Research in Library Science is conducted in many areas covering multiple questions, but one thing shared is data collection. Qualitative and quantitative information to support the question at hand are necessary to validate the needs or phenomenon or trends (Wildemuth, 2009). Transaction logs and focus groups are two valuable data collection techniques.
Whenever a person logs onto and begins to use a computer in the library, different kinds of information are automatically collected into transaction logs (Jansen, 2006). Sullenger (1997) recommends transaction logs “be examined by librarians to analyze how patrons use the catalog, what features they are using, and to see what areas of searching are problematic” (p. 21). Data can also be collected on “items viewed, sessions, site penetration; time online, users (trace evidence of, not individual information), navigational information” (Nicholas, Huntington, Jamali & Tenopir, 2006, p. 121). These data pieces provide useful information on usage patterns (Das & Turkoglu, 2009).
Transaction logs can be generated in two ways. The first is from the server’s side. These logs include data typically already collected on in-house. Data can also originate client-side using a specifically-written program to collect from the participants’ computers (Wildemuth, 2009). The former is more often used due to the abundance of data and less-costly features. Jansen (2006) describes a three step process to using transaction logs: data collection for a given period of time, preparing the data, and data analysis. He further breaks analysis into three parts: term, query, and session.
A major benefit to using transaction logs is that this is data already collected and waiting to be analyzed. This data is also free of biases that humans may interject (Jansen, 2006). Programs have been developed to help with data analysis. When information is uniform, data from multiple systems can be compared (Sullenger, 1997).
Quite a few challenges to using transaction logs exist. The first is the extreme amount of data available. The data have to be cleaned to rid the files of extraneous and incomplete information (Das & Turkoglu, 2009; Sullenger, 1997). The data are complex and take a great deal of time and effort to analyze (Jansen, 2006). Coding has to be developed and applied to see the patterns (Wildemuth, 2009). Due to this complexity, researchers need to carefully define terms and measuring devices (Jansen).
Transaction logs can be used in a variety of ways but some common ones are developing useful web page design (Das & Turkoglu, 2009), search engine usage (Jansen & Spike, 2006; Sullenger, 1997), as well as searching digital libraries and identifying browsing patterns (Nicholas, Huntington, Jarnali & Tenopir). Overall, the researcher is looking for algorithms and patterns that naturally generate from computer or system usage in the library. If a topic involves human interaction with the...