The root of my newest obsession: this article. The author, Seth Stephens-Davidowitz, a PhD student in economics at Harvard, innovatively uses location-specific Google search terms (Google Insights) to estimate that Obama lost 5 percentage points in 2008 because of “racial animus” or racism. How does he do this? Well, since self-reported surveys poorly capture racism, he looks at use of the n-word in Google searches as a proxy for an area’s level of “racial animus.” He then compares an area’s racially charged search terms to its votes for Obama, controlling for its votes for John Kerry in 2004. Although we could definitely argue whether Google searches including the n-word adequately capture racial animus, there appears to be something to this. To a methodologist, it presents a fascinating example of attempting to measure something that appears to be unmeasurable.
Davidowitz isn’t the only one using Google search terms for social research. NPR recently did a piece on sociologist Phillip Cohen’s Google Correlate explorations. Cohen has some interesting analyses examining the correlations among search terms and geographic space, such as how Google searches for certain political commentators such as Rachel Maddow or Rush Limbaugh correlate strongly with other “odd” search terms (like fennel salad) and also adhere to red/blue-state patterns. Is your interest piqued? Here are more of Cohen’s interesting Google correlations.
Guess which terms correlate more strongly with “cadmium” (a chemical element in the periodic table)? Here’s the answer. Hmmmm….interesting?
In sum, I agree with Cohen that “Someone — probably not me — should get serious about using this kind of data to connect search behavior with demographic trends, politics, culture, and other aggregate patterns of social behavior.” Any takers?