LibrarySearch Engines
Limitations
Size
too much information
A search for Barack Obama in December, 2008, retrieved over 106 million pages. When you are doing research, more is not really better—you could never review all of these pages, and most likely you will run out of patience within the first few pages of results. The enormous indexes that search engines build can backfire on us if we aren't able to narrow things down enough.
Relevance judgments
Search engines see your keywords, but they don't understand the context of your research. When a search engine's spider evaluates a web page, it is trying to get the sites that best match your keywords and have some kind of value. It cannot really know the quality of the sites, though, so it uses other signs of value. For instance, it might employ some of the following criteria:
- Do the search terms occur in the URL or title of the page? If so, the search engine assumes it is more likely to be relevant.
- Web developers can also insert a description and keywords that are designed to be read by spiders; these can influence a judgment of relevance. Other aspects of site design can also influence how the search engine ranks the site as well.
- Search engines often determine the value of a site based on how many other sites link to it. Google calls this determination PageRank (they've even trademarked the term). If a site has a high PageRank, it will appear higher in your search results; and when a site has a high PageRank, the sites it links to benefit as well.
It is important to understand that although these criteria can work very well for many purposes, they can be problematic for academic research. When you are looking for materials for a class, you are looking for reliable, authoritative, accurate, relatively unbiased materials, and the major search engines simply do not have a way to judge websites by these criteria.
What is more, these relevance judgments can be manipulated by savvy web developers. In fact, there is a whole industry that has grown up around Search Engine Optimization (SEO)—strategies for getting websites to be placed higher on results lists. And there are even cases of "google bombing," or manipulating the search engine to achieve certain results. A few years ago, a search for "miserable failure" returned as its top result the White House's official biography of George W. Bush. So, although appearing high on search results isn't meaningless, it is no guarantee of a site's quality, credibility, or authority.
And be sure to watch out for areas labeled "Sponsored Links," "Sponsored Results," or something similar. These are simply advertisements; they may or may not lead to good sites, but the only reason they are appearing where they are is that the website owner has paid the search engine to achieve that result.
Some search engines are trying to separate out more reliable results. Hakia, for instance, gives you a category for "credible sites" for some queries, and Scirus, a search engine for scientific topics, has a category it calls the "preferred web" (click the image on the left to see how it marks these results). These tools can help you, but the criteria they apply in making these distinctions is not always rigorous—and, since it is done automatically using an algorithm, you will still have to approach the results skeptically.
Gaps
Even though the major search engines have indexed millions of pages, many pages are missing. There are various reasons a page might not be in the index, including:
- the page is new, and the spiders haven't found it yet
- few sites link to it
- it is part of a password-protected site
- the author has included code instructing search engines not to index it
- the site is not designed in a way that the search engine's spiders can easily crawl and index.
According to recent studies, web search engines are still not adequately indexing U.S. government sites and collections of scholarly literature.
