Random Data

A database typically assigns some metadata (i.e., information about the data) automatically to each new record, information such as the date on which each record is created and a unique record number assigned by the database. The DALN also assigns a Universal Resource Identifier (URI) that serves as the Web address for that record. As a result, we can use a spreadsheet to build and access a set of narratives randomly, serving as a "control" of sorts for a search-based or "interested" approach to the DALN. Of course, I use the terms "random" and "control" largely outside of their meanings in statistical analysis or experimental research. A random sample of a convenience sample such as the DALN carries no predictive value, nor does it "control" any variable other than our own interests. But therein lies a use for the DALN Random Sets tool, which generates links to ten randomly selected literacy narratives in the DALN. Despite the various and largely inevitable geographic, social, and linguistic biases that characterize the DALN (see below), its size promises more diversity than smaller collections across many axes of difference: race, ethnicity, gender, sexual orientation, and socioeconomic status, to be sure, but also age, religious faith, employment status, education, political views, reading habits, and any of the myriad other ways in which people differ across common characteristics. Pursuing a particular axes of difference in our exploration of the DALN can lead to interpretive biases that random exploration of a sizable collection of literacy narratives may help reveal and offset. We may discover, for instance, that some characteristic we find in a set of narratives related in one way also appears in narratives selected randomly.