Wisdom of the crowd, or aggregated prejudice?

Roy makes some comments on newish Google features that let you move search results up or down in your own personal search results, and then have that remembered. These bring up some interesting thoughts for me.

I’m guessing Google is indeed hoping that if people do this for individual benefit, google can harvest and process all that data in the aggregate and then use to improve search results for everyone. Yeah.  I mean, that’s what I’d do.

I noticed this a few months ago, myself. I hardly ever use it, but occasionally eliminate an especially useless result from my list. Using this feature requires enabling and keeping your google cookies/login, which I try to avoid when doing google searches (but usually don’t succeed).

It kind of lets you use Google as a an integrated ‘bookmarking’ thing if you do use it. Let’s say I search for ‘brisket recipe’, and the 8th result was actually a really great recipe. I want to make sure to find it again, even if Google’s ranking changes three months later when I search for it again. I can promote it to be the first hit.  The typical user doesn’t know how to use delicious etc.; is grandma going to find this useful to accomplish a sort of bookmarking/tagging feature within Google?

And once I do decide that a particular recipe is especially valuable to me with the tag ‘brisket recipe’ (because what is a search query corresponding to a useful website but a kind of tag?), isn’t this useful information for google’s engines to use for other people’s search results? Especially aggregated with lots of other individuals choices — if hundreds or thousands of people decided that was an especially useful page for the search ‘brisket recipe’ or ‘brisket’ or ‘recipe’, that means something, doens’t it?

I mean, what is social tagging or any other aggregation of ‘social’ data in general but ‘codifying prejudices as if they were science’?  (And what is Google’s _original_ page rank, but an aggregation of individual prejudiced decisions to make links to certain pages with certain names?)

I guess the idea is that statistical agglomeration of lots of individual’s choices becomes science out of prejudice, if you have enough diverse people in your sample.  I’m sure it could be ‘gamed’ — the same way Google ranking ALREADY can be.

The biggest mistake many people make is thinking that Google’s ordinary ranking in the first place is some kind of objective science, unaffected by manipulation and ‘prejudice’.    Google from the start could be accused of “codifying people’s prejudies as if they were science” — and yet, it works out, more or less.  But it’s important to remember what you’re dealing with, especially as librarians, information professionals.

(A while ago I suggested that reference/research librarians should all get training in ‘search engine optimization’, not to exersize it themselves, but so they understand the kinds of manipulation that Google results are subject to, so they understand the field they are on in the first place. )