Every time you analyze data, you start with a research question and end with communicating an answer. But in between those start and end points are twelve other steps. I call this the Data Analysis Pathway. It’s a framework I put together years ago, inspired by a client who kept getting stuck in Weed #1. But I’ve honed it over the years of assisting thousands of researchers with their analysis.
Following these steps, in order, will help keep your data analysis on track and get you to the right place. And you’ll enjoy the journey a lot more knowing you’re headed in the right direction without getting sidetracked.
But getting sidetracked, off the path and into the weeds, is really easy to do. So what are these easy-to-get-lost-in weeds?
Weed #1: Getting distracted by all the analyses you could do
It’s exciting to approach a large, rich data set. One with many variables, one that holds a ton of information about the population you’re interested in.
And with just a little bit of software training, you could produce endless statistics, graphs, and tables about all these variables. It’s fun. It’s interesting.
It’s a weed.
It’s not going to get you any closer to your destination: an answer to your research question.
If you find yourself getting stuck here, I recommend the following to keep the data analysis on track. Write down your research question. Put it in some very visible spot that you can see every time you work on your data analysis. Maybe tape it to the wall over your computer. Use it to remind yourself what it is you’re here to do.
Weed #2: Doing the analysis you’re comfortable with
This is a weed we all get stuck in at times.
It seems to make sense to get the data to fit the tool you know how to use, rather than struggle with one you don’t. But sadly, it doesn’t.
You may have to make untenable assumptions in order to make that comfortable analysis fit. But even if you don’t, you just may not find the real effects in the data because you aren’t using the best analysis.
This one is harder to avoid because you don’t know what you don’t know. There may be a perfect analysis out there that you just haven’t heard of. This is part of the reason you want to put together your data analysis plan as an early step. Run it past someone knowledgeable and get feedback before you collect data.
Weed #3: Skipping steps then having to redo everything
Have you ever painted a room? You often hear the advice that most of the work, and what makes the paint job beautiful, is in the preparation. Cleaning the surface, filling holes, sanding everything smooth. All before you ever pick up a paint roller.
But these steps are vital to getting a good result. Skip them and you’ll just have to redo them later. And you’ll have wasted time in the process.
Weed #4: (Avoidable) surprises and complications
There are so many complexities in data analysis. Many of these can’t be avoided: abstract concepts, software limitations, and data issues.
But some can.
I’ve worked with many researchers over the years who were forced into a complex type of statistical model because of a design choice they unknowingly made early on. For example, they counter-balanced some factor in their design instead of randomizing. Or they chose to measure a variable on an ordinal scale instead of a continuous one.
There are also foreseeable data issues to consider. How likely is missing data? Is your dependent variable one that’s often skewed, like reaction times or household income? Consider these likely scenarios early in the process by planning the analysis early.
Avoiding likely problems is much easier than fixing them later on.