From the chaos of pre-Hammurabi times to Blackstone to the West Key Number System, doing legal research has come a long way. Founded in 1977, LexisNexis was the first real legal research technology system and became a dominant mode of research used around the world until Google came onto the scene in 1998.
The search engines that we currently use have great potential to “understand” what we’re looking for when we perform searches, but there are still some significant limitations. They retain many qualities of past search engines, but are beginning to learn, in a sense, what we want when running searches. After all, we’re looking for answers, not documents. Two of the most fundamental tactics for combing through vast reserves of information are filtering and sorting. Today we’ll show how these tactics are applied in advanced legal research engines, and how they can benefit your firm.
In most computer-assisted legal research, filtering basically means a keyword search: “show me all the documents that contain these terms.” It can reduce 10 million cases to a much smaller subset of cases that we are truly interested in, but ultimately, it is little more than a good guess.
With a relatively mechanical operation, filtering tends to be both overinclusive and underinclusive: a bit of a paradox. For example, out of 842 search results, only 23 will help you in your case (although you would not know that at first). If you perform a search and narrow the results down to, say, 98, your results are overinclusive because there are too many irrelevant results; and yet, they may also be underinclusive—not all 23 precedents will necessarily have made it into your list of results.
Though filtering may have been considered the means to a successful search 10 years ago, today it is very much a blunt tool. The case precedents that you are looking for are likely scattered among hundreds of search results without a system for locating them, and there is no real algorithm for the process.
Unlike filtering, sorting performs a search and sorts the best ones to the top of the results. This is extremely useful when you are looking for one specific thing as it will usually be brought to the top. However, when you are searching for fairly diverse precedents, sorting is problematic. So, how can we combine the best of both worlds? How do we filter and sort the most important results to the top of the list?
One example of an innovative search engine that goes beyond simple filtering and sorting is Fastcase, a service that looks for documents based on a combination of factors:
- Numerosity (how many times keywords are used)
- Proximity (how close the keywords are to each other)
- Density (keyword saturation compared to total number of words in the document)
- Diversity (a number of varied usages of all keywords, vs. limited usage of single keywords)
Through citation analysis, Fastcase also looks for cases that are frequently cited. If the case is frequently printed, viewed, or read, it also gets scored higher in the search. The result? What’s most relevant to your search ends up at the top of the list—something that was not possible with traditional text-based search results.
Cases are not all the same, nor should they look the same. With Fastcase, search results are no longer one-dimensional and can be sorted based on many facets. Using data visualization, results can now be plotted on a graph, letting you work with multiple search parameters at the same time. See how recent a precedent is, how frequently it is cited, and how relevant it is relative to the other search results, all at the same time. Now that is a truly multidimensional search.
Fastcase’s Forecite is another citation analysis tool that looks at which cases are cited. If a case is cited frequently by the search results but does not itself appear in the list of results, you will have the option of seeing it. It is important to consider all the options available to you; a good search tool must account for people’s different reasons for doing research. As our technology improves, we will see search become more semantic and move beyond simple filtering and sorting.
As law firms continue to evolve and grow, the legal research tools they employ will need to keep pace. Take your legal research above and beyond the traditional search engine in one hour with this free, on-demand webinar, and get a sample template for your legal research plan.