After that, you will start to minimize the number of your codes, to put them into categories and synthesise them into more inclusive and broader themes. In my case, this point occurred after 5 interviews. interviews), until the point when there is no need to create new codes, as the newly inputted data could be fully covered with the existing codes. Also, remember that you usually do line-by-line coding just for the first few sources (e.g. It is time consuming, but it’s worth it if you really want to apply this kind of coding in Nvivo 10. So what can be done about it? The following is a way around this limitation. You can clearly see in the above figure the problem with line-by-line coding in Nvivo 10 is. Figure 2 shows coding strips for an interview extract with each of its lines coded. The coding stripes in NVivo 10, however, do not correspond to the length of the coded text on the left hand side.
Although you can, of course, code any given line, word, or piece of text you want, the problem is that when you do line-by-line coding, you ideally want it to show next to the next you coded in a clear way, so that you can easily inspect the data (particularly because there may be so many codes in the text – I created 190 codes from my first 5 interviews in my study of English Language Identity!). Unfortunately, doing line-by-line coding in Nvivo 10 is not possible without having to “go around” its limitations. with codes being shown horizontally next to each line). If you are working on NVivo 10, or have Nvivo 11 but are not satisfied with how the codes are shown, read the following section on what else can be done to make line-by-line coding appear in a more “traditional” way (i.e. Once you created a big number of codes, constantly having to look at these numerous, vertically aligned and colourful, codes may be very exhausting. Therefore, as you can see in Figure 1 below, it is possible to easily distinguish between the codes, and attribute these codes to the text, even if each line is labelled with another code.Īlthough this is a major improvement from the previous version, which was much more “clumsy” in this respect, it is my personal opinion that it is still not as practical as it should be. The improvement that Nvivo 11 introduced is that the stripes clearly indicate the length of the coded text. In both versions of the software, after you code each line of the text, then go to VIEW > CODING STRIPES, the box on the right-hand side will show you the “stripes” indicating the codes (or, in Nvivo’s tems, nodes) that you used.
I will post a separate entry about line-by-line coding elsewhere, and for now I will assume that since you want to know how to apply this kind of coding in Nvivo, you have already made an informed decision as to why this approach suits you.Īt the time when I was involved in my study of Polish migrants’ English Language Identity, I was working with NVivo 10, which was much less adjusted to line-by-line coding than Nvivo 11 currently is. If you are investigating a “new”, or under-researched, phenomenon, such approach may enable you to effectively discover its various components, while also controlling your own assumptions about it, as the initial, detailed codes you are creating are, essentially, a summary of what was said in a given line of text. London: SAGE.), because it is particularly useful when we don’t want to impose a pre-existing framework onto our data, but rather to let new themes emerge from it, which constructivist grounded theory researchers mostly do. It is commonly associated with constructivist grounded theory (Charmaz, K. I find this technique extremely useful when I really want to “dig deep” into my data and extract as much as possible from it.
Line-by-line coding refers to applying codes to each line of qualitative data (e.g.
In this post, I explain how to do line-by-line coding in Nvivo 11 and how to overcome the limitations of NVivo 10 to apply this coding technique.