Data Reference Mystery Series

At the Atlantic DLI Workshop on April 22-23, 2004, Elizabeth Hamilton taught a session called Reference Mysteries. She discussed a model for conducting a data or statistical reference interview, including the types of follow-up queries to be asked of patrons, and then presented the group with a series of real reference questions.

Organized in small groups of three or four people, questions were disseminated to these teams to apply the data reference interview model and to propose materials or sources that would help the patron with her or his question.

The model consists of structuring the reference interview around four basic steps. First, determine if the patron is asking for data, statistics, or something else. Statistics are defined as quantitative summaries, facts, or figures that someone or some agency has produced from data. Statistics are often found in published tables in print or electronic formats. Data, on the other hand, is the raw material that is used to produce statistics. Data require processing to be of practical use.

The second step involves identifying any constraints that the patron has in mind. What is the unit of analysis or the object described by the data or statistics? Has the geography been constrained? Has a specific time period been requested? What are the characteristics that make up the data or statistics in the request?

The third step is to identify further information that you would like from the patron to help clarify her or his request. What is the intended use of the statistics or data? What resources have they already checked?

Fourth, compose a list of the resources or tools that might be useful to the patron in finding an answer. This will include the strategies employed in searching for material and resources as well as the tools used in the course of conducting the search.

This model treats the reference interview as a process and not as an end-state where an answer pops out of the information-gumball machine. The incorrect assumption with many statistical or data questions is that the outcome will consist of a single number or file at the conclusion of the reference interview. While this may be the case in some instances, there are many times when the patron will need to examine the resources that were located to see if an answer can be found.

With Elizabeth's permission, the series of questions that she circulated in her workshop session will be posted on this Web Log.