Lyn’s research hacks: Data analysis
We seem to be bombarded these days with media advising us about various 'life hacks'. A Google search provides the definition of a life hack as "a strategy or technique adopted in order to manage one's time and daily activities in a more efficient way". I’m always keen to work more efficiently on my research, so I thought I’d coin the phrase 'research hack', which I define as “a strategy or technique adopted in order to manage one's research in a more efficient way”. This is the first in a series of blog posts that provides hacks for various aspects of the research process.
Hack #1: Don’t forget the ‘garbage in, garbage out’ rule
The ‘garbage in, garbage out’ rule certainly applies to data analysis. It doesn’t matter how good your analysis skills are or how sophisticated the software is that you’re using – if the data hasn’t been carefully checked and prepared then the analysis you perform on it will likely be rubbish. Always take the time to ensure that your data is as clean and accurate as possible.
In terms of what this looks like in practice, for a quantitative data set, make sure you’ve cleaned and screened the data, checked for any data entry errors and assessed the normality of the data (if that’s important for the particular statistical tests you’ve chosen). If you start by running significance tests and then discover that there’s some missing data, or that some values have been incorrectly entered, then you’ll need to do that work all over again.
While you might not be dealing with numerical data for qualitative research, your data still needs to be clean. Make sure any field notes are understandable and check that your interview and focus group transcripts are accurate. Checking the accuracy of transcripts is particularly important if you outsourced your transcription.
Hack #2: Keep a log of your analysis – record everything!
I have a theory that age has made me a better researcher – not because I’m more experienced, but because my memory is getting worse so I always make a point of writing things down! The rule of thumb for any research project is “don’t assume you’ll remember because you won’t!” Record everything – particular tasks you’ve completed, decisions you’ve made (and why), next steps, challenges you’ve encountered etc. I think this is particularly important for qualitative research where a lot of potentially subjective decisions might be made, but it’s important on quantitative projects also. It doesn’t matter where you keep the records as long as you have a system – a memo in an NVivo project or a note in Evernote or research journal are all fine, but you need to pick a system that works for you. Keeping records like this may seem tedious, but when you’re at the writing up stage you’ll thank yourself for keeping track of things and it’s never failed to pay off for me personally.
Hack #3: Don’t jump straight into your quantitative analysis
A classic mistake in quantitative data analysis is jumping straight into statistical testing. Researchers tend to be obsessed with p-values – we’re super keen to know whether or not the results are significant. Jumping straight in can be to your disadvantage as you can miss patterns in the data that were unexpected, and you might not fully understand the statistical tests you’re conducting because you’re not familiar with the story that the data is telling. I always start my statistical analysis with some basic descriptive statistics, not just to assess normality and check for possible errors, but also to help inform any analysis I do later down the track. It might feel like you’re delaying getting started, but it usually pays off ten-fold.
Hack #4: Start qualitative analysis as soon as possible
If you’re a qualitative researcher, ideally you want to start analysing your data right away. You would never do this for a quantitative project as there’s not much point calculating means and standard deviations until all the data is collected, but for qualitative research you can actually start analysis the minute you get your first piece of data – and it may well be in your best interests to do so. Starting qualitative analysis early can inform any later data collection you conduct, and the sooner you start your analysis the sooner insights can occur. While it may seem slightly counter-intuitive, it’s a good habit to develop.
Hack #5: If you’re using research software, discover what tasks can be automated
One of the great things about using software is that it can speed up repetitive or administrative tasks and save time. By way of example, NVivo has an auto coding function that enables you to automatically group together everything a particular person said, or every response for a specific question. I know from my own experience that this really speeds up analysis. You could also use NVivo’s text search query to look for text that you may have forgotten to code. Likewise, SPSS has something called syntax which is its inbuilt programming language. If you need to run a particular set of analyses regularly, you can save the syntax and each time you need to run them, you just need to click a button and they will be completed all in one go. Syntax is also ideal for any analysis that is tedious or time consuming when using the drop-down menus. These are just a few examples of automation in data analysis software – it’s definitely worth finding out what options are available for the software you’re using.
If you’re keen to learn more tips for saving time when analysing data, our annual virtual event Research Accelerator 2021 is coming up on December 1st. We'd love to see you there!