Blue Heron in flight

Sunday, March 31, 2013


Can you trust a company, or for that matter a President, to be good and not evil because they tell you that they are or are not? A very good question. Could the company Google affect election results if they wanted and had the desire to do so, maybe in a subtle way that could never be detected? Absolutely. Consider the following matter:

Robert Epstein is a psychologist, the former editor in chief at Psychology Today magazine. He got in  a public kerfuffle with Google last year over page rankings and a malicious blog message. You can read the whole story here in the Washington Post but I will try to give it to you in a nutshell.

Epstein decides to create an experiment. He creates a fake search engine called Kadoodle and conducts a series of tests. Participants are shown search entries for a 2010 election in Australia and asked to rank candidates. Roughly 65% of the group favored candidates based on their internet search rankings. Epstein is presenting his research  at the 25th annual meeting of the Association for Psychological Science, Washington, D.C., May 2013. It it titled Democracy at Risk: Manipulating Search Rankings Can Shift Voting Preferences Substantially Without Voter Awareness. The following is the abstract from the study:
In a controlled experiment, web pages and search engine results from an actual election were presented to three groups of eligible voters. In two of the groups, rankings favored one candidate or the other. Preferences shifted dramatically toward favored candidates,with 75% of subjects showing no awareness of the manipulation. In a second experiment, voter preferences again shifted in the predicted direction,and the proportion of people who were unaware of the manipulation was increased by slightly altering the rankings to mask the favoritism. In a third experiment, a more aggressive mask was used to hide the manipulation, and no subjects appeared to be aware of it, even though voter preferences still shifted in the predicted directions. We conclude (1) that the outcomes of real elections—especially tight races—could conceivably be determined by the strategic manipulation of search engine rankings and (2) that the manipulation could be accomplished without people being aware of it.
Google is denying that they would ever engage in such actions and I have no reason to doubt them. But search rankings are being manipulated and are extremely important to people. Could Google be victimized or manipulated externally?
“Elections are won among low-information voters,” said Eli Pariser, former president of ­MoveOn.org and the author of “The Filter Bubble: What the Internet Is Hiding From You.” “The ability to raise a negative story about a candidate to a voter . . . could be quite powerful.”
Could Google's search results be manipulated from within, if it was in their economic interest and they had an inclination to do so? I finished the Dan Ariely book yesterday and found the chapter on cheating informative. People do cheat but they tend to cheat only a little bit, because the superego likes to reinforce broader social dictates. Perhaps future leaders of Google will have a slightly different moral composition than today's executives and take a different tack? Can you appreciate the ways that a person or entity could be harmed if they managed to run afoul of Google and let's say for the sake of argument, put on an internal enemies list?  One can only wonder what kind of controls are in place to avert this sort of behavior which can have dramatic impacts on many facets of our lives, not just elections. I am sure that a Google "push" for a consumer item could have a dramatic effect on a marketplace.
"The key lesson may be that search engines are not mere machines spitting out perfectly impartial results. They are driven by decisions, made by people who have biases. This does not necessarily make them evil — merely human.
“The more trust we give to these kinds of tools, the more likely we can be manipulated down the road,” said Panagiotis T. Metaxas, one of the computer science professors at Wellesley College who studied the Massachusetts election. “We need to understand, as people, as citizens, why we believe what we believe.”"
I am taking an online course in behavioral economics. It is interesting but much more work than I had anticipated. I watched a lecture today by Duke researcher Gavan Fitzsimons on psychologic resistance. He says that our ability to defend ourselves against subconscious influence is really quite limited. We humans are not very inclined to make rational decisions, no matter our best intentions.

Of course there are other search engines and other news media outlets that provide information. Although Google is by far the search engine of choice, perhaps we exaggerate their power and dominance. But even still it is easy to see how even a little search rank tweaking could affect not only elections but people's perceptions in general.

No comments: