Analysis is confusing
I’ve often found myself stumped when confronted with a bunch of data, whether quantitative or qualitative. I might know what I was hoping to find out (but not always!), but that doesn’t always translate easily into knowing what I need to do with the information I’ve collected.
This seems like a fairly common problem, and yet I’ve struggled with it for as long as I’ve done research. So, at least officially, since sometime in 2008 or so, when I took my first research methods class before applying to psychology graduate schools.
Quantitative data at least seems like it should be more tractable than qualitative data. The first one’s a bunch of numbers, right? That seems like it should be easier to understand than a bunch of fuzzy words and descriptions and ideas. That being so, I’ve taken numerous psychological statistics classes. Indeed, I my mom has stories of me helping adult students in my mother’s stats class when I was 8. Math makes sense to me. Figuring out what to _do_, what’s relevant, what’s useful? Much less clear. This is often true, I find: knowing how to use the information I have can be a daunting task.
One of the things I find fascinating about UX is that this is known to be confusing and hard and also a really important aspect of what we _do_. Rather than trying to look for statistical significance, however, we’re looking for ideas and guidance and places that are obviously painful and places that are working well. Statistical significance is somewhat… irrelevant to the questions we are trying to answer. Not ‘how much’ or ‘how fast’, but ‘what is happening’ and ‘why’.
Statistics always felt like it was supposed to be a thing that could be done on one’s own. Like I should just know what the best approach is. This is likely not helped by the fact that I have trouble verbalizing math; it’s not at all the same language in my head, and translating is difficult. Having trouble verbalizing math makes it difficult to discuss it, and to consult with others to figure out the right sorts of statistical methods to use beyond the basic stuff that pretty much always has to happen. It’s not even just about which methods to use, but how to correctly interpret things. Statistics is a lot more fuzzy than practitioners like to admit to, at least in the psychological sciences. It’s all ‘is this significant _enough_’?, ‘is there enough of an effect for this to actually matter?’, ‘Have I gone down a rabbit hole and wasted all this effort?’. This is _not_ helped by the fact that non-significant results rarely get published.
I’ve known of affinity mapping, and even tried to use sticky notes to figure out some of my data in the first UX project I did. Unfortunately, as I found out at the time, analysis of the data I get in UX research doesn’t really lend itself to being done alone. Much like statistics, I suspect. I’m not at all sure how UX consultants do their analyses, given this!
Thankfully, I now have a mentor and an internship! When I flailed at Mo earlier today during our meeting, she suggested that I obtain Gamestorming as a useful reference book, and that we should go ahead and do some affinity mapping on my data. I need a bit more data first, but this means I finally get some of the guidance I’ve desperately been looking for.
I’ve been reading Gamestorming today, taking frequent breaks so I have time to let things settle in before I continue reading. I’ve also been reading a Paper Prototyping book that I got at the suggestion of another helpful person in the Boston area UX community, Jen McGinn. Given that I sort of guessed at paper prototyping for the same project in which I tried to analyse my data using sticky notes, this book should be helpful.
I’m really looking forward to getting a chance to do affinity mapping on this project. I think it’ll make a huge difference for my confidence!
Previous, outreachy starts.
Next, initial interview analysis.