The US is experiencing a sea change. We are moving through a fundamental shift in how we think about racism. It is a personal and public journey … maybe even a revolution.
Americans are engaged in self-reflection. We are looking inward and finding attitudes and assumptions that are not consistent with who we want to be going forward. We are educating ourselves by watching free movies and documentaries about racial injustice on streaming services. We are reading books on the history of racism in America.
We are also reconsidering our silence — even apathy — toward racism. We are opting to take public stands against racism, whether with family, friends, or strangers.
Many Americans are intensely upset. They are protesting on the streets, expressing opposition to racism in all its systemic forms. Some are intentionally making themselves arrestable during protests to make their conscientious objections unavoidably visible. And a few are rioting, some to make their seething anger felt around the world, and others to rationalize selfish gains at the expense of others.
What is Going On?
Although it seems like a tremendous amount of change is underway, can we really know what is going on? The New York Times reports that more people are buying books on racism, but they are quick to add: “It is hard to know how deep or wide these responses run …”
Digital research methods (what we call ecosystem marketing analytics) can dig deep into recent events. They can tap into what people are thinking and doing right now. We can discover whether people are really changing by exploring their internet search data.
In this article, I use Google Trends search data. I go beyond basic steps such as reporting on search activity related to racism. I explore the underlying struggle that supports and opposes racism, and how that fight differs for critical US states.
The evidence is overwhelming: There is a sea change underway in America, and the evidence strongly suggests things will never be the same.
Prior Concerns About Racism
Chart 1 shows Google search activity for the term “racism”, either the word itself or expanded versions such as “what is racism?” The chart tracks interest in racism from January 1, 2004 through April 30, 2020. The reason for stopping at the end of April is to ignore, for a moment, recent events and focus on the past.
When people conduct a Google search on racism, they do so because they oppose racism and are concerned about its presence in society. This can be determined in Google Trends by examining “Related queries.” People who search “racism” have compassion for victims. They want to learn more about racism’s prevalence and policies to contain it. Also, racism searches are a strong indicator of a Democratic bias. More on this later.
Chart 1 reveals several patterns. First, interest in racism tends to drop off every year in the summer, perhaps because people are on vacation. The peaks are less consistent but tend to appear in November and April, partly due to elections and primaries.
Second, we can see an ebb and flow of interest in racism. Interest was high at the end of George W. Bush’s first term and then eased off during his second term. Concern eased further during Barak Obama’s first term and increased during his second term. Interest peaked in November 2016 when Donald Trump was elected, and then started to ease again.
Then Something Happened
On May 25, 2020 George Floyd died under the knee of a police officer in Minneapolis. Now the world, and the US in particular, is a different place. The sea change is shown in Chart 2. It tracks racism searches since the start of 2020.
Since May 25, there have been protests and riots. There have been unprecedented stances taken by sports organizations such as the NFL and NASCAR. The Supreme Court has issued surprise, non-conservative rulings despite a conservative majority. Confederate statues have been removed or toppled. The list goes on and on.
While the spike in racism searches was dramatic after George Floyd died, it eased off to about 40% of this year’s peak. However, that 40% level is still over six times higher than the median interest in racism before May 25th. In other words, a dramatic shift in concern about racism is taking hold in the US.
Racial Hatred Just Below the Surface
Standing in direct opposition to America’s concerns about racism, especially toward blacks, is the hatred captured by the search term “nigger”. The power of this search term was initially discovered by the researcher Seth Stephens-Davidowitz and published in his book Everybody Lies.
The comparison of “racism” and “nigger” searches in Chart 3 uses the same time frame as Chart 1: January 2004 through April 2020. Google searches for “nigger” are less volatile than “racism” searches, and both searches started to decline after Trump’s election.
When someone conducts a nigger search, they are acting out of intense negative emotions like anger, resentment, and loathing. For centuries this racial slur has been a highly offensive form of hate speech. Its meaning and intent are clear to nearly every adult in America. A few people use the term for research or academic purposes. The vast majority of its use on Google is to find nasty jokes, offensive pictures, or mean-spirited people who share a hatred of black people.
As Stephens-Davidowitz discovered, states with relatively high nigger search volumes consistently vote Republican. My research has verified and expanded on his finding.
Racism in 2020
When we look at search activity from January 1 through June 19, 2020, we see a dramatic shift in search behavior (Chart 4). While nigger searches increase after George Floyd’s death, the increase is minor compared to the leap in racism searches.
Interestingly, there was a spike in nigger searches when Kobe Bryant died. We can speculate that racists were threatened by the outpouring of grief and admiration for Kobe. On a relative basis, nigger searches peaked at 19 for Kobe but only reached 9 for George Floyd’s death.
The next few charts focus on the gap between nigger and racism searches. While the two terms “stayed in the same ballpark” from 2004 to Trump’s election, a relatively small gap opened after Trump’s election (Chart 3) and became a chasm with the death of George Floyd (Chart 4).
Deep Red States are Shifting
Based on the preceding charts it is tempting to predict a Democratic presidential victory in November 2020. But given the horrible inaccuracy in predicting the 2016 election, a much higher level of caution is in order. First, it is a long time between now and November. Second, the election is determined by electoral votes. So, it is critical to monitor search data as we approach the election. And it is even more important to examine search activity at a state-by-state level.
Chart 5 explores the gap between racism and nigger searches for 10 “deep red” states. My prior research (referenced earlier) shows these states have unusually high levels of nigger searches, and these searches were highly accurate predictors of Republican-voting states in the 2016 presidential election.
The blue column and blue horizontal line represent the overall US, where there was a 37-point shift favoring racism over nigger searches after May 25. This statistic captures America’s recent and intense concern about racism. It quantifies the balance between Democratic and Republican interests on a central issue that has motivated US politics since its founding.
The deep red states all experienced a sizable shift favoring racism after May 25th. The state with the lowest shift was Mississippi. But even in this reddest of red states, racism searches now have a 23-point advantage over nigger searches. After George Floyd’s death, Mississippi residents did not return to their prior search behavior where racism and nigger searches were roughly equivalent. Now Mississippi residents are much more likely to conduct racism searches rather than nigger searches.
Surprisingly, four out of 10 deep red states exceed the national average: Georgia, Tennessee, Louisiana, and Kentucky. Clearly people in these four states were dramatically moved by George Floyd’s death and the subsequent events, as reflected in their search behavior.
Together the four states represent 43 electoral votes. Since Trump’s margin of victory in 2016 was 34, shifting attitudes in these four states could place a Democrat in the White House.
While the swell of opposition to racism might mellow between now and the 2020 election, it seems unlikely. Shifts in search behavior are rarely this dramatic, and the effects of this sea change are likely to impact voting behavior for years to come.
Swing States Are Changing Too
My current research shows 11 states should be viewed as swing states (Chart 6). Trump won these states by relatively small margins, and their search activity only slightly favored nigger searches prior to May 25.
Six of the 11 swing states equal or exceed the overall US shift toward racism: Florida, Ohio, Michigan, North Carolina, Minnesota, and Iowa. Together they represent 94 electoral votes, which is 2.8 times greater than Trump’s 2016 electoral victory margin.
One state, Texas, is only 3 points behind the national average. Its 38 electoral votes also exceed Trump’s 2016 margin of victory.
Democrats Are More Activated
One possible explanation for such dramatic shifts in search behavior is that “silent” Democrats who were passive about racism or cynical about politics are now highly activated. This explanation implies a much higher Democratic voter turnout than in 2016.
Indeed, Democratic voters seem to be more engaged. For example, Georgia recently tripled its voter turnout for its primary elections compared to 2016. This happened in a large swing state despite coronavirus and widespread voting problems such as long lines in minority neighborhoods, insufficient ballots, etc.
Never Trumpers On the Rise?
Another possible interpretation of the shift is that many Republicans no longer tolerate racism. Of course, Trump is infamous for his racist comments and Tweets, so perhaps Republicans are abandoning him for this reason.
I could find no empirical indication that Never Trumpers now oppose racism. For example, even though John Bolton recently said he would not vote for Trump, he still refuses to vote for a Democrat. His reasons for not supporting Trump have nothing to do with newfound support for social justice reforms.
Instead, it seems there is a ceiling of racial intolerance in the Republican party, and it is low relative to the potential for backlash against racists. For example, Chart 4 shows a much higher peak in nigger searches for Kobe Bryant’s death than for the recent protests, riots, and police indictments.
Coronavirus and the Economy
The preceding analysis does not take coronavirus into account. For example, does Trump’s management of coronavirus and the economy moderate the sea change on racism? Further work is needed in this area. Initial indications are that coronavirus pushes Trump further behind, and that favorable ratings on the economy are not sufficiently offsetting.
In the Months Ahead …
The shift in search behavior on racism documented here, although an unprecedented finding, is still new. We cannot know for certain how long it will last and whether it will soon fade. Will prejudice and bigotry break through previous limits? Will Democratic support for social justice reform fade once again and have a limited impact on the presidential election?
As we seek to answer these questions, I caution readers about taking polls and surveys too seriously. It is important to remember that search-based research is far more accurate than voter surveys. They have massive sampling issues and people often lie on surveys. People may not feel comfortable saying they intend to vote for Trump or hate blacks. But their internet searches are much more likely to reveal the truth.
Going forward I intend to monitor search behavior and assess whether the recent sea change in support of racial equality is sustained or offset by other social concerns. For the sake of civility and justice for all, I hope we are at the start of a new era in social justice.