Google is performing a short-term experiment involving its users’ health-related searches.
The company, claiming it wants to better refine its health-search-related processes by “understanding how people search when they’re feeling sick,” wants to start differentiating between users searching for health-related topics purely for research purposes and those searching for those topics in order to find out more about a personal health issue.
In order to do so, searches for certain health topics will produce a small dialog box at the bottom of the screen, asking the user if they’re searching because they, or someone they know, are experiencing that particular health issue.
For example, if you search for “headache,” Google may ask, “Did you search because you or someone you know has a headache? Yes/No.”
Google says the experiment will be limited.
“Rather than make educated guesses about how many users are searching for because they’re sick, we’re running this experiment to collect real statistics,” Dr. Roni Zeiger, Google product manager, and Jeremy Ginsberg, Google software engineer, wrote in a May 13 corporate blog posting. “This is not a permanent change, but a short-term experiment. A small percentage of random health-related searches will trigger the poll question.”
According to Google, data collected in the survey will “be aggregated across thousands of users” and not associated with e-mail addresses or personal information. The company also claims that the survey data will not be used for advertising, but only to improve health-related search results and refine public health trends based on search queries.
That particular statement on Google’s corporate blog, it seems, is a preemptive attempt to head off any anti-privacy claims that critics occasionally levelwhenever the search-engine giant attempts a new data-mining initiative.
Google’s health-based efforts have been much in the news lately.
In late April 2009, as the world was temporarily seized by fears of a massive swine-flu pandemic, Google used its Flu Trends site to track the spread of the disease based on peoples’ searches for certain keywords.
“We found that there is a close relationship between how many people search for flu-related topics and how many people actually have flu symptoms,” the site explained at the time. “Some search queries tend to be popular exactly when flu season is happening and are therefore good indicators of flu activity.”
Google claimed that Flu Trends could indicate flu activity up to two weeks before its appearance on other surveillance systems. Within days of the outbreak, which killed dozens in Mexico but had relatively little effect elsewhere, Google had posted a Mexico-centric version of Flu Trends that followed flu activity across that country.
At the time, cooler heads argued that up-to-the-minute data collection by Google, as well as social-networking sites such as Twitter and Facebook, had the potential to rapid-fire spread misinformation and panic.