Skip to main content

Qualitative Data” is Experience

Being “data-driven” is a common refrain - often followed by an insistence that people consider both “quantitative” and “qualitative” data. The subtext is that, on their lonesome, numbers are incomplete, and possibly cold and insensitive. Human intuition, and compassion, is necessary to supplement dry quantitative analysis.

Logic is, in fact, incomplete - though not really in the way most people think. And the human perspective is integral to many decisions (as they impact humans) and should not be neglected.

But the best way to balance “quantitative” data (henceforth simply data) and human needs and perspectives isn’t to conflate the two. Doing so dilutes numerical explanations while skewing opinion, leading to an overconfident “data-driven” rationalization of what is actually a distorted analysis.

Data (in the modern “data science” sense of the term) is a series of observations, consistently recorded facts and information. It is inherently quantitative (even if it is categorical, messy, or unstructured), as each observation can be enumerated, and on that basis some form of statistics can be calculated (at least counts, aka quantities, for unstructured data - and that’s before applying whatever clever techniques to engineer proper features).

Qualitative information (N.B. not data) is about qualia - a concept even fuzzier than NLP and unstructured data. You can (and should!) dive into the sundry academic perspectives, but for a pithy contemporary take:

Tragedy is when I cut my finger. Comedy is when you fall into an open sewer and die.” - Mel Brooks

Comedy and tragedy theater masks

Apprehending qualia is personal, internal, and private. We may strive to describe the sensation of burning our tongue on hot soup or seeing the color purple - but a part of it remains ineffable. We rely on metaphor and appeals to empathy in the hopes that others can understand what we’ve been through.

Qualia are carried by experience, and specifically the subjective elements of it. And so, “qualitative data” is actually just experience. It’s not in any way commensurable to actual data - it’s innefable, and thus most definitely not enumerable.

Experience is valuable. But it is understood and shared in ways fundamentally different than with data. A major point of discussion on qualia is if they even exist. That takes a harder stance than is relevant here, as we are focused on the social function of experience, and not its epistemology or metaphysics.

But the uncertainty of qualia is a useful reference point. You’re probably pretty sure your qualia exist, but what about everyone else’s? Even if they exist, your best shot at understanding them is through comparisons to your own experience.

Making decisions (and thus impacting others) based on qualitative experience is an inherently selfish act. This does not mean it is always bad or wrong - but it does mean that, when you allow those factors to weigh your thinking, it is best to be mindful of the subjectivity. Don’t rationalize by treating feelings equivalently to numbers, and then massaging the whole thing together into a justification of what you wanted to do anyway.

Being data-driven doesn’t mean every decision is based on data - most things we care about are quite hard to measure. We should strive to use data responsibly, correctly, and when it is available and relevant. When decisions are beyond numbers, then absolutely consider the qualitative aspects, but in a way that acknowledges the subjectivity and limited perspective we all have, and owns the resulting decision as a result of that imperfect process.

If you’re interested in principled ways to apply lessons from qualitative experience, don’t just call it “data” - instead look at case studies. Approaches to case studies are varied and not without flaws, but they are part of an established scientific tradition and, properly done, recognizes the differences between “big N” and “small N” information.

Properly diambiguating between data and experience, and employing both in ways mindful of their strengths and limitations, is the difference between a data culture and a data cult.