Monthly Archives: August 2019

The 3 Things Democrats misunderstand about Republicans

By Kent R. Kroeger (, August 26, 2019)

In the late summer of 1994, I walked into a small, two-story Fairfax, Virginia office in the middle of a nondescript commercial complex of converted town homes, itself residing on the edge of a middle-class residential neighborhood.

The sign in the front window — Republican Party of Virginia — was faded from sunlight exposure and severely bent on the corners. After a few unanswered doorbell rings, I just walked in through its half-opened front door. Not a soul in sight.

“Hello!” I sheepishly bellowed. No answer. I walked halfway up some musty-smelling stairs, softly issued a last ‘Hello!’ and walked back down the stairs, picking up a couple of brochures at the receptionist’s desk as I walked out.

“Can I help you?” I heard as I opened my car door to leave. A large man approached me with a striking resemblance to Sebastian Cabot (a reference only someone over 50 can visualize).

“Yes, sir,” I replied. “I’d like to volunteer for the upcoming election if you need help.”
So began my 20-year-plus defection from the Democrats to the Republicans.

Only six years earlier, I volunteered for Jesse Jackson’s second presidential campaign and, four years before that, worked as a paid staffer for then-Iowa congressman Tom Harkin’s first U.S. Senate campaign. In that same 1984 election, as a Black Hawk County Democratic delegate for Jesse Jackson, I saw first-hand how the proportion of county delegates a candidate has does not necessarily translate into the same proportion of state convention delegates.

And there I was, in 1994, on a steamy Virginia weekday morning, offering my time and effort to help people like Oliver North — who was running for the U.S. Senate against Democratic incumbent Chuck Robb — get elected.

For the next 20 years, I attended Republican precinct meetings, helped organize neighborhood fundraisers, and worked the phones for the party’s get-out-the-vote efforts. When I left the Republican Party in 2016 to work on my brother’s campaign to challenge incumbent Republican Rod Blum in Iowa’s 1st Congressional District, I felt some ambivalence.

Though never an insider for either party, as a part-time activist for over 35 years, I have developed a strong sense about each party’s organizational and cultural differences. In most cases, these dissimilarities have remained constant over variations in time and geography (Iowa, Virginia, New Jersey). These observations are not, however, necessarily a reflection of the parties’ rank-and-file supporters. And this is not a thesis on the policy differences between the two parties. The following is merely the palpable imprints the two parties left on me.

And as I now reflect on these impressions, I have come to believe some of these differences may explain why, since Watergate’s aftermath, the Republicans have won more national and state elections across this country than have the Democrats, despite the fact that public opinion across many issues has become more ‘liberal’ since 1978.

That the Republicans continue to win elections on a consistent basis cannot be explained by blessing or luck. It is more likely because the Republicans have created a leadership and activist culture organized for just one purpose: winning elections. Where the Democrats have created the world’s most sophisticated fundraising and vote harvesting apparatus in American history, the Republicans have created the most successful political party.
As to how this could happen, I turn to the three specific organizational and cultural differences I’ve observed across the two parties. This first relies a bit too much on psychoanalysis, perhaps, but I believe is the most consistent difference between the two parties since 1978.

(1) Republicans have an enduring sense of being the ‘minority’

Calling Mitt Romney’s 2012 defeat to Barack Obama one of the most painful political losses in his lifetime, conservative radio host Rush Limbaugh found clarity in a moment of despair as he told his radio audience the day after Obama’s re-election:“I went to bed last night thinking we are outnumbered. I went to bed last night thinking we’ve lost the country. I don’t know how else you look at this.”

Fox News’ inelegant host Bill O’Reilly was more direct: “The white establishment is now the minority.”

Republican feelings of being outnumbered did not start in 2012, however. Or in 2008 for that matter. In fact, it can be traced as far back as Richard Nixon’s first presidential term as he reacted to the growing anti-Vietnam protest movement. “And so tonight — to you, the great silent majority of my fellow Americans — I ask for your support,” Nixon intoned in a televised address to the nation.

But if Nixon was anything, he was a realist — laced a pinch of dark paranoia. Having won his first presidential term by a whisker’s margin, he knew any majority he might rely on politically could never be taken for granted.

Nixon’s frequent lament about his mistreatment by other political elites would contribute to his eventual downfall in the Watergate conspiracy. More broadly, many of Nixon’s personal weaknesses were imprinted onto the psyches of Republican Party’s faithful. The psychological trauma dealt to the Republican Party from the double whammy of Watergate and the inconclusive end to the Vietnam War cannot be over-estimated. Despite a relatively close presidential election in 1976, the Republican Party’s status in Congress and across the state legislatures was in descent.

Around this time, we see the rise of terms such as ‘middle America,’ ‘liberal media,’ and ‘moral majority’ — terms still heard frequently in Republican circles (see Figure 1).

Figure 1:  Use of terms “middle America” (green line), “liberal media” (blue line) and “moral majority” (red line) in literature since 1950.

Source: Google Ngrams

The Republicans I knew in Virginia had an “us against the D.C. political establishment” mentality. Everything, I mean everything, was working against them. To this day, my conversations with Republicans often include references to the ‘liberal media’ and ‘coastal elites.’ The news media is against them (even Fox News, according to some particularly strong Trump loyalists). Washington ‘elites’ are against them. Hollywood is against them (try to name more than one scripted TV show today that openly espouses conservative principles). Social media is against them. The public school system is against them. Academics are against them.

These laments emerged in force in the mid-1970s. The term ‘Moral Majority’ was minted by conservative religious activist Paul Weyrich from a perspective of political vulnerability, not strength. It was a defensive mobilization strategy built on a deep insecurity, heightened to toxic levels by Watergate. “We are the moral majority!” was a Republican acknowledgement of their political minority status and the electoral necessity of a coalition with like-minded Democrats and independents. In the 1980 election, many of those like-minded voters would be called ‘Reagan Democrats.’

The Republican affection for the economic aristocracy also brings with it an attendant understanding that this socioeconomic bias puts them in a definitive minority status within the electorate. Without election activities engineered to attract weak partisans and suppress enthusiasm among Democratic voters generally, the Republicans are at a serious disadvantage to the Democrats.

And, in practice, this is has been the electoral strategy of the Republican Party since Watergate. Whereas I frequently heard within the Iowa Democratic Party leadership in the 2014 and 2016 elections that all the Democrats needed to do was “get out their voters,” I don’t recall ever hearing Virginia Republican leaders uttering an equivalent sentiment. Virginia Republicans in the 1990s had endured a century of Democratic Party domination of Virginia politics. For them to think they could just turnout ‘their voters’ (as if voters are owned) would have sounded suicidal.

More importantly, the Republican’s basic approach towards the electorate, motivated by this sense of minority status, makes them a far more effective competitor on election day. Where the Democrats believe the majority of Americans already love them, the Republicans entertain no such assumption.

Any attempt to explain why Republican candidates have kept winning elections since Watergate must therefore include, if not be dominated by, an understanding of their deep sense of their own minority status.

(2) Republicans are more tolerant than Democrats of opinion diversity within its ranks

In my experience with Virginia Republicans and Iowa Democrats, I witnessed two different management styles by the party leaders that impacted how opinion differences were handled within the organizations.

In 1994, when I joined the Republican Party of Virginia (RPV), there was open warfare between upstate (D.C.-area) conservatives, who tended to be more liberal on social issues, and the downstate conservatives, who were much more conservative. The religious right had just won the RPV chair position (Patrick McSweeney) but the battle was far from being won. What left an impression with me was how the Virginia Republicans more frequently engaged in bottom-up conversations. The Republican activists and volunteers I met, many of whom were already successful professionals outside of the party, were not overly deferential towards the party’s staff and leaders. In fact, they were often dismissive of the ‘political hacks’ running the state party. The organization less hierarchical and, frankly, more democratic.

The Iowa Democratic Party (IDP) was the opposite. The party leaders were either career activists that had risen up the ranks or were from academia; whereas, the activists and volunteers were typically students or young professionals early in their careers. This made the IDP more hierarchical and affected internal dialogues which were almost always top-down and less democratic.

Strict hierarchies tend to suppress diversity of opinion within an organization; flat organizations much less so.

Still, I encounter resistance anytime I suggest to a Democrat that the Republican Party is more tolerant of opinion diversity. ‘No f**king way,’ is a common response.
But, in fact, way.

Such resistance is rooted in a common belief that intellectual nimbleness is associated with higher levels of education. In reality, this is not necessarily true — at least not when it comes to politics. According to the analytics firm, PredictWise, the most politically intolerant Americans “tend to be whiter, more highly educated, older, more urban, and more partisan” and lean liberal, particularly on social issues.

This finding augments research by University of Pennsylvania professor Diana Mutz, who concluded in her book Hearing the Other Side that white, highly educated people — who skew Democrat and liberal — tend not to talk with people who disagree with them — which makes the probability higher that they will stereotype those with opposing views.
Self-selection bias impacts how we organize our lives, including who we talk to on a regular basis. The more educated and wealthier we become, the more we can control that dynamic and homogenize our surroundings.

But I posit another factor behind higher levels of opinion tolerance among Republicans. Most of the Virginia Republican leaders, activists and volunteers I met were already (or had been) successful business people from the private sector. They often had MBAs or business law backgrounds and were almost always private sector oriented.

Why does that matter?

Most businesses, large or small, try to avoid politics when making contact with current or potential customers. The big companies may hire powerful Beltway lobbyists to do the dirty partisan work, but at the point-of-sale, partisan politics does more harm than good. Republican party leaders and activists understand that better than their Democratic counterparts because they are more likely to have dealt with that issue in their professional lives.

In contrast, many Democratic Party leaders and activists have only worked in academia or in the non-profit, public advocacy sector. Planned Parenthood isn’t in the business of compromising its policy message in order to appeal to a broader audience. Greenpeace activists aren’t looking for common ground with Exxon-Mobil so they can work together more effectively. The Democrat DNA is fundamentally different and inherently resistant to compromise and opinion diversity.

Again, why does this matter? Because the electoral arena has more in common with consumer brand market competition than it does to an academic symposium on immigration policy or a non-profit funded public awareness campaign. Brands are forced to adjust to market realities in ways few advocacy groups ever encounter. “A wise man adapts himself to circumstances, as water shapes itself to the vessel that contains it,” says an old Chinese proverb. A more contemporary quote comes from business author Alan Deutschman who popularized the business catchphrase, “Change or die.”

The Republican Party is powered by a different culture and mindset than the Democratic Party. It makes the GOP more flexible on issues when it matters most — at election time. Republicans are betting at adapting to changing market conditions.

This cultural difference manifests itself when looking at the prior occupations of U.S. House members in the 103rd Congress and the current 116th Congress (see Figure 2). In the 103rd Congress, 41 percent of GOP members had a business background and that would rise to 47 percent in the 116th Congress. Only 18 percent of Democrat House members had a business background in the 103rd Congress, though it has risen to 30 percent in the current Congress. The other stark difference is in those House members with a public service or political activism background. As of today, 47 percent of Democrat House members have a public service/political background, compared to only 30 percent of GOP members.

Figure 2:  Occupations of U.S. House members in the 103rd Congress and the 116th Congress (House members can have multiple prior occupations)

The rise of Donald Trump is a living testament to the ability of the Republican Party and its rank-and-file supporters to embrace change and diversity. How can an evangelical Christian who abhors adultery embrace Trump? Apparently, its pretty easy when you believe you need to compromise on a candidate’s qualities in order to win elections. Does that sound like a quality commonly found among Democrats? I would say not, even if Joe Biden is still leading in the 2020 Democratic nomination race.

(3) Republicans hold themselves accountable for failure

Business people love quantitative performance metrics. They measure like their career depends on it, because if often does. And they always have that one, bottom line metric looming over them: profits.

The Republican National Committee (RNC) has a similar orientation. The party doesn’t value moral victories. The RNC cares about one metric: You either win or lose. It’s the Ricky Bobby theory of electoral politics: If you ain’t first, you’re last.

That’s why Republicans have no problem turning on themselves when their party loses an election. I wouldn’t call it introspection on their part. It’s more like, “We just lost, I need to fire somebody in this office.”

When the Republicans lost to Barack Obama in 2008, Senator John Cornyn, chairman of the National Republican Senatorial Committee, turned the spotlight on his entire party. “While some will want to blame one wing of the party over the other, the reality is candidates from all corners of our GOP lost tonight. Clearly we have work to do in the weeks and months ahead.”

There was little vote-shaming (i.e., blaming voters for how they voted) by Republican leaders after 2008 (though there was some of that after the 2012 election when some in the GOP pundit class blamed voters for wanting ‘more stuff.’).

After the 2012 election, the RNC commissioned an internal study to explain the election results and how the party needed to adjust going forward. The Atlantic’s Garance Franke-Ruta described the 2012 postmortem report as “an astonishingly frank document that calls for major changes in how the party addresses minorities, women, and its own campaign processes.”

No such study has ever been released by the Democratic National Committee (DNC) following their 2016 election debacle. Instead, groups outside the DNC have picked up the slack. One group,, led by Karen Bernal, a three-term Chair of the California Democratic Party’s Progressive Caucus, Pia Gallegos, Chair of the Adelante Progressive Caucus of the Democratic Party of New Mexico, Sam McCann, a writer, and Norman Solomon, co-founder of, produced their own report.

Unfortunately, such “independent” reports are more ideological treatises than true objective analyses — which is what the Democratic Party desperately needed after 2016.
Instead of critiquing their party’s policy and voter outreach strategies, as the Republicans did after 2012, the DNC and most Democratic leaders have blamed Russia for their 2016 loss to Trump.

Like a starving dog pack, the Republicans act collectively after electoral defeats; while, in contrast, the Democrats scurry about doing their own individual thing hoping that somehow it will all come together when the next election rolls around.

Republicans more collectivist? Democrats more individualistic? Isn’t that backwards?

Not at all.

There are many organizational and cultural characteristics distinguishing the Republicans from the Democrats. The three differences I’ve highlighted here may seem counter intuitive, which is why I believe the Democrats must acknowledge them if they are ever going to sustain any electoral success over multiple, contiguous elections.

As of today, it is the Democratic Party that shuns introspection and instead leans on undemocratic processes to push its agenda and pre-approved candidates through to its rank-and-file supporters. Underwriting this culture is the Party’s widely held belief that Democrats are the manifest majority in the U.S. and will become even more so as the country’s demographics change going forward. ‘All we have to do is turn out our voters,’ they tell themselves.

American’s electoral history is a graveyard for such hubris.

If the Democratic Party’s organizational culture doesn’t change, I predict more unexpected defeats in its future.

Please send comments to:

Democracy is messy, especially when you don’t respect the concept (The Tale of Brexit)

By Ken R. Kroeger (, August 20, 2019)

News networks are better at crushing (or lifting) outsider campaigns than they are at helping major candidates

By Kent R. Kroeger (, August 16, 2019)

I don’t blame Bernie Sanders for his crabbiness towards The Washington Post. He cares about his national media coverage as he knows it may be the difference between winning and losing the Democratic Party’s 2020 presidential nomination.

Sanders ignited a firestorm this month when he implied The Washington Post’s less-than-favorable coverage of his 2020 presidential campaign was influenced by the paper’s owner, Jeff Bezos, the founder, president and CEO of, Inc.

Immediate shock, outrage and dismissive eye-rolling animated the press corps.

“(Sanders) said what he said to get a cheap applause line at a town hall packed with supporters. The problem for Sanders, Trump and politics more generally is that many of the people who hear things like this from them don’t know better,” wrote CNN editor-at-large Chris Cillizza. “They actually believe there is some sort of conspiracy between corporate America and the news media. And when politicians — whether they are Sanders, Trump or anyone else in either party — stoke that sentiment, that’s dangerous. And bad for democracy. Full stop.”

“We’ve been tracking press coverage all primary long and Sanders has consistently been at or near the top of the field in terms of the volume of news coverage he’s received,” posted Nate Silver, the editor-in-chief of FiveThirtyEight, on Twitter.

These are the predictable responses one can expect from the national media when someone accuses them of systematic editorial bias.

Sanders, however, was wrong to suggest Jeff Bezos, or anyone connected to corporate, would intervene to impact WaPo editorial decisions. It just doesn’t happen that way.

The common interests of journalists and political elites drive the news media’s coverage of political candidates. They are alone a sufficient condition to power any conspiratorial-looking editorial process dedicated to helping one set of political candidates over others.

It doesn’t require late-night calls using anonymized cell phones or encrypted emails across secured networks. Journalists and politicians simply need a shared motivator to engineer an organic, successful, and legal conspiracy.

Perhaps ‘conspiracy’ is the wrong word for it. It is more like a confederacy built around an informal covenant. Members may not have secret handshakes, but they learn of their shared interests by going to the same schools, living in the same neighborhoods, attending the same parties.

The national news media and the Washington, D.C. political elite belong to the same club — a clique where you need to be invited, of course. Sorry, President Trump, your membership application has been misplaced. And, apparently, the membership renewal forms have been rejected for Hawaii House member Tulsi Gabbard and New York Senator Kirsten Gillibrand. They were once members in good standing. But not now.

The Systematic Bias of the National News Media is Undeniable

Recently, Michael Tauberg, an engineer by day and data journalist at night, published data on the tone of online news coverage for each of the 2020 Democratic presidential candidates from January to April 2019.

Figure 1 (below) shows the average news story sentiment for the major Democratic candidates. South Bend Mayor Pete Buttigieg, New Jersey Senator Cory Booker, former HUD Secretary Julian Castro, former Texas US House member Beto O’Rourke and Minnesota Senator Amy Klobuchar received the most positive news coverage over that period. More interesting are the candidates at the bottom: Gillibrand and Gabbard.

Figure 1: Average Sentiment of Coverage from Liberal News Sources

Source: Michael Tauberg

What exactly did Gillibrand and Gabbard do to earn so much negative coverage? Gabbard’s unforgivable offense to the political/media establishment is well documented. Rolling Stone magazine’s Matt Taibbi lays out a few reasons for why the establishment shuns Gabbard. She hates regime change wars and holds the Democratic Party — and the Obama administration, in particular — partially responsible for these never-ending conflicts. Gabbard hits the Democratic Party where it feels most vulnerable to the Republicans: national security issues. The fact Gabbard single-handedly stunted Hillary Clinton’s 2016 campaign momentum in late-February 2016 by resigning from a Democratic National Committee leadership position and endorsed Bernie Sanders did not help her status within the party establishment either.

But what about Gillibrand?

Her heresy is pure political insider stuff. She embarrassed former President Bill Clinton (and his wife) when she suggested Bill should have resigned for having an affair with a young intern.

Responding to Gillibrand’s admonishment, the former president told CBS News: “You have to — really ignore what the context was. But, you know, she’s living in a different context. And she did it for different reasons.”

That was Bill saying Gillibrand said what she said for purely political reasons. In other words, she’s no longer a FOB (Friend of Bill’s). If the Clintons are consistent on anything, it is punishing those who are disloyal; and Gillibrand, who they hand-picked for the New York Senate seat Hillary vacated when she became Secretary of State, has never been welcomed back into the Clinton or party establishment fold.

This becomes even more apparent in Tauberg’s analysis of news headline sentiment (see Figure 2). No other 2020 Democratic candidate possesses an average headline sentiment score as strongly negative as the scores for Gabbard and Gillibrand. Their outsider status is quantifiable and it won’t be easy for either to rise within this crowded field of candidates if their media coverage does not turn more positive.

Figure 2: Average Sentiment of Headlines from Liberal News Sources

Source: Michael Tauberg

But do Gabbard and Gillibrand deserve this level of negativity from the national news media? More importantly, who makes that decision? What objective editorial standard is in play that says these two candidates are going to hammered (or ignored) by our news outlet, while this candidate is going to get a far more positive treatment.

All candidates have strengths and weaknesses. Short of a candidacy by someone like David Duke or Richard Spencer, it is seems reasonable that an objective news organization would balance the tone of its coverage for all candidates, even if more popular candidates may get proportionately more coverage than less popular candidates.

But the Tauberg data shows how unbalanced the coverage has been through April with the 2020 Democratic race. The candidate differences in Figures 1 and 2 are not the product of chance — they are the product of an overtly sectarian and discriminatory editorial process.

CNN’s Chris Cillizza and members of the national news media ask us to believe those decisions are based purely on objective considerations by the journalists and editors themselves. The mere suggestion by Sanders that a corporate owner of a newspaper or broadcast news network could impact editorial content is universally mocked by those in the industry.

Responding to another opinion journalist’s criticism of corporate news decisions, MSNBC contributor Jason Johnson asks, “Do you actually think that network and site owners waste their time micromanaging a writer’s opinions? Or is that just some stuff you throw out because it hypes up your fans? I’ve never had anybody, on any outlet I’ve ever worked for even bother.”

There is an answer to Johnson’s first question and it is called ‘self-editing.’ Every writer of import does it — consciously and subconsciously. Writers know what their editors, readers and owners like to read, or at least the successful ones know.

Has Johnson not noticed writers and political analysts sometimes are around one day and gone the next on MSNBC? Has he asked himself why he doesn’t see Krystal Ball around the office as much anymore? [Krystal is now an anchor on The Hill’s TV news podcast Rising with the Hill’s Krystal Ball and Saagar Enjeti — an excellent morning program, by the way.]

The owners of major news outlets don’t micromanage because they hire people to do that. And the most important editorial control is not done at the micro-level anyway, it is done at a macro-level through the hiring and firing decisions by middle and upper management. In just the past twenty years, MSNBC has gone through at least two substantial ideological shifts. It was a mostly non-partisan, straight news organization in the late 1990s, but shifted to the progressive (anti-Fox News) left in the early-2000s with show hosts such as Phil Donahue, Dylan Ratigan, and Ed Schultz, only to turn into an establishment left news organization in the early 2010s with hosts such as Rachel Maddow and Chris Matthews.

Driven more, perhaps, by ratings than any ideological predilection on the part of its ownership, MSNBC is not atypical of other national news outlets. It is a corporate news organization guided by corporate necessities (i.e., advertising revenue and profit). That causal model systematically and inescapably alters the news content, largely in favor of advertiser and corporate interests. Lockheed Martin budgets a lot of advertising dollars each year and they are not going to use their money for underwriting spots on Democracy Now or Redacted Tonight. It just doesn’t work that way.

Changes in cable news coverage can change candidate support — but is it predictable?

Every political campaign I’ve worked on had a candidate and staff that complained incessantly about their campaign’s news coverage. One nasty news story could get a reporter barred from future interviews. But what candidates and campaign managers feared most — more than being on the receiving end of a negative story — was getting ignored by the media. Nothing kills a campaign faster than not being covered. Advertising and door-knocking can help build name recognition and promote a candidate’s core messages, but the credibility and visibility conveyed to a campaign through a national news media filter is irreplaceable — even in the age of social media.

How do we know news media coverage matters? We will need data.

The Data

The following is a very preliminary look at this question regarding news media influence on political candidate support. Using APIs to query a GDELT Project database on daily cable TV news coverage from January 4 to August 6, 2019, and downloading Democratic nomination polling data from RealClearPolitics’ data repository, I conducted a time-series analysis (vector autoregression) to determine whether or not changes in the volume of a candidate’s cable news coverage causes changes in a candidate’s popular support.

Spoiler alert: Changes in the volume of cable news coverage causes changes in candidate popularity, but the effect size is small.

First, let us look a how candidate support has varied since January (Figure 3). Of the five major candidates, they all received a significant popularity bump from the point of their candidacy announcement. In general, the surge in popularity lasted about a month and ranged in magnitude from 15 points for Sanders to 9 points for Buttigieg.

Warren is an outlier in that her popularity rise has been gradual and did not include a steep increase after her February 9th announcement; and, unlike the other four candidates, Warren has not suffered a significant decline from an announcement-related high — in fact, she has gained around 10 points. In contrast, Biden has seen his support decline from around 42 percent down to 30 percent. Sanders, similarly, has lost around 12 points from his March high, Harris 5 points and Buttigieg only about 2 points.

Warren may be the tortoise in a race full of hares.

Figure 3: Democratic Candidate Support since January 4, 2019

Data Source: RealClearPolitics; Polling data displayed and analyzed as 7-day moving averages

One other notable feature is the 8-point surge Harris experienced after the first debate where she challenged Biden on his position regarding forced busing. However, her lift from that debate has all but evaporated since the second debate when Harris was confronted by Gabbard over Harris’ law enforcement record while the California Attorney General.

Turning to cable news coverage, there is more day-to-day variation in the volume of coverage for candidates, particularly around specific events such as Biden’s late-April candidacy announcement and around the two debates (see Figure 4). Another interesting feature is how Warren and Sanders mentions, starting in late-May, is now moving in near perfect lock step fashion. The reason for this strong correlation are the two dominant narratives being employed by both MSNBC and CNN during this period: The first focused on whether Warren’s rise is stunting any further rise in Sanders’ support. The second focused on Warren and Sanders defending their progressive policy ideas, particularly on their ability to fund them.

Figure 4: Democratic Candidate Cable News Mentions (as % of all mentions) since January 4, 2019

Data Source: The GDELT Project; Polling data displayed and analyzed as 7-day moving averages

As yet, I’ve offered no evidence that candidate support and cable news coverage are causally linked. And that causal linkage, if it exists, may be non-recursive and going in both directions — a candidate’s rise in popularity may inspire more news coverage, and a rise in news coverage may increase a candidate’s popularity.

Before formally crunching the numbers, a visual inspection of how cable news coverage moves relative to candidate popularity might be helpful. I’ll focus on Mayor Buttigieg for two reasons: there are clear periods where his popularity and news coverage move together, while other periods where they seem to move in opposite directions (I’ve included charts for all five candidates in the Appendix below, Tables A.1 to A.5).

Figure 5 shows Buttigieg’s dramatic surge in support occurred at the same time as his cable news mentions increased. To only the most hardcore politicalphiles was the name Buttigieg familiar before the 2020 campaign. By the time of his April 14th announcement, he was near his campaign highs in both popularity and cable news coverage. But it is not obvious from Figure 5 if popularity causes news coverage or the other way around (or in both directions).

Figure 5: Buttigieg Support and Cable TV News Mentions

Data Sources: RealClearPolitics and the GDELT Project; data smoothed using a 7-day moving average

Since his mid-April highs, he has experienced significant declines in both variables, except for a period just after the first debate when his cable news mentions surged a second time. Another important period is between June 20th to 30th when Buttigieg’s cable mentions rose steeply even as his popularity was in steep decline. The negative turn was most likely a product of his city, South Bend, dealing with a June 14th police shooting where a Black man had been killed by a white police officer. Buttigieg’s appearance before a public hearing on the incident did not go well, according to many observers.

What this episode demonstrates is that any model of support and news coverage may require an interaction term or independent variable measuring the tone of coverage. Unfortunately, I am still working on generating a sentiment analysis of cable news stories for inclusion in a future analysis.

With that methodological caveat, any causal relationship I might find between popularity and news coverage here is likely going to be an underestimate of the true relationship. Research is, after all, an iterative process.

Feel free to skip to ‘The Results’ section if you don’t want the details on the statistical methodology.

The Statistical Method

Noble Prize winner in economics, Clive Granger, defines a relationship between two variables as causal (X1 Granger-causes X2) if prior changes in X1 predict future changes in X2, independent of past values in X2 and while controlling for other potential causal factors.

A common statistical method for testing for Granger-causality is vector autoregression (VAR). The beauty (and limitation) of this technique is that it makes few assumptions about the causal relationships between variables. Hence, these models often devolve into ‘everything but the kitchen sink’ specifications that eat up degrees of freedom (which are typically precious, particularly for the analysis here which has only 215 cases to work with).

Due to these issues, it is often recommended that VAR models be employed during the initial model and theory-building stages and that more explicit, theoretically-informed statistical models be used in the final analytic stages (e.g., dynamic Bayesian networks).

For my purposes, VAR is more than adequate to uncover the basic causal relationship between news coverage and candidate popularity.

Model Specification

The VAR models estimated here for popularity (Y) and volume of news coverage (X) specify a p-order = 7, which means the ten models (two models for each candidate) looks back 7 days to assess the relationship between Y and X. All variables were measured at the daily level and smoothed using a 7-day moving average. The variables were also differenced in order to meet VAR’s stationarity requirement. This means we are, in fact, testing whether changes in X cause changes in Y and vice versa.

The raw data, SAS code, APIs, model parameter estimates, and VAR diagnostic charts are available for download at:

The Results

The bottom line up front: For three of the five candidates (see Figure 6), increases in their volume of cable news coverage caused small but significant increases in candidate popularity. Also, for three of the five candidates, increases in candidate popularity caused small but significant increases in the volume of their cable news coverage.

Figure 6: Summary Table of VAR Model Estimations

Using Buttigieg again as our highlighted case, the VAR model predicting popularity has a model fit of R-squared = 0.71, compared to an R-squared = 0.39 for the news coverage model. This general pattern is consistent for all five candidates. However, not shown in Figure 7, the inclusion of news coverage volume added very little new information in explaining candidate popularity, as the R-square for all models only fell by 2 to 4 percentage points when news coverage was excluded from the model. This gives me great pause in suggesting news coverage is a dominant predictor of candidate support. Visually, we can see the some causal relationship is there, but, statistically, it still looks like just one actor in a much larger drama.

Figure 7: Summary of VAR Model Fits

One of the most informative outputs generated from a VAR model is what is called an Impulse Response Function (IRF) graph. An IRF describes the changes in the dependent variable along a specified time horizon after a one-unit shock in the independent variable. Both variables — candidate popularity and cable news mentions — are measured as percentages and theoretically range from 0 to 100 (variable summary statistics can be found in this project’s Github depository as part of the VAR output).

Buttigieg will once more be our exemplar.

Figure 8 indicates that a one percentage-point increase (shock) in cable news mentions of Buttigieg leads to a 0.40 percent increase in his popularity four days after later. No other lag parameters are significant in the Buttigieg model (that is, the confidence intervals include zero). That change in popularity might not seem like a game-changer, but cumulatively that translates into a nearly 1.5 percentage-point increase in popularity over 8 days. For a candidate whose support drifts between 4 and 9 percentage points, that is significant shift.

Figure 8: Impulse Response Function: 1-unit Shock in Cable News Coverage and Changes to Candidate Popularity over 8-day period (Buttigieg)

Data Sources: RealClearPolitics and the GDELT Project

Yet, without a measure of the tone of a candidate’s news coverage, the true dynamic between news coverage and candidate popularity is probably under-stated in the analysis here. As Buttigieg’s case makes clear, there are times when even a popular candidate with the news media will need to weather negative news coverage that will hurt his or her standing in the polls. That dynamic must be directly modelled.

My next step will be in further developing and capturing a tone/sentiment measure for news coverage. Presumably, this will significantly improve the model of candidate popularity.

Final Thoughts

Perhaps the most interesting candidates in this analysis are Bernie Sanders and Kamala Harris. There was little evidence of any causal relationship between their volume of cable news coverage and their popularity. It doesn’t surprise me that Sanders doesn’t even get a meager lift anymore from positive news coverage. It does surprise me that Kamala doesn’t. Just a visual inspection of the relationship between her popularity and her cable news mentions reveals what appears to be a strong (positive) relationship between the two (Figure 9).

Figure 9: Candidate Support and Cable TV Coverage (Harris)

Data Sources: RealClearPolitics and the GDELT Project; data smoothed using a 7-day moving average

Harris has three major spikes in cable TV news mentions. The first occurred soon after her candidacy announcement on January 21st — which was mapped closely by a surge in support. One interesting feature of this first surge is that the support level peaked about one week after the peak in cable news mentions. Buttigieg’s chart showed a similar dynamic (see Figure 5 above). Harris’ second cable news mention surge occurred after her first debate performance and was, again, mapped closely by a popularity surge. What is different from the first surge is that the peaks of both mentions and popularity occurred contemporaneously — most likely because the debate was widely watched on television and its impact on popularity more immediate.

Harris’ third cable news mention surge occurred after the second debate but did not witness the simultaneous positive movement in popularity, the reason fairly obvious — the second debate was not a good performance by Harris.

Looking at the charts for all of the major candidates (Figures A.1 to A.5 in the Appendix) makes one thing clear — the dynamics between cable news coverage and candidate popularity varies by candidate and can change over time within each candidacy. Campaign events (e.g., debates), gaffes (Biden is one of the candidates after all), and random shocks (e.g., the economy, mass shootings, the border crisis, the Middle East conflicts, etc.) add a level of randomness and unpredictability that no statistical model, no matter how well specified, can fully anticipate.

This preliminary look at the common daily variations in cable news coverage and candidate popularity — a two-variable model — does not come close to capturing the full complexity of a real world presidential campaign. There are many other factors in a campaign that affect candidate popularity: endorsements, advertising, social media, online news, Google searches, campaign rallies, retail politicking, etc.

We want to believe the news media, as one of the primary gatekeepers through which campaigns try to get information to the general public, is as powerful as the news media itself assumes. No doubt, the struggle of Tulsi Gabbard and Kirsten Gillibrand to get their messages to the voting public is hindered by systematic negative news coverage (or, worse, no news coverage at all). The Tauberg news sentiment data supports the contention that the national news media systematically favors some candidates over others and can crush (or lift) small, outsider campaigns if they so choose.

The news media will argue that is part of their job. If they don’t do it, who will? Voters can’t digest 21 different candidates. The field needs to be whittled down to a more manageable number and the news media is more than happy to provide that service.

“Somewhere, somehow, professional journalists have to decide who gets covered — and any formula they could choose is going to appear biased to someone,” says Columbia University journalism professor Jonathan Stray. “In the end, the candidates who attack the media are right about one thing: The press is a political player in its own right. There’s just no way to avoid that when attention is valuable.”

But the question remains, is the corporate news media an unbiased, neutral party in this process or does it play favorites? Bernie chooses the latter conclusion. I lean that way as well, but I am still surprised at how sketchy the data remains showing a strong causal arrow from the national news networks to candidate popularity.

  • K.R.K.

Comments and requests can be sent to:

All raw data, SQL commands, and SAS codes for this essay can be downloaded at:

APPENDIX: Additional Graphs

Figure A.1: Candidate Support and Cable TV Coverage (Biden)

Data Sources: RealClearPolitics and the GDELT Project; data smoothed using a 7-day moving average

Figure A.2: Candidate Support and Cable TV Coverage (Warren)

Data Sources: RealClearPolitics and the GDELT Project; data smoothed using a 7-day moving average

Figure A.3: Candidate Support and Cable TV Coverage (Sanders)

Data Sources: RealClearPolitics and the GDELT Project; data smoothed using a 7-day moving average

Figure A.4: Candidate Support and Cable TV Coverage (Harris)

Data Sources: RealClearPolitics and the GDELT Project; data smoothed using a 7-day moving average

Figure A.5: Candidate Support and Cable TV Coverage (Buttigieg)

Data Sources: RealClearPolitics and the GDELT Project; data smoothed using a 7-day moving average

Figure B.1: Impulse Response Function: 1-unit Shock in Cable News Coverage and Changes to Candidate Popularity over 8-day period (Biden)

Data Sources: RealClearPolitics and the GDELT Project

Figure B.2: Impulse Response Function: 1-unit Shock in Cable News Coverage and Changes to Candidate Popularity over 8-day period (Warren)

Data Sources: RealClearPolitics and the GDELT Project

Figure B.3: Impulse Response Function: 1-unit Shock in Cable News Coverage and Changes to Candidate Popularity over 8-day period (Sanders)

Data Sources: RealClearPolitics and the GDELT Project

Figure B.4: Impulse Response Function: 1-unit Shock in Cable News Coverage and Changes to Candidate Popularity over 8-day period (Harris)

Data Sources: RealClearPolitics and the GDELT Project

Figure B.5: Impulse Response Function: 1-unit Shock in Cable News Coverage and Changes to Candidate Popularity over 8-day period (Buttigieg)

Data Sources: RealClearPolitics and the GDELT Project

Cable TV news and Google are picking our next president

By Kent R. Kroeger (, August 8, 2019)

Medicare-4-All was the discriminating issue in the last two Democratic debates with Senators Bernie Sanders and Elizabeth Warren the most vocal proponents and the remaining candidates (less Hawaii Representative Tulsi Gabbard and New York Mayor Bill de Blasio) taking up the opposition.

To likely Democratic primary voters watching the two debates, Elizabeth Warren’s performance was particularly strong — though Tulsi Gabbard’s Cool Hand Luke impersonation offered us the best moment from either debate when she calmly exposed Senator Kamala Harris’ record as California Attorney General as incompatible with her presidential stump speech rhetoric. How could Kamala have been so unprepared to defend her own record?

However, most striking about the debate were the candidates relying on historically conservative talking points to dismiss not just Medicare-4-All, but the Green New Deal and college debt forgiveness, among other progressive policy priorities. At one point during the debate, my wife asked if Republicans were also being included in the debate. “No, honey. John Delaney is a Democrat.”

The Twitter accounts of these “Democrats” croon love poems to private insurance companies, market capitalism and ‘choice,’ while mixing in enough disinformation about progressive ideas to truly confound anyone trying to understand the disputed issues.

This is how they talk about Medicare-4-All:

The Republican National Committee couldn’t have written Tweets as dishonest as those about Sanders’ Medicare-4-All plan — a plan a Koch Brother-funded think tank concluded would reduce total U.S. health care spending by about $2 trillion over 10 years, relative to the current healthcare system. That is likely an under-estimate of the savings from Medicare-4-All.

A note to Joe Biden, Kamala Harris, Tom Perez and the Democratic National Committee: You know your party is out of energy and ideas when the talking points for some of your leading presidential candidate could pass for something Bill O’Reilly might write:

If you don’t believe our country can afford universal healthcare for all or that we can get off fossil fuels sooner rather than later, you are not a Democrat anymore. You are a Republican. Which is fine. There is a party for that.

There is nothing wrong with Republicans running as Democrats and Democrats running as Republicans (though I cannot think of a single example in my lifetime of the latter happening…John Anderson in 1980?).

Who has the right to tell anyone what party they should belong to. This is a democracy after all? Right?

Who is driving the Democratic Party train?

Years ago, the first assignment I gave undergraduate students in my introductory political science classes was to describe, in a short paragraph, how the American democracy works.

The students’ paragraphs were earnest, typically focusing on voters and elections and describing a system where people’s policy preferences are reflected, if only loosely, in the policies passed by their elected representatives. Most of the students’ paragraphs were held aloft with Jeffersonian/Madisonian language (‘all people created equal,’ ‘liberty,’ ‘consent of the governed,’ ‘freedom’) and mixed with a healthy dose of American exceptionalism (‘the freest country in the world’).

But there would always be at least one student that would write something along these lines:

We don’t live in a democracy. The corporations pick the candidates and the government reflects corporate interests.

No, Alexandra Ocasio-Cortez was never in one of my classes — She was a toddler when I taught at the university-level.

As a young instructor, I felt a duty to guide students away from cynical views on our democracy and towards something realistic, but also optimistic. In one the class slides I provided a definition of democracy by which to judge the American version:

A government, existing at the consent of the people, who, through elected representatives, satisfy these conditions: (1) Full participation — all citizens are included in the political system and have the capacity to participate, (2) Voting equality — each citizen’s vote is equal to another’s, (3) Full knowledge of choices — citizens understand all available political choices, (4) Popular agenda control — citizens control the political agenda.

The paragraph was just a reformulation of political scientist Robert Dahl’s definition of an ‘ideal’ democracy from his 1989 book, Democracy and Its CriticsAs an exemplar, no democracy could meet Dahl’s idealized version, and that was his point. All democracies are imperfect, some more so than others.

My conclusion at the time was the American democracy, while flawed, fulfilled all four of those observable conditions.

Over the years, however, my views have crept closer to the cynical view, perhaps to an even darker place. And nowhere has my own cynicism grown darker than when observing the corporate news media’s current role in the democratic process.

The recent Democratic debates are illustrative. In the July 31st debate sponsored by CNN, former Vice President Joe Biden spoke for 22.5 minutes, Senator Kamala Harris for 17 minutes, and Senator Cory Booker for 13 minutes. The other seven candidates each spoke between 9 and 11 minutes. These differences are not products of chance. The debate was designed by CNN to give the top 2 or 3 candidates — in terms of polling numbers — the most face time with the 11.3 million Americans viewing the debate. Campaigns will pay around $200,000 for only 30 seconds on a national cable TV network. In essence, CNN gifted Joe Biden about $4.5 million in advertising and to Harris about $2.5 million.

Lesser known presidential candidates already struggle to gain traction with voters and donors. By tilting the process towards the “popular” candidates, CNN is effectively enforcing an Overton window among political candidates.

The cable news media’s sentinel role is apparent when looking at its coverage of the 2020 Democratic nomination candidates over the first seven months of the race (see Figure 1). Over that period, Biden has received 35 percent of all candidate mentions, followed by Sanders (15%), Warren (12%), and Harris (11%). A significant drop-off occurs after that, as almost three-quarters of mentions go to the top four candidates. That may be the maximum number candidates a news network can handle at one time.

Figure 1: Weekly cable news clips from Dec. 30, 2018 to July 21, 2019 (28 weeks).

Data source: The GDELT project

A different relationship exists for online news candidate mentions (see Appendix Figure A.1), as there is no clear leader and the top four candidates are more bunched together. Over the same seven-month period, Sanders received 15 percent of all candidate mentions, followed by Warren (14%), Biden (14%), and Harris (11%).

Is it wrong for cable news networks to be political gatekeepers? Isn’t that one of the core functions of the press? Who wants to learn about John Delaney or John Hickenlooper or Michael Bennet if they don’t need to?

There is a strong argument rooted in microeconomic and decision theory literature that journalists should be filtering out minor candidates and focus their reporting on more viable ones. There are over 20 Democratic candidates, after all. We can’t possibly get to know every one of them with any meaningful depth. The cable networks are providing a genuine service when they ignore some candidates and direct their attention towards others.

“Somewhere, somehow, professional journalists have to decide who gets covered — and any formula they could choose is going to appear biased to someone,” concludes journalist Jonathan Stray. “In the end, the candidates who attack the media are right about one thing: The press is a political player in its own right. There’s just no way to avoid that when attention is valuable.

Yet, why don’t I trust the cable networks and now the high tech and social media companies to make these decisions?

I think back to my most cynical students. How can we defend a democratic system that systematically marginalizes voices outside the mainstream? Perhaps we don’t live in a democracy. Perhaps the corporate news organizations really are the ones picking our presidential candidates and Election Day is simply a tool to legitimize the their choices.

An over-simplification? Obviously. Wholly inaccurate? I think not. And is this power restricted to the corporate news media? Absolutely not, as the high tech and social media giants also have become active participants in the American electoral system and evidence of their power is rapidly being documented by researchers.

Is Google already too powerful?

Speaking at New York City research summit in January 2013 organized by Google, Robert Epstein, a Harvard Ph.D. in psychology and behavioral sciences, said, “Google has likely been determining the outcomes of upwards of 25 percent of the national elections worldwide since at least 2015.”

The same academic, testifying in mid-July before the Senate Judiciary Subcommittee on the Constitution, said that Google’s search algorithm produces a Search Engine Manipulation Effect (SEME) which likely changed vote decisions among undecided voters in 2016 to the net benefit of 2.6 million votes for Hillary Clinton. Looking towards the 2020 election, Epstein testified:

“SEME…leaves people thinking they have made up their own minds, which is very much an illusion. It also leaves no paper trail for authorities to trace. Worse still, the very few people who can detect bias in search results shift even farther in the direction of the bias, so merely being able to see the bias doesn’t protect you from it. Bottom line: biased search results can easily produce shifts in the opinions and voting preference of undecided voters by 20 percent or more — up to 80 percent in some demographic groups.”

Skepticism is always advisable before accepting social science findings, but shouldn’t the burden of proof be on Google, rather than social scientists, to show that the SEME does not systematically alter election outcomes? Can you imagine if pharmaceutical companies were allowed to market drugs to the public before doing research on their safety? No, of course not. So, why should the standard be different for Google or the giant social media companies?

But before we go to Mountain View, California, breaking down the doors at Google headquarters to seize their search algorithms — which we should do at some point — let us not forget the older and probably bigger dog in this fight…television news, particularly the cable networks (CNN, MSNBC and Fox News).

Political science research has generally found that national news outlets are powerful actors in shaping our political environment (Summaries of this research can be found hereherehere and here). Not only do the media expose us to events and people we would not otherwise experience directly, they frame and interpret the political world for us as well — and that may be their greatest influence on our political system: Frame the policy agenda and change the electoral outcome.

But don’t take academia’s word for the importance of the news media on politics — just ask Jeff Zucker, president of CNN, about his news network’s role in the election of Donald Trump in 2016.

While stopping short of saying CNN unduly influenced the 2016 election outcome, in a 2018 CNN interview with former Barack Obama adviser David Axelrod, Zucker admitted his network televised too many Trump election rallies unfiltered and in their entirety. No other 2016 candidate was gifted so much free publicity — some estimates indicate Trump received a $1 billion advantage over Hillary Clinton in free media coverage. More certainly, Trump’s challengers for the 2016 Republican nomination never had a chance to gain traction with voters as long as CNN, MSNBC, and Fox News kept broadcasting Trump rallies live (see Figure 2). In another study, researchers found that CNN, specifically, provided Trump with eight times as much coverage as the number two candidate in the 2016 Republican nomination race, Ted Cruz.

Figure 2: Bought versus Free Media in the 2016 Presidential Election

Graphic source: The New York Times

“I do not believe that’s why he’s president of the United States,” Zucker told Axelrod. “A lot of people want to assign that blame to us and to me. If only we had that much power, especially on the Republican side. I do not believe that’s why he’s president of the United States. But I do think we made a mistake.”

Zucker is being too modest. Along with MSNBC, headed by Zucker’s former understudy, Phil Griffin, CNN showered Trump with attention from the moment Trump came down the escalator at Trump Tower to announce his candidacy. They may not have realized what they were doing — in part, because many at the news networks believed Trump was more of a sideshow than a serious presidential candidate — but by the time it became apparent Trump had a punchers chance of winning the presidency over Clinton, the damage had been done.

In an interview with last year, media critic Stephen Miller said, “There is no figure in media more responsible for Donald Trump’s legitimacy in media, which led to his rise in politics and to the presidency, than Jeff Zucker.”

And you thought the Russians and their $100,000 Facebook advertising buy elected Trump.

Candidate popularity and cable news coverage move in lock-step

A cross-sectional analysis of recent data from RealClearPolitics’ presidential poll databases, cable news content (i.e., candidate mentions) from the GDELT project, and Google search trends data reveals the strong relationship between the cable news media and candidate popularity. And we also see evidence of a relationship between candidate popularity and Google search trends. This first relationship is not news. Data journalist pioneer Nate Silver and other researchers have measured the strong relationship between cable news and candidate popularity.

In my 2020 Democratic nomination data, cable news coverage of candidates explains 95 percent of the variance in average levels of support between Dec. 30, 2018 and July 21, 2019 for the 12 candidates in my dataset (see Figure 3). All candidates are near the regression line — the only exception being Sanders who over performs in popularity given his level of cable news coverage. One explanation for that could be that Sanders is a known quantity given his 2016 presidential run and the news media doesn’t find him as ‘newsworthy.’ However, Biden is perhaps even better known than Sanders, yet his news mentions and public support level are well explained by the regression model. Sanders is most likely different for reasons other than just name recognition.

Figure 3: Cable TV news coverage and Democratic candidate popularity

Data sources: The GDELT project and RealClearPolitics polling database

In contrast, Google search volume, while highly correlated with candidate popularity (see Figure 4), does not offer any new information regarding candidate support. When both cable news mentions and Google searches are included in the same regression model as explanatory variables, only cable news mentions maintains statistical significance (see Figure A.3 in the Appendix below). Google search volume also appears to have a diminishing return on candidate popularity as a search volume’s relative index approaches 100 (its maximum possible value for any single candidate).

Figure 4: Google search volume index and Democratic candidate popularity

Data sources: Google Trends and RealClearPolitics polling database

Interestingly, Google search volume on candidate names generates a divergent candidate ordering from the amount of cable/online news coverage a candidate receives (see Appendix Figure A.2). Since late-December 2018, Biden has received 19 percent of all candidate name searches on Google, followed by Harris (15%), Sanders (14%), South Bend, Indiana Mayor Pete Buttigieg (11%) and Senator Beto O’Rourke (10%). New candidates receive relatively more attention on Google than they do within the news media.

It is important to note that my data cannot tell us much about the direction of causation. Is media coverage causing candidate popularity? Or is candidate popularity driving media coverage? The answer is likely that both statements are true, which could explain why the correlation between the cable news coverage and popularity is is so strong. As a candidate becomes more popular, the cable news media adjusts their own coverage to meet the apparent public demand. Similarly, a candidate’s popularity can rise dramatically when cable news dedicates its coverage to a relatively unknown candidate —Mayor Buttigieg is a good example.

We haven’t even mentioned the role of big donors within the Democratic Party. Harris famously courted Hillary Clinton’s mega-donors in the Hamptons in Summer 2017, a time when she was relatively unknown nationwide. Any understanding of the nomination process cannot ignore where the money comes from and the expectations that follow.

“Maybe Harris has what it takes and will surge ahead of the pack in a few years to win the right to dethrone Donald Trump,” The Guardian’s Ross Barkan wrote two years ago. “It’s too early to tell. But her Hamptons gallivant with Clinton plutocrats is a dispiriting reminder that the Democratic party thinks all can be as it once was, and the status quo isn’t worth being ruffled. Donors can still vet candidates and propel them forward in the press. Anyone beyond the upper crust isn’t a serious agenda setter.”

It is no longer too early to tell. Yes, mega-donors can create viable presidential candidates at a brie and Chablis-filled party nestled in one of the wealthiest neighborhoods in America. Would we be talking about Harris now if the Democratic Party’s donor class hadn’t sanctioned her run? Of course not.

Should we be worried?

I refuse to contribute to the angst industry, so forgive me if I am not worried about Google now, or in the future, controlling our political system. Google picking our president is no more or less loathsome to me than CNN doing it. Cloistered elites have been picking U.S. presidential candidates since the country’s inception — why would it change now?

Still, the possibility of Silicon Valley tech czars deciding who gets access to the nuclear launch codes should not sit well with anyone. Do we resign ourselves to the notion that voters are just ancillary participants, by the Founders’ design or due to their own human limitations, in a political system where their role is more adjunctive than substantive? Are we mere driftwood being carried along in the political currents?

2014 study I’ve cited many times remains my bedrock evidence when arguing that the U.S. is not a democracy if the policies passed by our representatives don’t reflect our collective interests. The study by political scientists Martin Gilens of Princeton and Benjamin I. Page of Northwestern has its critics (as it should), but the daily anecdotal evidence only strengthens its core finding: “When the preferences of economic elites and the stands of organized interest groups are controlled for, the preferences of the average American appear to have only a minuscule, near-zero, statistically non-significant impact upon public policy,” concluded Gilens and Page.

As Johns Hopkins associate professor Yascha Mounk wrote in The Atlantic, “Public policy does not reflect the preferences of the majority of Americans. If it did…Marijuana would be legal and campaign contributions more tightly regulated; paid parental leave would be the law of the land and public colleges free; the minimum wage would be higher and gun control much stricter; abortions would be more accessible in the early stages of pregnancy and illegal in the third trimester.”

At the risk of blaming the victim…

We live in a technology environment that deliberately herds news consumers to a relatively small set of news sites, despite consumers having — theoretically — hundreds of news websites from which to choose. In the process, voters have grown more partisan and less tolerant of alternative points of view.

Regina Lawrence, executive director of the University of Oregon’s Agora Journalism Center and George S. Turnbull Portland Center, cites citizens’ own selection bias as an important factor in the polarization trend. “Selective exposure is the tendency many of us have to seek out news sources that don’t fundamentally challenge what we believe about the world,” according to Lawrence. “We know there’s a relationship between selective exposure and the growing divide in political attitudes in this country. And that gap is clearly related to the rise of more partisan media sources.”

However, search engines like Google also contribute to this cycle by reinforcing people’s news consumption patterns through their profiling algorithms (i.e., your Google search history impacts what you see every time you use Google).

It doesn’t need to be this way and the best place to start changing that dynamic is by regularly putting ourselves in a position to hear perspectives much different from our own. Don’t use Google every time you search for news stories or information. Other search engines exist, such as and (both of whom protect your privacy by not storing your search history).

Information source diversity is our greatest defense against elite domination of our political system. The power of CNN, MSNBC, Fox News, Facebook, Twitter and Google, among others, to control our opinions would be significantly mitigated if we sought alternative news sources.

Forcing yourself to seek out new viewpoints sounds nice in theory but is difficult in practice. It rubs against our basic human nature. Even for myself, as I preach the importance of opinion diversity, my own news consumption patterns in the past three years have grown narrower (…does The Jimmy Dore Show even count as a news show? The fact that this is even a legitimate question indicates how far mainstream news has fallen.).

We need to fight our instinct to drift towards conformity and predictability. If we don’t, CNN, MSNBC, Fox News and Google will indeed pick our next president.

  • K.R.K.

Comments and data requests can be sent to:

APPENDIX: Summary Data Tables

Figure A.1: Weekly online news clips from Dec. 30, 2018 to July 21, 2019 (28 weeks).

Data source: The GDELT project

Figure A.2: Google search volume (on a relative scale with a maximum value of 100) from Dec. 30, 2018 to July 29, 2019 (29 weeks).

Data source: Google Trends

Figure A.3: Linear regression model for RCP poll average (Dec. 30, 2018 to July 21, 2019) with Google search relative volume and cable news coverage as explanatory variables (n=12 candidates).