Tripodi's subtitle calls attention to the central role of algorithmic manipulations in today's media environment, but her account is informed by history as well as her own ethnographic observations, so recent high-tech manipulations are situated in a much deeper and broader context. In 2017, Tripodi writes, she set out "to understand how conservative voters sought out information they felt they could trust. ... My goal was to better understand how Trump voters made sense of the contemporary news environment and how search engine optimization might play a role."
To research this, Tripodi immersed herself in with two representative groups in Virginia. "I had no intention of studying extremism," she writes. "I had no idea that the way information is tagged and categorized would take me into a media ecosystem fueled by conspiratorial logic. I did not expect the content in which I immersed myself to influence my own mindset, and I certainly did not expect to witness the violence of the Unite the Right rally."
She was, in short, greatly surprised by what she found: "Quite frankly, I did not realize how bad it already was, how bad it still could get, and how vulnerable we all are, myself included."
Yet there's a sure-footed quality to "The Propagandists' Playbook." However covert, sweeping and powerful the manipulations Tripodi explores may be, they do not disorient her account, and they need not disorient the rest of us either — with the help of her clear-eyed analysis. The book is organized as a set of seven "steps," and I chose largely to follow the chapter-to-chapter thread for clarity's sake in my conversation with her, which has been edited for clarity and length.
You organize "The Propagandists' Playbook" in a set of seven steps, and the first one is a commandment to "Know your audience." In that chapter you describe "the five F's of conservatism." What led you to that formulation, and what are they?
This book is based on research. I'm an ethnographic researcher, and I did months of research and interviews and content analysis of the news and information that people rely on. So I used grounded theory to identify pertinent themes and trends, and then based on those category I created this construct of the five F's of conservatism. I asked people to describe what they mean by "conservatism," and their definitions centered around these concepts over and over again, and these concepts were also central to the news and information that they were reading. So the five F's that I describe are faith, family, the armed forces — which constitutes the military and the police — firearms and a free market.
The second step is: "Build a network." As you describe it, this network has a long history, going back to the early days of radio, but only moved beyond radio in the 1980s and '90s as a result of the rise of televangelism and Reagan era deregulation, including repeal of the Fairness Doctrine. First of all, what is this network like today?
What's really important about understanding the network of right-wing information is that it's not just in one space. There's a lot of great research that looks at television or at ways that news and information travels online or thinks about YouTube or social media. What I demonstrate in my book is that these are highly interconnected forms of information, and that this has been going on for some time. Things that they write about in their news coverage or in books they would talk about on radio, and then that became television, and because they had a lot of practice at building this network, adding the layer of internet information was not too challenging. So it looks a lot like what it looked like since it started, with the exception that they have adapted to the ways that people get news and information in the 21st century.
Second, you studied the network through immersion in it, and you identified what you called two central conspiracy theories as a result.
One is that the left is dangerous — that those on the left are increasingly intolerant, scary, dangerous and disruptive to society. The other is that the media works in tandem with the left, and as a result the traditional media cannot be trusted. What's fascinating is these are also not new conspiracies. I show in my book how these notions of media distrust have been around since conservative media started, but the fear of the left being increasingly dangerous was really focused and emphasized in the 1960s, during the civil rights movement.
The third step is to "Engage in their form of media literacy," which you describe as founded in "scriptural inference." What does that mean, and why is it so central to how conservatives make sense of the world?
One thing I think is really important about my book is showing that conservatism is not just a worldview, it's also a media practice. Specifically, it's a form of media practice that leverages individual interpretation and emphasizes direct engagement with the literature. And whether that be the Bible or the Constitution or the Federalist Papers, or the memo that Trump released when he was being impeached the first time — his memo with [Volodymyr] Zelenskyy — this call to engage directly in the text is really rooted in the Protestant Reformation, specifically the Protestantism formed within the United States, which was about elevating individual interpretation in favor of an expert telling you what to do.
This form of media literacy, this way of interacting and engaging with the media is also the way conservative media talk to their audience. They don't just say, "Trust us, we know what we're doing." They actually activate this form of active inquiry, they utilize hermeneutical methods in their newscasts. You'll see this as a regular strategy that Tucker Carlson plays out. He'll put the quotes behind him that he wants people to focus on, and they leverage that form of media literacy. That's so important because it's different from the way other people, including progressives, engage with the media.
Step four is "Understand how information flows," and step five is "Set the traps." These are clearly closely connected, as you write that "Conservative elites … leverage a niche understanding of SEO strategies and methodologies to maximize the exposure of conservative brands, causes, and content." I'd like to ask about two specific examples you describe, and what they show us about the general strategies. The first involves Nellie Ohr, who was used to portray Trump as a victim of an attempted Democratic coup. What happened there, and what general strategies were involved?
Nellie Ohr is a great example of what I describe as "keyword curation" and "strategic signaling." The first part of understanding how information flows is not exclusive to conservative content creators. It's a basic understanding of how algorithms work, and what's important about that is to recognize that inputs — your keywords — are driving the output that any search engine's going to bring back to you.
How did Nellie Ohr, the wife of an obscure Justice Department official, become a "curated keyword" used to "perpetuate a conspiracy theory about an attempted coup" against Donald Trump?
So "Nellie Ohr" was this curated keyword that was adopted and essentially created leading up to and during Trump's first impeachment. Keyword curation works by relying on what scholars refer to as a data void: When little to nothing currently exists online, that hole or that gap can be easily filled with other content. Nellie Ohr is the wife of Bruce Ohr, who was a Department of Justice official at the time of Trump's impeachment. But because she worked at Fusion GPS — and Fusion GPS was behind the now clearly poorly-researched dossier — they created this whole narrative that the impeachment surrounding Trump's desire to have Ukraine interfere with the 2020 election was a way of unseating this president who was rightfully in power.
So a series of articles were written about Nellie Ohr, exclusively within the right-wing media ecosystem, and they all linked back to each other. A lot of them used the same copied-and-pasted text and made the same allegations, and then those same allegations were then covered by more mainstream outlets like Fox News. So during the impeachment trial, Rep. Devin Nunes used his time in his opening remarks to say, "We shouldn't be paying attention to this — what we should be paying attention to is Nellie Ohr." By activating this phrase, people were like, "Who is Nellie Ohr?" Then you go to Google and search for Nellie Ohr and the only thing returned is these conservative information systems that are perpetuating this conspiracy theory about an attempted coup to take out the president.
Could you say a bit more about the creation of data voids? I think that's a concept people are not generally aware of.
This comes out of Microsoft research: The notion of data voids is that sometimes there's not much existing on the internet around a subject or phrase. So data voids can get filled for a variety of reasons. Some of them can be filled by news coverage, for example. When a mass shooting happened in Sutherland Springs, Texas, no one had ever written about that town, and it was essentially a void: a Zillow listing and information about the population. So these voids, especially when there's a news event, are really ripe for bad information, because people are trying to get things out as quickly as possible and mistakes can happen. So that's one way a void gets filled.
The other way that voids get created and filled — we see this a lot in advertising — is that if you're trying to sell a product, you want to create a name for a product that doesn't already exist. Otherwise, if people search for your product, they're going to get the more established product. So they're taking this concept from advertising and applying it to news. So the data void is tied to problematic information in that if nothing exists online, it's easy to fill it with a bunch of information, especially if you have an already existing network of content creators.
The second example I'd like to ask about is the pushback against Black Lives Matter, which was a process in several steps. What happened there?
Black Lives Matter was the creation of activists who were trying to demonstrate the unfair treatment of Black people in the United States, in particular when it comes to crime and policing. What's fascinating is that you can see, using Google Trends data, that a way to respond to Black Lives Matter was to create alternative hashtags that could compete with it. So after #BlackLivesMatter rises you see the creation of #AllLivesMatter, which was trying to use this colorblind concept that everyone's equal so all lives should matter, not "just Black lives." Then that turned into #BlueLivesMatter, a catchphrase created to support the police and the armed forces, and not only did that activate the five forces of conservatism, but it also began to trend, it became a quick response to #BlackLivesMatter. We can see that it was created in response, because #BlueLivesMatter didn't exist before #BlackLivesMatter, according to Google Trends.
What did that creation sequence prove or demonstrate? You draw some conclusions — could you talk about the insights you gained from observing that?
A lot of times people will say, "This has nothing to do with #BlackLivesMatter, this is just talking about how these lives also matter." What you can see from the data is that if this was not a response to something, then it would have been created simultaneously with, or even before, the Black Lives Matter hashtag. The fact that it was lifted off the "lives matter" mantra and then appropriated for various groups activated those terms again whenever a Black person was killed by police. They were showing up very clearly in response to Black Lives Matter hashtags following extreme instances of police violence toward a Black person.
Want a daily wrap-up of all the news and commentary Salon has to offer? Subscribe to our morning newsletter, Crash Course.
This example of appropriation makes me think back to something else you wrote about. Another example was how "feminism" tags were used to spread conservative ideas, and that conservative sites often ranked higher in searches for "feminism" than liberal ones. What was going on there?
So that's looking at YouTube videos. I was trying to answer the question, "Don't people on the left also do the same thing as people on the right?" That's a great question, an important one. In order to answer that, I worked with a data scientist: He wrote a script and we looked at the top 10 content producers on YouTube from the left and the top 10 on the right. So we looked at the channels of people with millions of followers, and his script looked at how the content creators were tagging their content. So this wasn't how YouTube was tagging their content, it was about how the creators themselves were tagging their content.
Tags are important, because algorithms aren't people: They read in tags, they read in metadata. The tag is important because it helps an algorithm attach significance to content. It says, "Oh, you're looking for 'feminism'? Oh, this says 'feminism' — this is a match."
What we found when we looked at conservative content creators and progressive content creators is that content creators on the left had no idea how tagging worked, and they used very literal or strategic tagging, I guess you'd call it. They'd have these very literal tags that described what their content was. But conservative content creators recognize, "Well, some people might be looking for this stuff, and if we're trying to push back against these ideas, we need to also tag our content this way." Prager University, for example, which runs a conservative YouTube channel, has more videos tagged as "feminism" than tagged as "conservative." This demonstrates that they just have a more nuanced understanding of how keywords and tagging work than content creators on the left.
I asked about that because it seemed parallel to the appropriation of the "X lives matter" theme.
Absolutely. We didn't look at that tag specifically, "Black lives matter" or "Blue lives matter," but the appropriation of keywords — taking a concept that doesn't actually belong to you, but you're pretending that it does through metadata, then your content is going to be associated with that tag, even if it has absolutely nothing to do with that tag.
Step six is "Make old ideas seem new," which is particularly focused on how discredited racist ideas have been reintroduced. That adds another dimension to what we've just been talking about. Then step seven is "Close the loop." You describe the example of PragerU videos: "By providing textual evidence out of context, these videos invite conservatives to think critically about lines of text provided, but not question the broader cultural narrative in which those texts were created and now exist." How does that apply to the example you explore of how conservatives have subverted Martin Luther King Jr.'s message?
Conservative content creators have galvanized around a single phrase lifted from the "I Have a Dream" speech that allows them to take all of Martin Luther King's work out of context.
A huge number of conservative content creators have galvanized around the phrase lifted from the "I Have a Dream" speech, that King had a dream that his children would be judged not by the color of their skin but by the content of their character. So by focusing in on this one very specific line, it allows these political elites — whether that be media pundits or politicians — to take all of Martin Luther King's work out of context. King very much advocated for civil rights under the notion that Black people were not being treated equally in the United States, and I find it interesting that he's now being used as an example of conservative embodiment, when at the time he was classified as a Communist threat and was monitored by the FBI as a potential domestic terrorist. It's a classic example of taking one line and pretending it means something that it does not.
How does this fit into the framework of "closing the loop?" What do you mean by that, and what does it tell us?
So the "loop," I think, is two things. One of the things I describe in the last chapter to close it all together is the cyclical nature of these narratives about outside agitators and radical leftists, which have been around for a very long time. I show how this well-worn path of disinformation has been flowing through this information landscape for the last hundred years.
The other thing I talk about is what I refer to as the IKEA effect of disinformation. Business scholars have found that when people put together low-quality furniture on their own, they're more likely to value it, and think that it's better quality than it actually is. The same tangible, do-it-yourself quality of saying, "Well, don't trust us, go online and Google it for yourself" — or "DuckDuckGo it yourself," whichever one they're saying — activates audiences to take part in this scavenger hunt, not really recognizing that because of the keywords that have been provided to them, specific returns are going to be provided to them, and that these have been written and vetted by those who are telling them to go out and do it themselves. So this is how the loop actually closes, and why it's all interconnected.
Finally, what's the most important question I didn't ask, and what's the answer?
One thing I'm worried about is that people will say, "Well, this is why I don't go to Google," as if it's their fault. While Google has its issues, I'm not a techno-apologist — it's definitely selling our data, there's problems with the platform — the information-seeking process, whether we go through Google or whatever search engine you choose, is ultimately going to return us largely the same information if we aren't critical about the keywords we start with. So one thing I think we need to be more mindful of is not thinking the fix is going to come from tech companies, but rather thinking about how the fix is contingent on different social interactions with these search engines.
So that's a message for consumers, but also for progressive producers.
Sure. People will say, "Isn't this happening on the left?" And I would say, "Sure, it could." Anyone can use search engine optimization. But it isn't, based on the data I have. It isn't happening to the same capacity. And then, part two is to be mindful. That was my dedication: "To the information seekers everywhere: be mindful where the journey leads." I think a lot of people go, "Be wary of what you're seeing on Facebook" or "Be careful of what you're seeing on Twitter" or "Don't trust what you're seeing on TikTok." So a lot of people will see things, and then go, "Oh, let me go find out." Then they'll take these same concepts and they'll go to Google, and often what's returned to them is the same bad content they saw on Facebook, Twitter or TikTok.
So if you're just kind of input-in/input-out, taking these same ideas and just searching for them, without really recognizing how that works or understanding that search engines aren't neutral arbiters of truth — if you're trying to make sure you're getting the right information, you need to take a little more time in assessing the quality of your sources, and you need to understand that Google is not a helpful librarian.
