UPDATE ON COLABRATION #
For multi-user online/cloud-based cooperation on PDF coding, see those tools.
- dedoose (17 USD/user/month， 10USD for stu)
- FlowUS 息流 (15CNY, 9 for stu)
QDA can be used for literature review, but not bibtex management. For bibtex/latex management, see Reference and Knowledge Management (Zotero, Qiqqa).
I wrote this blog in a short time, please leave a comment if you find any mistakes.
As a PhD student, I found QDA method and software are very useful. However, I need to choose software carefully due to the lock-in effects, which is not a problem for Zotero-like reference management software. It is almost not possible to migrate between different QDA software. (I saw a migrator from Atlas to NVivo, but not sure.)
When 3 QDA software applications are compared, MAXQDA will be tested later. This is due to its user-unfriendly installation and it is digtis&stats-oriented idea without good visualization. Its price is 90~200 EUR for 2 years. Actualy, MAXQDA is my favourite, but they are not good at advertising it, especially the great visualization part and matrix.
When I choose between NVivo and Atlas.ti, it is a tradeoff. Atlas is 75 EUR for 2 years while NVivo is 170 EUR. They look similar, work similar. Thus, I spend some time to compare them.
Qiqqa is not actually a piece of QDA software, but it gives a lot of support for daily research such as auto-summary (“annotation report”), topic modelling (expedition & themes). Thus, it is used as a benchmark method in this blog.
The highest/best grdae is 5. “?” is not checked yet. This is a very subjective comparison, please try them by yourself.
NVivo has 2 week full-feature trial time. Atlas has unlimited trial time with limitations when saving projects. Qiqqa is free for one computer with some minor limitations.
I really need a hierarchical organization of papers, like in Zotero. Atlas provides “code tree”, but does NOT support duplicate code names on different tree branch. Besides, I don’t know why, the content in the main window can NOT refresh automatically after I modified somewhere else, which is ridiculous.
Qiqqa provides a very good in-doc search, multi-highlighting, click-to-sync view. Atlas provides a similar one which is also very good. I have not find a convenient way to do a “ctrl + F”-like search & multi-highlighting & sync while viewing a document. Qiqqa’s multi-highlighting is shown below (see also ref management):
I have not find a way to automatically generate SLR matrix in Qiqqa. I appreciate it if someone can leave a comment for steps in Qiqqa, if possible.
MAXQDA have a menu option to search selected text. However, there is no shortcut. What is worse, Ctrl+C has some compatibility issues. Possiblly with win10+libreOffice+hibernate. See their link)
The results are displayed well and we can click to sync view. It would be much better to highlight string in results (instead of capitalize) like Qiqqa or Atlas.
Nvivo sucks at doing this kind of quick query.
Though Nvivo is not giving many features with grade 5, it is giving a good overall performance compared with Atlas and Qiqqa, in my opinion. OCR is actually mandatory, but we can use other pdf software to handle it as pre-treatment before putting into QDA.
Considering backup, Atlas puts many files in \AppData\Roaming\ folder (similar like Qiqqa), while MAXQDA & NVivo has one single file for a project. I prefer the later one.
TIP BEFORE USING #
- Import data using pdf files with names “auth yyyy title …”, do NOT import from Zotero ris or Endnote etc., as it is complicated to search file attributes (such as authors, years) in NVivo. Simple search is done using filename.
- We can use sets to have multi shortcuts one file. Folders/groups of sources/docs contain different sources like in our operating system. We canNOT leave one item in both folders while we can in Zotero. The copied files in different folders are considered as different items.
- When doing coding, remember to select all parental codes, which will ease later analysis.
- When creating codes, it is better to use some keywords from parental codes so that sub-codes can represent parents directly by name. (no need for MAXQDA)
MAXQDA TUTORIAL #
MAXQDA FOR SLR/SM #
For Systematic Literature Review (SLR) or Systematic Mapping (SM), we use “sets” as SLR categories (one paper can be in different sets wrt SLR ???).
method 1 overlapping codes (suggested) #
This method is also good to see methods/solutions vs. topics/problems. The generated graph is the same as method 2.
(Similar like NVivo)
method 2 using sets (not suggested) #
Coding: put the doc into SLR sets; code the title with topic.
Query & Result:
(pictures below are from official tutorial, so the codes have no “topic 2”)
!(https://cdn.staticaly.com/gh/ol-storage/pub@master/previous_image/code matrix 2.png)
!(https://cdn.staticaly.com/gh/ol-storage/pub@master/previous_image/code matrix 3.png)
NVIVO TUTORIAL #
Videos to watch (please watch in the numbered order):
NVIVO FOR SLR/SM #
There are 2 methods to create matrix.
method 1 - matrix coding (suggested) #
Coding: ensure overlap, e.g., by selecting two codes for the title.
Benefit: we can see the numbers directly and see all the papers & code details by double clicking the number.
- Ensure overlap. In the pdf file, t2 and ttt2 codes need to have overlap to make this work, for example, the title is selected and coded with both codes (t2 and ttt2).
- Do NOT overlap in other places other than title or cite key.
method 2 - using codes & cases (not suggested) #
Use “cases” as categories of papers, such as “method”, “tool” etc., cuz the cases can be used as rows, and nodes/codes (paper topics) can be used as columns in SLR or mapping matrix.
Benefit: auto summarize.
OBS: the same as using only codes, see above.