An official website of the United States government

Official websites use .gov

A .gov website belongs to an official government organization in the United States.

Secure .gov websites use HTTPS

A lock ( ) or https:// means you’ve safely connected to the .gov website. Share sensitive information only on official, secure websites.

You are viewing ARCHIVED CONTENT released online from January 20, 2017 to January 20, 2021.

Content in this archive site is NOT UPDATED, and links may not function.

For current information, go to



MODERATOR:  Okay.  Good afternoon and welcome to the Washington Foreign Press Center’s on-the-recording briefing on Elections 2020 and the Fourth Estate: How Journalists’ Election Coverage Affects Voting Behavior.  My name is Jen McAndrew and I’m the moderator. 

First, I will introduce our briefer and then I will give the ground rules.  Today’s briefer is Dr. Sharon Jarvis, Associate Professor of Communication Studies and Associate Director of the Annette Strauss Institute for Civic Life at the University of Texas at Austin.  She will discuss how media coverage of elections (inaudible) stories make people want to vote, based on research in her book, Votes that Count and Voters Who Don’t.  We greatly appreciate Professor Jarvis for giving her time today. 

And now for the ground rules.  This briefing is on the record.  The views expressed by briefers not affiliated with the Department of State or U.S. Government are their own and do not necessarily reflect those of the U.S. Government.  Participation in Foreign Press Center programming does not imply endorsement, approval, or recommendation of their views.  We will post the transcript of this briefing later today on our website, which is 

Professor Jarvis will give opening remarks and then we will open it up for questions.  If you have a question, you can raise your hand virtually in the participant field or submit it in the chat field.   If you are called upon, you can unmute yourself and turn on your camera to ask your question.  If you have not already done so, please take the time now to rename your profile with your full name and media outlet.  And with that, I will pass it over to Professor Jarvis. 

MS JARVIS:  Thank you so much, Jen.  Thank you for this invitation, and thank you all for your time this afternoon.  I am delighted to be with you.  I’m going to take one moment to share my screen.  And Jen, may I see a thumbs-up?  Terrific.  So we’ll go ahead and begin. 

The title for this afternoon is “The Fourth Estate: How Journalists’ Election Coverage Affects Voting Behavior.”  And I’d like to begin on an interactive note.  What you see here on slide two, you will see again on a later slide.  But let’s all imagine a reporter, one of your colleagues, from The New York Times is on the phone.  She wants your take on the 2020 presidential election.  Please think for one moment:  What would you say?  What would you say in response to this question? 

Now, why on Earth start a talk on the public conversation surrounding voters in this way?  And I have three key points, and I’m going to go ahead and give them to you at the outset.  First, since 1972, in the United States, the reification or the glorification of political strategists and pundits has sidelined the role of voters in election coverage.  Second, voters in the United States have become the collateral damage of a battle between journalists and strategists to control the political conversation.  And then third, it hasn’t always been this way, and it doesn’t have to be in the future. 

How can I support these claims?  We’ll spend a lot of the time today during this briefing discussing findings from my recent book that came out in 2018, and then we’ll also address a grant that followed up on this on how to talk about threats to elections. 

But I’m going to begin with your audience with a little bit of context for the research.  And I think that laying some background is very helpful in this regard.  In the United States, voting is a contested method and symbol.  If we go back to our founding fathers, it’s remarkable and worth noting that they did not put an affirmative right to vote in the United States Constitution.  Subsequent democracies did put an affirmative right to vote in the constitution, in the founding documents.  What is more, democracies founded throughout the 20th century, many of them automatically registered people to vote upon a certain birthday, making that a default mechanism, and that is something that we simply don’t have in the United States.  We have centuries of ambivalence over the opportunity to vote. 

And that notion, the contested nature of voting as a method or a symbol, has led to a couple hundred years – few hundred years – where voters have few natural advocates.  And let me share more here. 

Politically, elected officials in the United States have engaged in strategies to make it difficult for folks they don’t want to vote and easier for folks they do want to vote.  And here’s an example of one barrier to participation from the state of Louisiana in the 1950s, where voters would have to take a literacy test in order to be able to vote, and we could agree if we spent time on this that many of these items don’t necessarily test a fitness for voting, they just represent a barrier. 

Additionally, candidates – and this is research that appears in the book – as recently as the 1940s and ‘50s in the United States talked about voting as a value, but more recently it’s become politicized.  And we see two quotations from two Republican nominating conventions where Eisenhower after World War II treats it as a value, and then we see Romney in 2012 really questioning if voting is helpful or harmful.  What is more – and this is very important for those of us who study civic education in government in the United States – our national school system as recently as the 1950s and 1960s used to teach young people how to vote, but much of that curriculum has been pulled back.  And so when we think about it, not only did the founders display ambivalence about who should vote, but our politicians, our candidates, and our schools have not contributed to a culture where, for the last 60 years at least, there are advocates for voting. 

And then when it comes to reporters – and I want to be incredibly generous here – voters simply can be challenging to cover, and here are a few examples from the United States:   

There’s a famous headline in 1948 where the Chicago Tribune wanted to be first with the story, and they went with polls that were faulty, and they declared Thomas Dewey the president – again, in 1948 – when indeed Harry Truman, the gentleman in this picture, actually was the gentleman who won when all the votes were counted.   

We saw a similar pattern in the 1964 presidential election where LBJ, Lyndon Baines Johnson, was so far ahead it was unthinkable for reporters to not share that he was going to win, and they announced that perhaps he would before folks on the western states could vote in the United States.   

We saw it again in 1980, and here we have two broadcasters where they called Ronald Reagan’s victory such that Jimmy Carter, who was the incumbent president in 1970 – 1980 – running against Reagan, actually conceded before folks on the west could vote.   

And we saw it again of course in the 2000 election where we had several projections go back and forth, and indeed we didn’t know the winner for quite some time in that year.  So voters are difficult to cover because news norms – and I’m teaching to a group of professionals here – the timeliness – the timeliness and the desire to have that scoop makes it difficult, particularly in the United States, a country with few advocates for voters and centuries of wondering who actually should have the opportunity to engage in electoral participation.  

And so introducing those notions, a lot of political scientists often come to these early calls by suggesting, what does it matter?  All of the folks that were called other than Gore actually ended up winning.  But I remember reading something from a more sociological point of view, and it appears here in red at the bottom, and these scholars, Kurt and Gladys Lang, after studying the 1964 early calls, suggested that no two elections are exactly alike, and in the United States, in a country with all of this ambivalence, the matter of regulating the dissemination of returns on election day should be debated less in the terms of the number of voters affected than the impact on the legitimacy of the process.  And that was a paragraph that has haunted me for quite some time, and so it led to my writing this book, and the book has three parts.   

The first part:  How has electoral participation been portrayed in print news coverage of presidential elections from 1948 to 2016?  And to answer that, I looked at how the words “vote,” “voter,” and “voting” were treated in news conducting a content analysis.  I learned two primary findings, and so the next step was:  Would other people see what I saw or was I just an academic studying these words?  So I led to – I conducted experiments and face-to-face focus groups.  And then finally, to interpret what I learned from the content analysis and the experiments, I did interviews with 54 elite journalists who actually wrote for the papers in content analysis.  And one of my favorite interviews was with a gentleman who covered the 1956 Republican nominating convention for the Los Angeles Times, and he gave me three hours.   

And so these were the parts that contributed to the analysis that I’d like to discuss today.  For those of you who might be wondering, these were the newspapers I examined, again between 1948 and 2016.  I looked at the words “vote,” “voter,” and “voting.”  There were over 3,000 – 36,400 of them in my project, so I did some macro quantitative work and then I did a stratified random sample of a smaller number of them, almost 4,000, where I coded for a variety of factors.   

The primary finding from the macro level analysis – and this probably isn’t going to surprise a group of news professionals – is that these three words – “vote,” “voter,” and “voting” – are largely unremarkable unless they’re endangered.  And you see the largest attention to these three words come in 2004, and that was the election right after the 2000 debacle where it was called for Gore, then Bush, then Gore, then Bush, and then it went to the Supreme Court.  And so a first finding is it’s very easy for journalists to overlook voters, and we can discuss how that might be important – I’m not trying to suggest that elites aren’t important but – other than when we’re in trouble.   

A second finding has to do with word choice, and my specialty, again, is at the intersection of persuasion in language, in political settings, or in news narratives.  But if we look at how U.S. journalists have used these words over the years, from the year 1948 to 1968 largely, we see the word “vote” being used more commonly than the other words, and “vote” picks up again in 2008.  And if we have time we’ll continue to talk about how those words were treated better, more favorably in the news – votes count, votes matter, votes can be sought, votes can be calculated.  Yet we also see in the later years of the analysis the word “voter” takes off, and I believe that many of the journalists who I spoke with had the very best of intentions, but a focus on the voter in these later years, the personalization was often quite thin.  It – voters could be duped, voters could be fooled, voters might not have a lot of information.  And so this shift may be beyond our mindset, but my experiments certainly pick up when everyday people read about votes that contribute, that are sought, that are counted – that encourages people to want to vote and be less frustrated with the media, versus a thinner personalization that encourages frustration with the media.  It simply isn’t something that encourages people to vote. 

And then a third finding from putting all these words together is there’s a journalistic temptation – and again, I understand this one – to address who might be the new voters.  And so just this morning as I was scanning the news, there were several articles about how Indian Americans might vote, or how Asian Americans might vote, or how various groups might vote – for many years of the analysis, how Latino voters might vote.  And intriguingly, in the focus groups that we conducted, even those are occasionally valuable.  They’re very well received by those individual groups.  There is a little bit of a backlash from the rest of voters in that the thin personalization that we saw on the earlier slide becomes primary as they get frustrated that other individual groups are the only ones who might contribute to the electoral outcomes.  

So here’s a little bit from the content analysis.  In the spirit of time, I’m going to hop over that to get to the experiments.  Again, from 1948 to 1968 and again in 2008, and part of that was the increase in the word “vote.”  Part of it also was journalists were very open, that when they were faced with a compelling candidate like Barack Obama, they just spent more attention to the crowds that Obama was amassing.  There was one frame, and we call that the mobilized participant frame.  In contrast from 1972 to 2000, voters were more likely to be subsumed over – under public opinion polls and discussed as if they could be predicted long in advance by strategists.  And we called that the isolated spectator frame. 

And when we held experiments and follow-up focus groups, we found when voters had the opportunity to read article (inaudible) made a difference, the mobilized participant frame, they were more positive in their responses; they were more likely to mention the importance of voting and they were more likely to talk about citizens in the United States as having efficacy.   

In contrast, when participants read the article where they were subsumed into public opinion polls, as has been the case 1970 to 2000, they discussed elections as negative gains, they voiced frustration with politics and the media, and they bemoaned the power of the press, parties, and money in campaigns.  And what’s notable here is that our Republicans, after reading the first frame, after reading the mobilized participant frame, were not – didn’t voice anger at the media, and that’s a difficult finding to see.  In contrast below, both Democrats and Republicans voiced frustration with politics and the media when they read the isolated spectator frame. 

So again, we did the content analysis to get a sense of how voters were really portrayed.  We held the experiments.  We talked to journalists, the individuals who wrote many of these stories.  And the first finding was sort of a really huh, wow, we don’t do that.  So I – my stance was I’m a university professor, you’re a news professional; please walk me through my examples.  And so I would give them examples of the two frames as well as other snippets from coverage.   

And a second finding from the interviews was that journalists really feared that strategists take away the news narrative and it is the journalist’s job to take us behind the curtain and expose what’s really happening.  And I had several people who were so incredibly honest in that they hadn’t realized how that might shift the mindset such that the gaze, the reportorial gaze is just on the folks who are working for the campaign as opposed to all the individuals who might have the opportunity to vote. 

Three other findings here that sort of speak to the differences between the portrayals:  They believe that a sense of patriotism as reporters were returning from World War II and/or a good candidate, Barack Obama, amassing crowds could lead to just more favorable coverage in general, which would extend to voters.  The journalists I talked to expressed a normative connection to the public in that they thought through this and thought, “Well, we really would like to engage in reporting that can help the democratic condition in the United States.”  And then they kind of closed with “If you can give us more empirically backed research projects, that might help us not only change the mindset in the newsroom, but change the mindset with our editors that we could write different types of stories.” 

And I’m going to hop over – I’ve been speaking for 20 minutes.  I’m going to spend just a moment here and a moment on my next project.  But in the conclusion of the book, I encourage not only reporters but all of us, anyone who cares about promoting the role of citizen participation in the news, to engage in three strategies.  And it’s watching our own language, and the three of them are these here.  Steven Pinker is a psychologist at Harvard, and he’s regularly regarded as one of the most hundred innovative thinkers in the United States.  And he shares that once we know something, it’s very difficult to talk about it without our knowing what we know.   

And so an example that I used is my husband is a passionate basketball fan, and if you sit next to him at a game, sometimes we don’t know exactly what he’s talking about, and/or some of us may love the financial channels of news, and yet if one of our children were to walk into the room, they wouldn’t know everything that we know, so they wouldn’t know what they – it’s very difficult to catch up.  And what Pinker shares is in reporting, in education, whenever we really want to socialize people in a positive direction, spelling things out helps.  And that is one of the points that I address in the conclusion. 

A second notion is the theory of rhetorical sensitivity that comes from Rod Hart in the late 1960s, early 1970s.  And his point here is not everything has to be super-circulated.  And it’s provocative, in that we think about how much attention the current President of the United States gets.  And the point here is some of it is incredibly warranted, yet some of the attention, if we go back to just the amount of attention he was able to get in 2015 and 2016, perhaps every once in a while reporters might have wanted to look away.   

And the third point in watching our own language is – this one’s from Kathleen Hall Jamieson; she’s at the University of Pennsylvania, and she did a wonderful study where she showed that journalists didn’t pay attention to the 1996 presidential campaign in the United States, and that was Bill Clinton as an incumbent and Senator Bob Dole as Republican challenger.  And journalists just thought the election was in the bag, and so there was far less reporting that year and the lack of reporting correlated – can’t say causation – correlated with low voter turnout that year. 

And so as we kind of pull it all together with some concrete communications strategies – spelling out how and why voters matter, being mindful that everything that’s provocative might not need to be shared, but then on Jamieson’s point, even if journalists are bored that there’s a lot of important things in democratic elections to share, particularly the voters’ role – all of this led to a project that was funded by the Democracy Fund – and I’m closing down right now – and they’re a group in Washington D.C. that does nonpartisan research on maintaining democratic values.  And the folks who called me were from the trust in elections condition, and their question was, how could you build on this book to tell us how to talk about threats to elections in ways that people still want to vote?   

And we engaged in two sets of experiments.  I’m happy to talk about this in the Q&A.  The first was we tested two types of articles, one that only addressed the threat and one that addressed the threat with a paragraph that addressed a solution to the threat, and the conditions were aging voting machines, cybersecurity, and voter roll purges.  And in all instances, the articles that mentioned the solution were more likely to inspire people to feel that their vote was meaningful and to feel that they could make a difference.  And that solution part, as in my book, leads to a little bit more trust in the news, and so we look forward to continuing to research this as we can amass a variety of findings that giving that solution as opposed to constantly making people feel threatened could be a way that newsrooms could develop greater trust with their readership. 

Final thing I’m going to say:  We didn’t just study how journalists talk about threats, but we also talked about how people who had large Twitter followings discussed threats to elections.  And this studied both solutions and the tweets were in the days heading into the 2018 campaign, and we believe that the threat tweet to the left was very well-intentioned.  This individual tweeted, “If your vote didn’t matter, they wouldn’t be trying so hard to keep you from voting.”  But when we tested that in a set of experiments against this other one – and this was – the person on the left was actually a rapper and the person on the right – musician – was a mobilization group.  Just that extra phrase of “requesting a provisional ballot is required by law” truly inspired the feeling that people did matter and that their votes did count. 

And so I just thank you for your time.  I look forward to hearing your questions.  I’ll turn it back to Jen. 

MODERATOR:  Thank you, Professor Jarvis.  This is a fascinating discussion of the way journalists can really have an impact when writing about elections.   

So we will now start the Q&A.  If you have a question, you can go to the participant box and virtually raise your hand and we will call on you.  You can also submit your question in the chat box in writing and your question will be read.  So I will turn it over to questions.  Do we have any questions? 

(No response.) 

MODERATOR:  Very surprising not to have questions from this very active group. 

Professor Jarvis, while we’re waiting to see if anybody would like you to elaborate, maybe you could talk a little bit more about what journalists – how journalists can approach writing about threats to the election in 2020 in a way that is still constructive for readers. 

MS JARVIS:  Absolutely.  And if anyone is interested on this, our white paper appears on – if you were to Google my name, Jarvis, and “electorally speaking,” we posted both a one-pager and a white paper that – I believe it’s 50 pages – and we share our stimuli.  So there you can see we took actual articles on these three topics – cybersecurity – and that came from The Washington Post, where we just cut one condition where the paragraphs from that article that only addressed threat, and then the other condition – there had been a paragraph about solutions that – what government was trying to do about it – that we included in the second condition.  And it’s – it can be as small as a 25-word paragraph that alerts the reader to what government is doing or to what other organizations are doing to really inspire a sense of faith. 

A second condition had to do with aging voter machines, and I believe that this one came from USA Today.  And it was very important for us to have a sense of external validity in the project, such that we weren’t just writing stories that academics would write, that we were using real stories that actually aired.  And so in the voting machines, aging voting machines article, one just featured the threat and it was a story about a situation in North Carolina where the voting machine was old and simply stopped counting votes after the nth vote.  But then the solutions article started with all of that content and moved in the direction of the state of Colorado actually has been highly regarded as to what they’re doing for election hygiene.  So there was a bit of a solution, and that had an effect.   

And then the third condition was on voter roll purges, and it was a story about a gentleman in Ohio who had served in the military and lived in his house for 15 years but hadn’t voted recently.  And he wanted to go vote against the legalization of marijuana – that was one condition – and he had been purged.   

And then the other condition, just to have a little bit of information that was solutions-oriented – and that one actually had the most profound effects, in that I think it was sort of surprising for people to read an article about a gentleman who had served in Ohio being purged in that most of the time when we see voter roll purges, we don’t see it in a rural area in the middle of Ohio.  It’s more likely to be the story’s written about urban areas. 

MODERATOR:  Thank you for that.  I do have a question in the chat field from Pearl Matibe from Open Parliament in Zimbabwe.  Her question is: “Have you seen any trends on voters after publishing articles on foreign interference or cyber breaches?” 

MS JARVIS:  This is fascinating.  We – thank you very much, Pearl, for that question.  I haven’t seen anything published.  I have a couple instincts from our data.  The work with the Democracy Fund – so often when we’re involved and our journalists are involved, these items are very top of mind, yet we have in this project, in the experiments heading into 2018, we would ask open-ended measures of “When people say that elections are rigged, what do you think they mean?” or when – “How do you feel about elections in the United States?”  And when we give voters the opportunity to respond, these concerns aren’t as top of mind as partisanship.   

Now, it could change going into 2020.  Of course, things could change.  But yeah, on one hand, it might be slightly comforting that these aren’t the top-of-mind answers for most individuals.  Yet the other hand, of course, is that they’re very, very important, and definitely deserve attention in the news. 

MODERATOR:  Thank you.  If we have any other questions, please raise your hand virtually in the participant field, or submit in the chat box.  We’ll give it another minute.  If we don’t have a question, we’ll go ahead and wrap up.  Oh, looks like we have another one. 

This is from Novi Magazine, Serbia, Marija Sajkas – apologize if I mispronounced your name:  “Do you have any concrete examples about the best ways for journalists to report about the 2020 election?  Is there something in particular that we should pay attention to?” 

MS JARVIS:  I really appreciate – thank you for that question.  And I’m smiling for – the final chapter of the book that I addressed begins with Barack Obama, and I say something to the effect of former President Obama had a bit of an obsession with voters in that he knew his elections looked a little bit different than traditional Democratic constituencies.  And he knew that Clinton was in trouble – Hillary Clinton was in trouble in 2016, and that in the reporting that I read, she simply wasn’t getting the attention of the rallies that Donald Trump was getting.   

And so toward that question, for someone who has empirical support that telling voters they matter matters, I think it’s okay for reporters to mention both Trump and Biden are truly asking for the vote, and that hasn’t always been the case.  And when reporter – when – our data suggests that when people see that, it opens up a different mindset, that they’re contributing to something versus when Trump’s key strategist or Biden’s key strategist tells us on cable news that it’s in the bag for someone. 

A second point – and I do not want to be partisan in any respect – I’ve heard it a touch more from Biden’s supporters, but it opens up the frame in that I believe a lot of reporters, the people I talked to shared – these are their words, not mine – that it was really easy to believe the polls in 2016 and not think about how the polls would connect to turnout.  And the Biden campaign, at least some of their surrogates, are inviting us to think that in the weeks or days leading up to the election that the polls may be accurate, but the outcome is still an uncertain question.   

And so to the extent that that could appear in the news rather than forecasting it and predetermining it, reminding voters that everyone who turns out does indeed influence how the results will go.  Wonderful question.   

MODERATOR:  Thank you.  We still have time for one or two more questions if we have any.  You can submit in the chat or raise your hand. 

(No response.) 

MODERATOR:  In the meantime, Dr. Jarvis, I might just ask you if you could elaborate on the partisan differences – and I know you’ve done some research on even differences within the Republican Party between Trump supporters versus Republicans who do not support Trump.  Maybe you could speak to that a little bit, in terms of their media participation or their reaction to this type of coverage. 

MS JARVIS:  Sure, absolutely.  Thank you for that question and I believe it goes back to Pearl’s somewhat, if I were to spend more time with Pearl’s question, which now I have the opportunity to do so. 

The research in my field of political communication and also in political psychology is really moving in a direction that we interpret so much through partisan screens, so much through partisan screens, in that folks on the right trust media outlets from the right, social media from the right, and their friends on the right.  Similarly, folks on the left, same.  Media outlets on the left, their friends on the left, social media on the left. 

And so when we’re faced with new information, the data just continued to show that we begin to interpret it in means of how we think our favorite politicians will interpret it.  I recognize that’s a step aside where we were going with this.  Jen mentioned an article that I published a few years ago, and the data came from 2016 – May of 2016 – when we had a poll out in the field of over 2,000 individuals in the United States, online poll.  And we tried to track their attitudes on the Democratic candidates and the Republican candidates.  And if we go back to May of 2016, it’s a really interesting time in that the elections, the nominees, were not a foregone conclusion.  Again, I work in Texas, and many people in Texas thought that somehow, Senator Ted Cruz would beat Donald Trump.  Other people were thinking that perhaps the Republican National Committee would come out and do something such that someone who had never held elected office was going to be catapulted to the White House.  Then we had data again from 2017, which is six months after a very surprising election.   

So in the project where we compared early Trump supporters to traditional Republican voters, what we learned is that early Trump supporters had a far more contracted conservative news diet.  Traditional Republicans would seek out five or six news outlets regularly, and it was our Republican – or early Trump supporters who sought out fewer news outlets.  And then that led us to read into our open-ended data to get a sense of how else they were different.   

But that was a key finding.  So not only are we in a moment where we interpret media through partisan screens, but the early Trump movement were seeking out a far smaller set of outlets.  Traditional Republican voters were more likely to be reading Wall Street Journal.  The early Trump supporters were more likely to be parked on Fox News.   

MODERATOR:  Thank you for that.  It looks like there’s a little bit of a follow-up to Pearl’s question, and she’s asked:  “Could you speak on the quality, quantity, or approaches journalists have taken in covering the U.S. ‘mouthpiece,’ in quotation, of each presidency, either as the White House spokesperson or as in the State Department?  Has it leaned partisan?” 

MS JARVIS:  Pearl, yet again a wonderful question.  I have colleagues who study the presidency specifically, and so I really don’t want to speak out of turn on this one.  I can kind of follow up on based on wherever – so you’re speaking kind of to the supply side of how journalists are covering?  The only thing I can kind of speak to with integrity and out of respect for your question is the demand side in that however – whatever folks in the White House are saying are partisan —  

QUESTION:  Yes.  No, maybe I can just – I guess I can quickly clarify.  I think you’re quite right.  I was just trying to find out whether the way or the manner in which journalists have covered each term of a presidency when that has switched from either Republican or Democrat, how that may have impacted voter behavior.  So has, for instance, the White House spokesperson or the State Department spokesperson as kind of the official information channels, if you will, or information platforms impacted how journalists have covered, and has that impacted voter behavior, as opposed to other countries where it might be viewed as propagandist, so to speak?   

So basically, what journalists might publish, what might – journalists might write as a result can either sway voter behavior.  We see this in other countries.  I just was wondering whether you’ve seen how journalists report on these two institutions, the White House and State Department, in an election year.   

MS JARVIS:  I appreciate that, and I don’t – I appreciate that, and I – wonderful to hear was your voice and thank you for (inaudible).  I don’t – again, I don’t study the presidency in a way that I can do a good job here.  I can speak to the side a tiny bit in that – I recognize this isn’t going to the heart of the question or treating it with the respect it deserves, but if we go back to 2016, if we want to link back to voter behavior, it was only 77,000 votes in three states – Wisconsin, Pennsylvania, and Michigan – that separated a Clinton victory from a Trump victory.  And this isn’t speaking to the mouthpiece by any regard, but I heard voter behavior twice or more, so I had to go into this instant.  Voter – I’ve always lived in safe states.  I was born and raised outside of San Francisco, California, which is a blue state.  I currently work in Texas, which is a red state.  And when we think about voter behavior, the news and quite frankly districts signal to people more or less power.  I’ve always wanted to do studies in Ohio or Florida in that they’re routinely up for grabs, for Michigan, Wisconsin, and Pennsylvania to be up for grabs and really seminal here. 

So I can’t speak to that, but when it comes to voter behavior, and is anybody treated as more important than others, we do have data to suggest that – many reporters talk about politics as a game, yet the game frame in Ohio and Florida is a frame that hands far more power over to the voters than would a game framed in a state like California or a state like Texas.  And again, I thank you for your question.  I don’t want to speak out of turn and portend to have any expertise that I don’t.   

MODERATOR:  Thank you.  There was also a follow-up from Marija Sajkas at Novi Magazine in Serbia, which may be a little bit outside of your research area, Professor Jarvis.  But the question is:  “Do you have any insights into ways activities on social media and influencers, closed groups, influence voting?” 

MS JARVIS:  Absolutely.  And the key finding, and it’s about a decade old, is that we’re most likely to trust what we see from our friends.  So once something starts getting circulated by friends, and by that friends or ideologically aligned or families, that has just the greatest impact.  Subsequent findings have split hairs on that finding, so we have more specifics, but the key group is getting it from a friend is more likely to influence our attitudes and our behaviors. 

MODERATOR:  Okay.  With that, I don’t see any other questions coming in to the chat.  So we will be posting a transcript of this briefing later today on our website.  We will also make the PowerPoint slides available to the participants, who may want to refer back to the research.  And I want to thank you, Professor Jarvis, for this really excellent briefing today that’s very useful for our journalists.  And those who do share a story, please send us an email at if you plan on writing something.  Thank you, everybody, and thank you, Professor Jarvis. 

MS JARVIS:  Thank you.  Thank you very much.  Thank you.   

U.S. Department of State

The Lessons of 1989: Freedom and Our Future