Let N-word = No value

THE DEPTH of racism in American life is a malign gift that keeps on giving, a mystery that seems to affect our every social interaction, figure in every equation of our lives and our fortunes as Americans. In a new study, a Harvard researcher suggests that we go to Google to discover just how deeply racism is entrenched in the national discourse.

But the foundational data in his findings presents a Brownie snapshot when a digital video is called for, a sharp image of one moment in time that offers us, at best, an outdated perception that does little or nothing to tell us conclusively where we are today. There’s good reason to think there are some things a search engine can’t help you find.

In the research paper released on June 9, Seth Stephens-Davidowitz, a Harvard University Ph.D. candidate in economics, finds that, by using the N-word as a proxy for racial attitudes, the Insights tool of the Google search engine can be used to offer a meaningful projection of the impact of race on the 2012 presidential election.

◊ ◊ ◊

The findings, he said are "based on the hypothesis that individuals with higher racial animus will be more likely to type the racial epithet into Google."

“Racially charged search, in contrast, is a robust negative predictor of Obama’s vote share,” he writes in the paper.

Stephens-Davidowitz distilled his findings in a June 9 piece in The New York Times. It’s the Times piece that’s the basis of this response.

Stephens-Davidowitz found that racial animus cost Obama between 3 and 5 percent of the vote in 2008. He won the election anyway, with 53 percent of the popular, and with a seven-point bulge over Arizona Sen. John McCain.

Stephens-Davidowitz uses the clinical language of research throughout the study, and somewhat less so in the Times report, but he avoids addressing the more disquieting question his research makes inevitable: As bad as the impact of racial animus may have been in 2008, how much worse is it likely to be now, with three-plus years of Barack Obama as a known presidential quantity, with three-plus years of a sputtering economy, and with two years of a passive-aggressive racist political meme created and cultivated by the Tea Party wing of the Republican Party?

We barely submerged racial animus as a catalyst for political action four years ago. Will our better, less reactionary, more analytical angels step up to the plate again, this November?

◊ ◊ ◊

Correlation, of course, isn’t necessarily causation. It’s not a good idea to follow Stephens-Davidowitz’s findings out the window. There’s reason to have a healthy skepticism about its conclusions. The first problem is the most obvious: Not every search for the N-word and its variants is a hunt for kindred intolerant spirits; it’s a certainty that some or many of those searches (we’ll never know how many) were conducted by researchers, scholars, standup comics looking for material, organizations that monitor hate speech, or the simply and indifferently curious.

Until Google Insights is able to filter these searches on the basis of someone’s political or philosophical identity, what we’re left with is a compelling, even provocative societal assessment driven by numbers that, ironically, don’t quantify anything dispositive.

Another problem is that Stephens-Davidowitz apparently has no control for the age of people who input the N-word in their Google searches. That’s crucial when you’re trying to determine the mindset of American voters in an election. Since there’s no age limit on who can use Google to search for anything under the sun, we can’t know who among the universe of N-word searchers was even old enough to vote in the first place. That’s a hole in the study’s methodology that Stephens-Davidowitz can’t patch.

◊ ◊ ◊

The researcher betrays a fascination with Google Insights’ information-gathering potential that’s almost breathless. Defending his use of the Google Insights tool, for example, he writes that “Google, aggregating information from billions of searches, has an uncanny ability to reveal meaningful social patterns. ‘God’ is Googled more often in the Bible Belt, ‘Lakers' in Los Angeles.”

For anyone who’s visited or lived in either of those regions, that’s not even an insight, Google’s or anyone else’s. It’s hardly “uncanny,” it’s pretty much common knowledge, and what you’d expect given the quotidian concerns and diversions of life in those regions of the country.

Christian religion has for generations been a social and cultural foundation of life in the Bible Belt states; and in Los Angeles, the second-biggest media market in the United States, the relatively high winning performance of the Los Angeles Lakers (five NBA championships since 2000) made discussion of the team’s fortunes an everyday thing for years before Google even went public.

◊ ◊ ◊

BUT STEPHENS-Davidowitz most deeply undercuts the value of his own findings when he writes that “I used data from 2004 to 2007 because I wanted a measure not directly influenced by feelings toward Mr. Obama.”

Stephens-Davidowitz anchors his findings on a prediction of how well Obama would have done if racial animus had not factored into voters’ decisions in 2008. He says his conclusions are “based on how many votes John Kerry received in 2004 plus the average gain achieved by other 2008 Democratic Congressional candidates.”

But that prediction of an ideal Obama election performance (votes without racial influence) in 2008 is based on data culled from the years between 2004 and 2007 — plus Stephens-Davidowitz’s best educated guess.

How does he make any real linkage between what Americans are thinking vis-à-vis race and Obama’s chances for re-election today and what they thought years before the senator from Illinois came on the national scene?


The researcher gives us speculation about how the absence of racial animus might have altered what’s already happened, and he bases it on data from years before it happened.

It would be more valuable if he used known search data about race animus toward Obama in 2008 as a predictor for the outcome of the 2012 election. That comparison, closer to like-for-like, would make more sense. If you’re trying to give people a meaningful “predictor” of how racism could affect the current election, why not use such search-related data extracted from results of the presidential election that preceded it?

◊ ◊ ◊

STEPHENS-Davidowitz uses unfiltered antecedent information and a statistically-based surmise to buttress his view of a contemporary national perspective. The results of such a study can only be as current as, and no less speculative than, the data used to create it.

It’s of course a sad given that race has played and will play its grim, subversive role in American presidential politics; the conservatives have seen to that. But this study, as boiled down in The Times, suggests strongly that Stephens-Davidowitz is recording a current census based on old addresses.

The boogeyman of racism will make its presence known; a more meaningful survey of that boogeyman’s place in the national political discourse would use data from now to tell us what we’re thinking now, would speculate on a future election by including data from the previous one, and it would preferably use more emotionally granular, informative data than found on a search engine in a hunt for use of what may be our ugliest national word.

---

Image credits: Google logo: ©2012 Google. Table: Extract from the study by Seth Stephens-Davidowitz. Lakers logo: © 2012 Los Angeles Lakers/NBA. Obama: Pete Souza/The White House.