One of the most interesting classes of the Digital Media and Politics program at Stanford was given by Laura Granka, who works both at Stanford and at the Google User Experience Team. She explained how search engines work and how they influence politics and society thru their sites’s ranking.
One of the most interesting things she said is that people have a huge trust in Google, to the point that in an eye tracking experiment results were artificially changed and people kept clicking on the first result. This is a big responsibility for Google, who knows that its ranking has many effects in the information people are getting.
These are the three main issues that search engines have to deal with in this respect:
1. Relevancy in search results. “Search engines’ information retrieval algorithms sort the results by an estimate of the probability that it will fulfill the user’s information need, thereby potentially reducing the cognitive effort and time costs for searchers”, Granka said. Theoretically, search engines could alter results to their interest (or a Government’s interest). But this would get people frustrated when they don’t find what they are really looking for, so search engines would get pennalized if they did so.
2. Diversity in search results. This has to do with having different sources of data and different sites where you can get results from, in order to avoid nepotism. Search engines are also encouraged to have as much diversity as possible, but this has more to do with ethics than with profesionalism. I would say that there are some exceptions, as Google China, which is censored by the Government in order to eliminate certain results. Google thought it was better this way than no giving any results at all to people living in China.
3. Transparency in the way search engines work. This has to do with algorithms, which are obscure formulations. Granka said that search engines are somehow transparent because they blog about their algorithms (for example, what Matt Cutts does for Google) and also use general information retrieval techniques. “People don’t really need a 100% transparency”, Granka said.
(The video is a short interview with Granka)