Monday, June 26, 2006
Alejandro Diaz has written a better honors thesis ("Through the Google Goggles," a study of search engine bias) than you or I could have written. Really food for thought. (It was written a year ago. I just came across this through an article on Aaron Wall's site.)
In essence, it elaborates at length on a few paragraphs written by Page and Brin themselves, who anticipated bias in both commercial and non-commercial search results based on the fact that search engines themselves are profit-making enterprises.
Make no mistake: it would take an idiot to believe that large media conglomerates don't have interests. I notice even the little things, myself. I might even be a bit of a conspiracy theorist. Let's just say that I don't think MySpace is going to be getting a lot of favorable press by any of the media corporations not owned by Rupert Murdoch, anytime soon. First there was the parody on Saturday Night Live that "exposed" it as a place for perverts to hang out. And since there have now been a lawsuit and an impending marriage of a 17-year-old American girl and a 20-year-old Palestinian who have never met in person, I wouldn't expect the salacious tone of the coverage to abate. Yep, there is bias in what media outlets choose to report... especially about one another.
Media bias is something a fair number of postsecondary, undergraduate, and even graduate students spend time studying, of course. As forms of media have morphed, the subject seems to get more difficult to stay on top of. Much of what we see online today (like what you're reading now) is much shallower than traditional news reporting. Yet there is so much more of it, and a variety of ways of accessing it. For now, the searcher does seem to have a lot of choices, but the recurring worry in Diaz's thesis is that non-controversial spins on certain topics are more likely to find their way onto the user's screen. "Adversarial" voices may be crowded out.
The natural response is: do search engines enforce, or merely reflect, social biases?
Recall that PageRank is a kind of hypercaffeinated online analog of academic peer review. John Rawls, the most-cited philosopher of our age, may not be the "greatest" or the "best" philosopher, I always said when I was a keen student. But the citation indexes told us he was the big guy. And someone like Foucault, brilliant, controversial, and shaven-headed, came into his own for different reasons. Like a rock star... he got cited plenty when his time came.
The filtering process by which respectable opinions came up through respectable journals produced by academics at respectable universities would tend to support the relatively "mainstream" liberalism of Rawls, and that happy scenario supported thousands of little Rawls' and hundreds of thousands of senior-level students who needed contemporary, big, thick ideas to read alongside Aristotle. Everyone wins, except controversy, perhaps. Except when it's the kind of controversy that sells (thus, Foucault). Diaz, with his Stanford degree, implicitly buys into this process, with his respectably controversial thesis. Alejandro Diaz is highly marketable, as are some other well-known (Harvard educated no less) critics of search engines.
These types of questions could lead to endless circular debates about what truly has "substance" and what doesn't, and about what material is being "hidden" from us. Material is actively hidden from people in repressive societies, to be sure. But I'd argue that there are two main barriers to balanced, substantive fact-finding efforts, using search engines or other media sources for that matter. (1) Research skills; (2) The funding or encouragement of research and dialogue on relevant topics.
On the first point, no one should ever expect to get a deep take on controversial topics from a basic web search. I mean, no one ever thought Reader's Digest or The Economist were the last word on life, did they (as opposed to useful starting points or just fun reading)? Did they? Gulp.
Cue: the invisible web. How ludicrous is it to use a basic web search engine for really looking into an important policy question? Try typing some complex topic into Google News. Of course, you'll only see two or three recent stories. Nothing archived. Nothing from better academic journals, etc. You might do a bit better if you used Google Scholar. A lot better with a multifaceted search of periodicals and newspaper archives over a long period of time. Not that you or I have time for that sort of thing, but someone should.
Even here, there is no shortage of variety in search results if you type, say, "rising tuition fees in Ontario" into Yahoo Search. A variety of perspectives, including research papers (including some in Word format), are included on the first few pages of SERP's. The reason some other material might not be as prominent might be as simple as: the relevant organization did not put it online; the relevant organization isn't advertising in the margin, as a lead generation method so they can attract supporters and fund further research; the relevant organization hasn't prioritized its research program; the relevant organization had its funding pulled; rival organizations have done a better job of gaining mainstream media attention, which influences search rank. (That's the real toughie, to me: "legitimate" spokespersons and research papers get cited, and rival organizations may be relegated to the wilderness, by journalists. That's an issue for j-school or for soul-searching by news organizations -- not for search engines.) So this is a mixed bag. Search engines do reflect media bias and societal bias to some extent, but there is room for users to bust through this if they simply refuse to stop at one or two searches.
On (2), then -- if political battles are waged at the level of funding and de-funding organizations with different points of view, then it's tough to imagine how a search engine might fight this. Eric Schmidt (as cited in Diaz's thesis) may be too disingenuous by half when he explains that there is no news bias at Google because these are "just computers." (I once heard a Google News product manager say the same thing.) But he has a point. Google isn't a magazine. It doesn't create editorial content, or tell you what to think. It indexes pages, and ranks them. You can change the methodology of indexing and ranking, but that's always what it will do. Of course the algorithm has a bias, even if that bias is coming from some measure of social authority or acceptability, or user interest.
"Controversial folk" -- even those rotting in jail for decades -- have always found a way to get to the heart of matters, and have often found ways to disseminate their points of view, in circumstances far more constrained than those enforced by today's search companies. At the end of the day, the concern seems not to be that "we, the savvy searchers and holders of postsecondary degrees" will be duped when we type generic queries into the wrong search tool for our purpose, but rather that "they," the poor unsuspecting consumers with the AOL dialup accounts and small vocabularies, will have their minds warped by giant corporations (in a way that "we" never could, because we "get" how the SERP's are constructed). Perhaps. And that leads to a voluminous literature and some action on awareness-raising on behalf of the downtrodden that dates back to the late 1960's. :) It's just that we have bigger fish to fry this week...
A bolder act than writing a paper on news bias, then, might be to take the Church Lady act into society. If search engines exhibit bias, it's often based on the money and power "out there" that foster popular points of view and crowd out others, as opposed to the money and power interests of media organizations themselves. But yes, there's that, too. Without media literacy, we're all blithering idiots. Let's just hope the unsuspecting masses are savvier than "we" give them credit for. And that institutions of higher education will continue to teach deep research skills so people can uncover a variety of perspectives on problems like, say, click fraud. Did someone say search engines only present the "sunny side" of issues?
Good night, and good luck. Again.
View Posts by Category
Andrew's book, Winning Results With Google AdWords, (McGraw-Hill, 2nd ed.), is still helping tens of thousands of advertisers cut through the noise and set a solid course for campaign ROI.
And for a glowing review of the pioneering 1st ed. of the book, check out this review, by none other than Google's Matt Cutts.
Posts from 2002 to 2010