Skip to main content

University Library, University of Illinois at Urbana-Champaign

Ways of the Web: Filter Bubbles and the Deep Web: So what?

This guide is a companion to the Ways of the Web Savvy Researcher Workshop.

The Pros of Filter Bubbles

Filter bubbles are not all bad. Here are some pros of filter bubbles:

  • Personalized results provided by websites and search engines can give us results that we want faster.
  • Websites that we commonly go to are easier to find, as they appear higher up in search lists.
  • Location detection is helpful when you simply want a restaurant that is close by. 
  • Our login information can be saved so we don't have to keep retyping it, if we are on a private computer.

Information Literacy

It is important to consider Filter Bubbles and what they mean for Information Literacy. If search results are skewed and we are unaware of it, this affects our ability to access, evaluate, and use information. We need to know if search results are biased in order to be critical in our selection of information to incorporate into our value system. To be effective at information access, we need to be aware of when certain information might not be readily accessible to us. Here are some frameworks to contextualize this problem:

Is the Filter Bubble real?

Some people might argue that effects that Eli Pariser calls the "filter bubble" are overblown. Pariser's argument is based on anecdotal evidence and algorithms - how they work - rather than scientific evidence about how personalization on the web affects how we think. Yet, there is empirical evidence showing that the effects of personalization on our search results are very real and significant. While more research is needed in this area, anytime results are personalized to fit our clicking history, there is the chance that some other important information will not reach us - reason enough for concern. Here is a challenge to Pariser's argument:

Google's PageRank

There is another reason to be wary of search results and question their objectivity. Google's PageRank - which refers to the system for ranking or ordering search results - is based not on the quality of information on a diagram showing how Google PageRank workswebsite, but rather on the website's popularity, that is, on how many other websites link to that particular website, and how high those websites are ranked.

Information literacy is about selecting and evaluating information for its quality, but if we accept search results in the order in which they appear, we are really getting information that is most popular, not necessarily high quality. Of course sometimes the most popular websites are great sources of information, but not always.

To the right is a depiction of PageRank, which is based on what are called "backlinks" - the pages that link to a particular search result. Each circle represents a web page, and the arrows represent links to that webpage. The percentage represents the rank of the page. As you can see, the smaller circles - the ones with the lower rank, have fewer pages linking to them, or the pages linking to them have low percentages. The pages with higher percentages have lots of other pages linking to them, or have other pages with high percentages linking to them. For example, the orange page does not have very many pages that link to it. However, the page that links to it has a very high percentage, or rank, relative to the other pages. Thus, orange also has a high PageRank.

So, in evaluating a website, it can be useful to know what webpages link to it. A tool for analyzing a website based on its backlinks can be found at alexa.com (See the link below.)

Image: http://commons.wikimedia.org/wiki/File:PageRanks-Example.svg#mediaviewer/File:PageRanks-Example.svg