Once the research results have been carefully gathered and recorded using some filing system, the data needs to be able to be explained and interpreted in order to attempt to answer the initial research question. The type of analysis and the means of presentation need to be considered at an early stage – even before the results themselves have been obtained – in order to be certain that the research method(s) chosen can actually be used to generate sensible results. This chapter in the book outlines some key steps to ensure that your data analysis is appropriate for your needs. The companion website shares the experiences of some active researchers in how they approach data analysis in their own research areas.

Ok for me, the main purpose of the analysis process is to take that raw datawe’ve collected about that subject we’re interested in, and turn it into information we can work with. It’s the process by which we attribute meaning to the results which we’ve collected and that helps inform our findings and the conclusions we might reach in our projects. I think there’s a number of key decisions that researchers have to make in terms of the analysis process. One of the first things that they have to ask themselves, after they’ve collected the data, is “is that data complete?”, do they have enough data to answer, or that relates to all of the research questions. If it isn’t complete, they may need to identify some gaps and consider some further data collection. Also, as a researchers project develops, they will often find some of the data they’ve collected becomes surplus to requirements, as the focus might slightly change or be refined, so there’s some key decisions about what to leave out of the analysis process. Then, I think researchers need to revisit the plans around their data analysis methods and make sure they are fit for purpose, and a key consideration there is whether the data analysis methods they proposed to use align well with the types of data they’d collected, and what they want to be able to show. So for example, if someone’s collected a lot of statistical data, they need to make a decision about whether they use descriptive statistics, that show general patterns and demographics, or whether they need to go further and use inferential statistics that might show the relationship between different variables or factors. If they’ve been undertaking qualitative data collection, then they need to be sure that they’ve got a robust method for analysing that qualitative data. One of the big criticisms of qualitative research, when we do interviews and focus groups is that the researchers sometimes look as if they are cherry-picking the quotations or the descriptions that have been provided to them to fit the argument they want to make. So you need to have a method of qualitative data analysis that lets the data speak for itself. Now that’s probably going to involve some kind of thematic analysis that brings to the fore a range of views and opinions. It may involve some type of analysis that looks at individuals’ experience, but whatever data collection, whatever data analysis you want to undertake, you have to be sure that you’ve got a robust method that lets the data speak for itself and allows you to find your conclusions. In terms of the relationship between research questions, our data and conclusions, I think the key issue is one of alignment. We need to be able to show that the research questions that we’ve asked, we’ve collected relevant data against, and that the data we’ve analysed is being used to make informed conclusions that relate to those original research questions. Now, there’s two, I think, two key considerations here in terms of thinking about the relationship between data analysis and conclusions. The conclusions we arrive at have to be able to be seen, and should be based directly upon the data analysis we’ve undertaken. Conclusions in a research project are not the place to bring in new information or to start talking about issues we haven’t actually investigated. I think another key consideration in terms of the relationship between our data, our analysis and our conclusions is to be absolutely upfront about the gaps that might still remain, when we’ve got to the end of the research process. These might be areas where we’ve partly answered some of the research questions which is absolutely fine, or it may be areas that have emerged through the data analysis that we didn’t set out to investigate originally, and we might identify those as areas for further research in the future.