| The case citation is a precise method of marshaling legal arguments in ajudicial opinion. It should be utilized more effectively in full text on-line legal databases as a means of tracking the chains of reasoning used in legal decision making and of improving the effectiveness of case law retrieval. Conventional text retrieval systems are based on inverted word files and provide generic indexing, not contextually related concepts. It is difficult to predict what words a legal writer will used to describe application of a legal rule. The command languages require translating a query from natural language into a formal structure and are difficult for the inexperienced or inexperienced user. The Boolean search does not weight search terms, and it retrieves an unordered set of documents which often contains too many documents for the relevant ones to be identified in a reasonable amount of the searcher's time. This dissertation uses citation analysis in a set of state law cases to determine whether certain predictable patterns of citation practice can be identified which link sets of related documents. It recommends improvements in retrieval and search features available on full-text legal database systems. In order to test the use of contextual citation analysis in case law retrieval, a specific field of law was chosen. An exemplar case, recognized as a leading precedent case with a longer than average subsequent citation history, was compared to above and below average law cases to test normative theories of how case citations are used in judicial opinions. Results indicate that recognized parameters of case law citation such as jurisdiction of court, level of court, age of citation, co-citation, and number of citations near the cited case can be useful for weighting retrieval output. |
