Data nerd? Stats nerd? Or maybe just a fascinated observer of my journey from longtime professional storyteller to statistical storyteller and scoliosis researcher?

This is the story of the early chapters of my emerging studies about the impact of exercise on chronic scoliosis pain and mobility issues in adults. And if you ARE a fellow data nerd - I have some takeaways for you to add a layer of statistical magic to your data storytelling!

My first attempt at creating a survey instrument measuring the relationship between scoliosis pain and exercise, yielded mixed results - reliable yes. Consistent? Meh….

But the best news? I loved every stinkin’ minute of it - from literature review (yes, I’m a nerd who loves lit reviews!), through construct and item creation, scale distribution, data and statistical analysis, and project write-up.

It was the summer and fall of 2024 and I was uncovering statistical stories in a dataset from a survey of my own creation. The survey consisted of 18 questions, a potpourri of item types: multiple choice, Likert, dichotomous, and what turned out to be the most effective types for my data, visual analog items. The survey was aimed at adults with chronic scoliosis pain who exercise, even occasionally, and accordingly, I’d later name it the “SE-18.”

Back to all the stats fun!

Sure, the entire scale could not be tested in one fell swoop (R, SPSS) for reliability and exploratory factor analysis (EFA) because apparently I’m a far better book writer than I am a scale item writer - so far. Ultimately though, for the scale subsets that were run, reliability scores were good, and the EFA revealed some potentially useful common denominators in adult scoliosis pain: mobility, neurological, and stress effects.

3 Research Project Takeaways

In analyzing every last correlation coefficient and creating colorful data visualizations from all the beautiful survey data dots, 3 stats analysis lessons emerged. No matter what your specific data storytelling role (if that’s your jam) here are 3 takeaways from my imperfect but impactful scoliosis/exercise survey instrument pilot study.

#1: Am I Measuring What I Think I’m Measuring?

Simply put, this is a core question in the creation of a survey construct. At the beginning you decide what it is you are measuring, and how it may or may not show up in the data. Then, when you’re analyzing your results, you keep asking.

Based on the running of the stats, my scoliosis/exercise scale was trying to measure multiple things in one scale - the nature of scoliosis pain, its mobility issues, its impact on daily life, way too many angles of the subject of exercise, and even some toe dipping into mental health effects. My scale had a major multi-tasking problem! (I already have a plan in place to reign it in for rev 2.0 and beyond of my scale.)

Takeaway Tactic: Is the data you’re analyzing not behaving the way you’d expect it to, even though you’ve cleaned and recleaned it, and run your other normal checks? Are you sure it’s measuring what it says it’s measuring? Or do you have a case of multi-tasking data, trying to be all things to all people, all in one dataset?

#2: The Real World: Data Edition

This is where our paths diverge a bit, since as the researcher I got to decide the who, how, and what of data collection for my pilot study. In brief, for rev 2.0 of my scale, I’ll be casting a wider net over a more varied population for a longer period of time (and followed up with qualitative focus groups).

But as a data analyst/storyteller, data collection probably looks quite the opposite of that autonomy. So what tactics can statistical storytelling and a scoliosis pilot study offer to those tasked with analyzing OPD (other people’s data)?

Takeaway Tactic: When you inherit a dataset, ask yourself: who was included, who was excluded, and what questions were asked (or not asked)? Pretend you’re a casting director for a show called "The Real World: Data Edition" - were all the right characters even in the room? And if not, how might that affect the story your data is telling? Being aware of the origin story of your data gives you an edge in how you analyze, interpret, and ultimately communicate it.

#3: Theory, Theory, Theory (the part of my scale pilot study that I owned!)

In my master’s program, I had one professor in particular who reminded us over and over, to never trust the stats alone, especially if the theory doesn’t sync up. Fortunately for me, and as mentioned earlier, I love literature reviews! So this repeated reminder by Dr. Bolin to know your subject matter inside and out was music to my ears.

In light of my various statistical analyses learning curves, if you don’t mind, I’m going to take the W on this one. My literature reviews, plural, the foundation of my survey instrument, were possibly its saving feature. The deep understanding of my topic of interest helped me sort out more than a few “huh??” moments.

Takeaway Tactic: Even if you’re at the mercy of someone else’s dataset, you have the option of learning more about the theories driving the data. I’m not suggesting a full-on lit review… unless you’re really into that sort of thing of course. But generally speaking, understanding your story-driven stats can add valuable context during data analysis. Sometimes being a better quantitative researcher means becoming a qualitative one in the process!

Overall, where do you have control in your data analysis to potentially help guide your stakeholders, your clients in collecting better and more useful data? How can you go beyond data analyst to a consultant role (over delivering is one of the hallmarks of my 20 plus year career serving clients!), applying critical thinking, creative pathways, understanding the human behavior driving the data dots, and other “AI divergent” skills? As a human data analyst you are not at the mercy of your data.

Better Theoretical Foundation > Better Data Collection Procedures > Better Data > Better Analysis > Better Interpretations > Better Stories > Better Data Visualizations > Better Client Outcomes

PS: Learn more about my scoliosis/exercise survey instrument and ongoing research project!