A Methodology For Documenting A Certain Type Of Hate

HuffPost created a database of hate incidents in which the perpetrator said some variation of "go back to your country." Here's how we did it.

After the 2016 presidential election, which left many in America worried about rising bigotry, HuffPost partnered with the Documenting Hate project, a media collaborative led by ProPublica. The project includes a database of tips: thousands of reports from victims and witnesses of hate incidents from around the country. Combing through that database, HuffPost noticed the prevalence of the phrase “go back to your country,” as if America’s bigots were all speaking from the same script.

We started keeping track of these “go back” incidents, filing them away in a separate database of our own. Our hope was that it might help us examine what hate looks like in the era of President Donald Trump.

Then, earlier this year, Trump himself told four congresswomen of color, all of whom are American, to “go back” where they came from. It felt like our project took on a new kind of urgency.

Ultimately, our investigation into these incidents — which you can read here — showed how ubiquitous hate has been in the experience of so many in America these last few years. And just how traumatic those experiences can be. And perhaps most concerning, we found that in 20 percent of the hate incidents we examined, the perpetrators invoked the name of the president or his campaign slogans.

So how do you show the impact of racist rhetoric on the country?

To start, HuffPost searched for incidents occurring between the launch of Trump’s campaign on June 15, 2015, and June 15, 2019. We searched LexisNexis, a database of media stories, for news articles using the phrase “go back to your country” and other variations of the phrase. ProPublica combed public records it had received from police departments around the country for incident reports containing this phrase.

Lastly, we put out a call out to readers, asking them if they had ever experienced this kind of harassment.

At last, we built a database to house all these stories. When the president himself used the phrase, it marked a sort of poetic bookend to our months of work, bringing attention to its pervasiveness across the country.

Building The Database

First things first, we had to determine what questions we would be asking of the data. In addition to the information that ProPublica provided for us — victim demographics, incident location, date of the hate incident and so on — we also wanted to know more about the perpetrators: their demographics, what “go back” phrase they used, and whether they invoked the president or his campaign slogans. We were then able to write a standardized data dictionary to make data cleaning easier later and so other reporters could be involved.

After HuffPost received hundreds of reports from ProPublica and its 32 project partners, we analyzed each report to see if it fit the criteria to be included in our database.

Then, it was time to talk to the people who submitted their stories.

Verifying Their Stories

With a couple HuffPost reporters working on verifying the database, we kept track of which victims were interviewed with a separate table detailing how many emails and calls each reporter completed in a single day. Most importantly, each case had a status ranking, which ranged from “contacted – awaiting response” to “interviewed – credible.”

In interviews, HuffPost reporters asked victims for more details about the incidents and for any corroborating evidence. After sending out multiple emails and making many calls, each report was assigned a number that corresponded to a document of our interview notes.

While written documentation or video of the incident is desirable for verification purposes, in these incidents, it’s not always attainable. HuffPost decided that a victim’s statements without documentation would be credible if there was another witness there, and based on the amount of detail they provide. Some were taken on a case-by-case basis.

Next, HuffPost went hunting for news. After brainstorming the most commonly used variations of the phrase we’d seen, we settled on searching our predetermined time period on LexisNexis for: “Go back to Africa,” “go back to Mexico,” “go back to China,” “go back to your own country,” “go back to your country,” “go back to where you came from,” “get out of the country,” “get out of my country,” and “get out of our country.”

We read every resulting article, and input them into our database. Because these cases were already fact-checked by reputable news organizations, we did not need to independently verify these cases any further.

Last but not least, we analyzed incident reports from six police departments for sentiments of “go back to your country.” ProPublica acquired these incident reports through public records requests asking for, “Any and all incident reports (including descriptions of the crime) that have been marked or investigated as potential hate crimes motivated by anti-white or anti-black bias since Jan. 1, 2016.” Of those agencies that returned reports, only 16 cases had the “go back” language we were looking for.

Cleaning, Fact-checking And Analysis

Try as we might, building our own database was still messy. Once every story was in one place, we went hunting for errors. Nothing was safe — typos, extra spaces and different terms for the same attributes were all meticulously cleaned and standardized. HuffPost utilized pivot tables, filters and OpenRefine for this process.

Even further, we had to fact-check the database as efficiently and ethically as we could. Every tenth report in the database was fact-checked against the original submission or article for accuracy. We were glad we made multiple copies of the data as we went along for this.

Finally, HuffPost analyzed the data to find patterns and anomalies using pivot tables and SQL. Results were fact-checked twice: once to determine possible story avenues and once for data visualization.

All told, we collected 800 reports of people being told to “go back” over the past four years. Taken together, these reports paint an alarming portrait of America headed into the 2020 election, and the trauma of those who have experienced being told to “go back” to where they came from.

Clarification: This article has been updated to reflect that ProPublica submitted public records requests for hate crime reports, and then searched for the phrase “go back to your country.” The initial request for records was not limited to incidents that included the phrase.

Popular in the Community


What's Hot