PDF Print E-mail
Chapter 6 - Computer Concepts and Legal Applications

Full Text Searches
“Regardless of the size of the library, no human will browse more than 20 documents to find what they are looking for.” Paul Nelso, VP of Excalibur Technologies.

By searching a full text document for important words or phrases, you can gather useful and pertinent information about your case. The greatest search advancements are being made in this area. A variety of word searches are at your disposal, such as synonym searches, proximity searches, phrase searches, and wildcard searches. The searching of documents using full text in the practice of law is extremely important. Equally important is the ability to increase the precision of your search so that the number of return “hits” is manageable.

Recall and Precision. The goal of full text searching is to retrieve the information in your case that is relevant. In full text parlance, this is referred to as recall and precision. Recall means to retrieve all information that is connected that could be useful. Precision relates only to those documents that are useful.

These two concepts are inversely related. As the precision of your search increases, the recall of the number of documents decreases. If your request is too precise, then only one or two documents will be recalled. For example, if I search for “car” then I would “recall” all references to car regardless of which witness discusses a car. However, if I search for “corvette” then the search is more precise and the number of occurrences would be significantly lower. The problem with too precise of a search is that if the “corvette” is referred to as a “car” by a witness, then you will not locate that particular reference if you only search for “corvette”.

Every search feature of full text systems supports recall or precision. For example, how close key words are grouped in a document would increase precision but decrease recall. Wildcard searches would increase recall but decrease precision. The importance of this concept is that if your search is too broad, then the number of documents recalled will take a long time to review, whereas if your precision is reasonable, then the number of “relevant hits” will be manageable. Recall and cross-referencing the text with keywords can strengthen precision. For example, some full text software offers you the ability to add synonyms. When you are conducting a search, the synonyms would automatically be included in the search.

Searches. Below are a number of different search concepts and features that assist in your full text searches. Be aware that they are not part of all full text software products on the market.

  • Boolean Searches. Boolean search refers to a search for information using “AND”, “OR”, and “NOT” commands. For example, if you want to search a number of different documents, such as depositions and witness interviews for the terms “driving” and “beer”, then the search term would be “driving AND beer”. If a document contained these two terms, then it would be a “hit” and retrieve the document and the location where the terms are located. The more terms you include in the search with AND such as “driving AND beer AND night AND juvenile”, the narrower the search. If you wish to include more documents, then use the OR command. In this example, any documents containing “beer” or “driving” would be returned. This obviously enlarges the number of documents that will be located.

    If you wish to exclude documents, then you need to use the word NOT. For example, if the important information involves driving and beer but that will return unneeded documents involving a prior boat accident involving beer and driving, then the search phrase (beer AND driving) NOT “prior boat accident” would return only the relevant documents. A quote around the words “prior boat accident” generally directs the computer to locate these three words together. Unfortunately, there is no standard full text engine or command structure followed by all full text developers. Though many use the standard Boolean search format, they will change how to formulate search requests. For example, some developers require that combinations of words, such as General Motors, be in parenthesis or quotes, while others do not require this. The software manual for a particular software product needs to be consulted for the appropriate search syntax to use.
  • Single Word Search - Type in a single word for which you are searching. For example, if you type in the word “fired”, all occurrences of this word in the full text will be located.
  • Wildcard Search - The wildcard symbol (*) can be used with any search to retrieve different conjugations of the same word. Using an asterisk means that any word ending or beginning with the specified letters will be retrieved. For example, use the search request *tion to retrieve all words ending in tion or *auto* to retrieve all words containing the letters auto. For instance, if you type in the search phrase *process*, it would locate the words process, reprocess, processor, processing
 

Find Legal Software


Sponsors






eDiscovery Alerts

Click here to sign up for ediscovery e-mail alerts that provide news on the latest electronic discovery and evidence issues.