I’ve coached many managers through large-scale improvement programs in the last couple years. Regardless whether it’s an upgrade in a telephony system, a reorganization, or changing the way we reward and recognize our staff, it seems we have plenty of numbers along the way to measure how well we’re doing at change. How many action items are in red? What’s are the milestone achievements? How much will we save in changing the way we work? How far will our customer satisfaction measure move once the change is complete?
But there seems to be an underlying inferno of missing information that we’re simply not capturing. We can’t touch it or measure it using our most powerful excel or mini-tab graphs (Note: This article is not tackling hypothesis testing!). It’s more difficult to capture and harder to analyze. I’m talking about qualitative data.
Whilst there are many types of data, qualitative – and it’s opposite quantitative – can be the easiest to distinguish, but the hardest to display and analyze. Simply put, quantitative data details quantities, numbers and categories whereas qualitative lists somewhat intangible qualities, descriptions and anything that isn’t instantly measurable. However that doesn’t mean we should ignore it as a data set. Let’s pretend for a moment we’re conducting a project on improving how our own personal work-spaces are set up. We might collect the following data.
- On a scale of 1 – 10, how positively would you rate your work space? (For me, about a 6*)
- How many items are on your desk right now? (36).
- How many categories of things are there? (17 pens, 2 speakers, 14 reference books, and 3 cans of nondescript, store brand diet soda. I have 4 categories of items)
- How many types of soda do you have? (2 different store-brand types; one that claims to of received an education in medicine, and one that claims to have a passing resemblance to the moisture on a grassy mountain).
This quantitative data makes my work-space sound quite pleasant! I can count, categorize, and most importantly create charts, graphs and trends for all the above for how my work space is set up.
That data gives me some important information. But might not give me the full picture of what’s going on. Now consider:
- How would you describe the conditions of your work space? (messy, a little claustrophobic, and disorganized*)
- How do you feel in your work-space? (Over-whelmed and cluttered*)
- How delicious is your nondescript, store brought soda? (Not very. But it’s fizzy and caffeinated)
*=Ok, none of those are true of my actual work space. But this qualitative data that describes specific qualities of my space that you simply couldn’t measure in numbers for improvement or summarize trends in their raw format, yet it’s still important in describing what my work space looks like if we want to make an improvement. It’s still data, even though it’s not yet in number form.
So how do you deal with Qualitative data?
Whilst you can’t create themes from a set of raw qualitative data, you can turn it into quantitative data. Let’s imagine that we’ve started some problem solving, and decided that the quality of my soda is has a huge impact on how I feel about my space. (Spoiler alert: It has way more impact that it should have). We have to understand more about this soda to get a better picture of how to improve my work space.
While you can’t calculate the mean or the median for the descriptors of my nondescript, store brought diet soda, I can work out a mode. To do this, I use my favorite qualitative analysis tool, the humble word cloud. A word cloud will generate a visual image, with more frequently used words being bigger and more prominent than those used occasionally. This has a really big impact for describing an environment. Recently and organization I worked in was going through a big change.
Their projects milestones were being met, all their actions were green, but they couldn’t get the improvement to ‘land’, no one was using the new way of working. We collected some quantitative data about how they felt about the change, and sure enough most reported to be feeling ‘unheard’ and ‘uninformed’. Needless to say, with this new data, we changed our approach.
You can also turn qualitative into quantitative by categorizing the words in your descriptors. For example, using the above descriptive text, I was able to quantify if each of the words were positive, neutral or negative, to generate a general overview as to how I feel about my soda.
Since qualitative data helps identify trends and point in the right direction, I can also look at comparing variables of my qualitative data. I have a hypothesis that the temperature of my soda has an impact on its taste. By comparing how many times the temperature (warm, room temperature, cold) is mentioned in the same comment as how good it tastes (tasteless, yummy, delicious), I can conclude that the colder the drink, the more delicious it is, and potentially therefore, the more positive I feel about my work space.
This is an incredibly simple example, granted. But only through gathering qualitative data could we begin to conclude that in order to increase the positive rating of my work-space, I need to buy myself a miniature refrigerator… we might not have achieved that with the quantitative data alone.
I’d love to know your thoughts! What are your favorite ways of displaying, analyzing, and talking about qualitative data? What are some pitfalls you should definitely try to avoid?