Naturalistic Data

Naturalistic data can be defined as data that make up records of human activities that are neither elicited by nor affected by the actions of social researchers. The test for whether data are naturalistic is if the social researcher is ill, interviews and ethnographic observations would have to be cancelled, but therapy sessions, parliamentary debates, and everyday phone calls would still take place.

From the start, social science has worked largely with self-reported materials (e.g., surveys or interviews) or researcher-manipulated materials (e.g., experiments). Occasional critics argued for unobtrusive measures, yet in practice this attempt took the form of indirect measures of behavior (e.g., garbage as a cue to consumption habits) or the use of materials such as diaries with little discussion of their analysis. Although ethnographers often collect material through observation, this collection is typically in the form of fieldnotes or interviews that embed researcher categories into the material, making it particularly hard to recover the original patterning of the interaction.

The analysis of records of people interacting was stimulated most fundamentally by Harvey Sacks and the tradition of conversation analysis that he and his colleagues developed. This work started to exploit developments in audio and now videorecording technology.

Advantages commonly offered for working with naturalistic data include the following:

• 1.
  It does not flood the research setting with the researcher's own categories, which are embedded in the questions.
• 2.
  It avoids encouraging participants to provide normatively appropriate descriptions, as interviews often do.
• 3.
  It does not leave the researcher to make a range of potentially problematic inferences from the data collection arena to the topic of study. Claims about health helplines, for example, are not dependent on what callers or nurses say about health helplines.
• 4.
  It can open the researcher to novel issues and concerns that were not predicted at the start of the research.
• 5.
  It is a rich record of people living their lives, pursuing goals, managing institutional tasks, and so on.

Naturalistic data can allow readers and referees of research reports to access transcripts of the material and, increasingly, web-based audio and videorecords.

In practice, a variety of sources of reactivity may arise, for participants will be aware of the recording and may have some idea of the research questions. Moreover, the processes of transcription themselves involve theoretical and analytic judgments that prioritize some phenomena (e.g., overlap between speakers) and downplay others (e.g., regional accent). Although naturalistic data may not be orchestrated by the researcher as in an interview or experiment, it is not entirely independent of researcher's categories and judgments. For this reason, the term naturalistic data is preferable to natural data.

It is likely that social researchers will increasingly work with naturalistic data as new digital technologies make it simpler to capture and work with high quality audio and videorecords of people living their lives.