromised or has a vested interest in the outcome.
, the comment that follows must indicate how the authority cited in the statement is biased or has a vested interest in the outcome, providing a link to a reliable source that details that bias or spells out a clear conflict of interest with the authority being cited.

The written comment itself must adhere to rigid syntactical guidelines, which are exclusive to each technique in the site glossary. If a comment does not meet this standard, it is revised or deleted from the database.

With that said, there isn't any actual scientific formula that is 100% objective for identifying propaganda. There are objective measures that can be devised and employed, but ultimately there is always some degree of subjective judgment involved when making these determinations.

Commitment to Nonpartisanship

In keeping with our commitment to nonpartisanship and fairness, we analyze all content using the same standard and let the evidence dictate our conclusions. We do not concentrate our analysis on any one side of the political spectrum.

As a result, we strive to present an equal balance of comments reflecting all sides of the political spectrum, featuring a current balance of 49.73% / 50.27% comments associated with Democrats / Republicans respectively. As this is a real-time rating, it fluctuates as new content is added to the database.