Data analysis mistakes to avoid

Lyn LaveryIt’s the little things that count in life, and this is very much the case with data analysis, where failing to pay attention to small details can result in quite large problems! Here are the top five data analysis mistakes that we see researchers make. These are all easily avoided, so do keep them in mind for your next project.

Not preparing data well: When it comes to data analysis, it’s very much a case of “garbage in, garbage out”. It doesn’t matter how good your data analysis skills are, or how flash the software is you’re using – if the data you’re analysing hasn’t been checked and cleaned, any analysis you conduct on it will be “rubbish”. For qualitative data, check the accuracy of all data carefully, particularly if you’ve outsourced your transcription. For quantitative data, screen and clean the data before beginning, checking for any possible data entry errors, outliers etc. While this step may feel like you’re wasting time, there’s nothing worse than getting partway through your analysis and discovering that the data you’re working with is full of errors, and that you need to start over.

Failing to keep a log of the analysis: Even if you’ve got the best memory in the world, always keep a clear record of your analysis tasks and decisions. This is especially relevant to qualitative projects where numerous (possibly) subjective decisions are made, but it’s also important for quantitative analyses. Keep a research log that records things like coding decisions (qual) or why certain variables have been combined (quant) – I can guarantee you won’t regret it!

Losing sight of the research question: Any data analysis you conduct needs to be very focused on the question(s) you’re trying to answer. While this probably seems like an obvious piece of advice to give, it’s amazing how many researchers lose sight of this. The data collection stage can be overwhelming, so when it comes time to start analysing, it’s not uncommon to have lost sight of the original question you were trying to answer (which of course may have been developed months or even years prior). The first thing you should always do before starting your data analysis is revisit your research question(s)—this will help enormously with your analysis, regardless of whether it’s qualitative or quantitative. In fact, we’d recommend keeping your research question prominently displayed on your wall or desk throughout the analysis process.

Failing to be consistent: Research is hard enough, so why make it more difficult for yourself by ending up with a muddle of differently named files, participant identifiers and other information. Consistency is important at all stages of the research process, but particularly so when it comes to analysis. Ensure file names, unique identifiers, pseudonyms, and variable labels are used consistently wherever they appear, and spend some time deciding on these up-front so that you’re not constantly changing them throughout.

Thinking that software is a “magic wand”: No software will do the analysis for you. As the analyst, you need to be making the decisions about what statistical tests you will run, or what approach you will take when creating and applying codes. Software will certainly assist with this, but as the researcher you will be “driving” the software, so you will need to be knowledgeable about the data analysis process. It’s also important to keep in mind that you don’t have to use software. On a smaller qualitative project, it may be completely appropriate (and manageable) to opt for good old-fashioned highlighter pens for example.

If you do decide to use software, then make sure you use it well. Software can really speed up your data analysis, but it can also hinder it if you don’t use it effectively. Academic Consulting’s training programme has a range of courses to help you become a more efficient data analyst, and avoid some of the mistakes outlined in this post. We have courses covering NVivo and SPSS, through to qualitative and quantitative analysis, suitable for all levels from beginner to expert. I hope to see some of you there!