Research, Creativity, and Data Saturation

Research is the cornerstone of nonfiction, from business reports to academic meta-analyses. At some point in a project, and probably around the beginning, we have to pore over the books, click through the articles, and immerse ourselves in the facts and figures. At a later point, we have to decide when to quit reading and start writing.

That’s where data saturation comes in.

We’re talking specifically about qualitative research that starts to yield no new information, whether it’s from the literature or firsthand data collection (e.g., interviews, lab experiments).

An extremely unhelpful professor of mine told me that to prepare for my oral exams, I needed to “know everything about everything” and be prepared to talk about it. What he should have said is that I needed to recognize when I had reached data saturation.

So, how do we know when we’ve gotten to the point when we can gather our notes and start adding our own contributions to the field? It’s tempting to ask how many experts we should know the names of and how many articles we have to read.

But it’s not about the amount of data. It’s about our relationship with the data.

Sign #1: We’re Not Learning Anything

The first sign is the most obvious: we’re reading or hearing the same things over and over. Instead of taking notes with new ideas or learning new jargon and theories, we’re just nodding along thinking, “Been there. Done that.”

Moreover, it’s not just the facts we’re seeing repeatedly but also the various opinions of those facts. Say we’re looking into climate change and fossil fuels, and we know the major studies and their conclusions as well as the reactions to those studies, ecological, political, and otherwise. We know the common recommendations for energy reform and the price of their implementation. More than just having a lot of data, we have “rich data,” and even some of the subtlest nuances are things we’ve seen before.

Sign #2: We’re Noticing Patterns (i.e., Themes)

At this stage, we know enough to ignore the true outliers (e.g., crackpot theories, useless experiments) and see the way things work in the field. A common pattern in medical research, for example, is that various theories and assumptions will dominate the literature for a time, get debunked, and then reappear. So, we hear that coffee is good for us, coffee is bad for us, coffee makes no difference at all, coffee is good for us.

This is where data saturation transforms into the beginnings of our argument. If we can judge the data with some objectivity, we’re going to see connections, motivations, aspects of the human condition, financial trends, or whatever else we’re looking for. Moreover, when we read yet another study or interview another group, we find that these patterns continue.

Sign #3: We’re (Accurately) Guessing What People Are Going to Say

Thus, our understanding of the data becomes predictive. It’s essential at this point not to stop reading or listening with care, however. If we let ourselves be blinded by a hasty generalization or pet theory, we run the risk of not absorbing the data.

This is how conspiracy theories get popular. People believe they’ve unlocked some great mystery and so only want to interpret the data that specific way. I’ve always liked the observation by Abraham Maslow, “When the only tool you have is a hammer, you tend to see every problem as a nail.”

But if we keep our minds open to the information as we receive it, we can avoid this pitfall and be certain that when someone says or writes something we’ve seen time and again, we’ve uncovered a true pattern, not just a convenient “nail.”

This is also where a good editor can help by pointing out unintended biases or unexplored aspects of an argument. Try a free sample of that sort of help at ProofreadingPal.

Sign #4: We See How It All Fits Together

The ultimate goal of research, one can argue, is to see how something works. It’s not just seeing the “Big Picture,” but also knowing how the little things work (or don’t work) together. At this stage, we know we’ve reached data saturation because what we’re reading is both familiar and makes sense.

Again, we may fall into the trap of thinking we have enough data for our argument when we actually need to keep digging. A good failsafe is to make sure we’re getting our data from varied sources but still aren’t finding anything new. This is an obvious example of the value of diversity in research.

A good researcher never really stops researching, but reaching data saturation means we can pause and make our observations, adding our voice to the chorus of authors trying to figure things out.

Julia H.