Many researchers code in NVivo. It is where the transcripts live, where the coding happens line by line and where months of careful work are kept. Once the coding is done, many want to see the shape of that work: the themes laid out as a map they can read, refine and present.
Themescape works alongside NVivo for that part of the analysis. It replaces none of your coding software. Import from your coding software reads a project exported from NVivo and writes the Themescape template files from it, so you can see your themes as a map and carry on refining them. This article walks through the move, from preparing the project to tidying the files afterwards.
Before you export: use pseudonyms
The import brings participants across from NVivo cases. Whatever names your cases have in NVivo are the names that will appear in the participant column of the template, in the reader and in anything you present from it.
So the first step happens in NVivo. Rename participants and cases to the pseudonyms you will report before you export. It is much easier to do this once, at the source, than to correct every row afterwards, and it means identifying names never reach your template files at all.
Exporting from NVivo
NVivo exports whole projects in the REFI-QDA format, an open exchange format for qualitative projects.
- On a Mac, choose File, then Export Project.
- On Windows, choose Share, then Export Project, then REFI-QDA Project.
Save the exported file somewhere you would keep the rest of your study's data. It contains your coded text, so it deserves the same care as your transcripts.
Importing into Themescape
On the viewer's Start a project screen, choose Import from your coding software, then the NVivo tab, and select the exported file. The viewer reads it in your browser on your own computer, writes the template files from it and shows a short import report. Nothing is sent to Themescape, and your NVivo project is not changed.
From there you can open the viewer straight away, or download the CSV files and carry on in the template. Most people do both: look at the map first, then keep the files.
What comes across, and what stays behind
The import maps NVivo's structure onto the template's.
- Top level codes become themes, and the codes beneath them become codes and sub-codes.
- Deeper levels are folded into the sub-code, since the template has three levels beneath the project.
- Coded text from text documents comes across as quotes.
- Participants come from NVivo cases.
- Code descriptions become working definitions in the Framework file.
Some things stay behind. Coding on PDFs, images, audio and video is counted in the import report and left out. Connections between codes do not come across, because NVivo does not export them. You can add them afterwards in Work mode, choosing Empirical or Theoretical for each and noting why.
Tidying the template files
An NVivo project is organised for coding, and the template is organised for themes, so a little tidying is usual.
Check the top level. Because top level codes become themes, any top level folder you used for organising, such as a parent code for each interview question, will arrive as a theme. If that is not what you want, restructure it in the Data file by changing the theme names on those rows.
Check names. Themes, codes and sub-codes are matched as plain text, so two spellings of one name become two items. Look through the file tree for near duplicates.
Fill in what NVivo did not have. The template has columns NVivo does not export, such as refined_quote for presentation wording, illustrativeness for rating how well each quote shows its code, and preferred for your favourites. None are required, and you can fill them in as you go.
Read the definitions. Code descriptions written in NVivo become working definitions. They are a good starting point, and worth rereading now that you can see each one beside its quotes.
Keeping NVivo and the template in step
Once the files exist, decide where the coding lives from now on. There are two sensible choices.
You can carry on in the template, treating the import as a one off move and keeping NVivo as the archive of the coding so far. Or you can keep coding in NVivo and import again when the coding changes. A fresh import writes fresh files from the NVivo project, so definitions, favourites and connections you refined in Themescape will not be in them; keep your saved progress files and copy those refinements across.
Either works. What causes trouble is coding in both places at once, because the two will drift apart.
Other software
MAXQDA and ATLAS.ti projects come in the same way: export the project as a REFI-QDA file ending in .qdpx and choose that program's tab under Import from your coding software. If you use other software that can export coded segments to a spreadsheet, you can also arrange them in the template's columns by hand: one row per coded quote, with its participant, the quote, its theme and usually a code.
Saving your work
Themescape keeps no copy of your project, so save your progress before you close the tab. A save downloads the template files and, if you have arranged anything, a layout file, all named after your project with the date and time. Keep them with the rest of your study's data, and load them next time to carry on where you left off.
Charlotte Price is the founder of Themescape.
