Words are our DATA in qualitative research. We collect the words of others if we interview or use documents and we collect our own words if we observe other’s behavior or are in the field researching.
I’ve done most of my own research in field sites using ethnographic methods.
What I want to focus on in this post is how to ensure that your data is as free as possible of inference, or your interpretation. Interpretation is our analytic process upon the data. And what we strive for is data that is as free as possible from our own inferences or interpretations before we apply the interpretive lens when we begin analysis.
The date we collect needs to be reliable. While there are other terms you could use instead of reliability, and regardless of which term you use, the crux is that without reliable data, we can not claim to make valid interpretations. Therefore, reliability in the qualitative domain means your data reliably represents what happened, what was said, what was seen, etc, without an interpretation. We strive for the data to be as reliable as possible so the data hasn’t been potentially tainted with a layer of researcher interpretation BEFORE the interpretive analysis begins.
Reliability should strive for third-party verification, in theory. While, it would be ideal to have multiple observers, the reality for many in the field, is that it is often a solo endeavor. In this case, we still strive to create data that could be third-party verified- that our collected data represents how another would see the scene.
One way to do this as a solo observer is thru the use of low-inference vocabulary as referred to in Carspecken, F. P. (2013). Critical ethnography in educational research: A theoretical and practical guide. Routledge. This video walks thru an overview of why low-inference vocabulary is a mechanism to ensure, to the best our ability, reliability in collection of observational data.