New Apps on the Block: Atlas.ti 7 and NVivo 10

I want to be brief and to the point. I am doing a large project and so far I am:

(1) satisfied with DT’s macrostructural approach:

  • good document search, tagging, AI and other features
  • good linking feature to specific page in the document.

(2) very disappointed with DT’s microstructural approach:

  • cannot tag a specific quotation, cannot do much to assist micromanaging a long project.
  • Cannot provide TOC of documents, the list could go on and on.

In view of these limitations I urge users to complement—not abandon—DT with two new tools for research: Atlas.ti 7 and NVivo 10. These past months have seen the Mac version of these two professional Windows applications. Yes, they are expensive, but there’s a 30 day trial. And who knows, it might even be a match later on.

I am testing Atlas.ti as I write this post, and it looks promising. Let the conversation begin…

Just curious – Did something you read or hear lead you to expect DEVONthink would have these features? I don’t recall seeing any feature comparisons suggesting there were similarities or overlaps between DEVONthink and either Atlas.ti or NVivo.

Good approach. :wink:

Such a conversation should begin by recognizing that the comparison is not apples to apples, but apples to oranges.

The DEVONthink applications are general purpose document management/information management software that can be used for many purposes and in many ways. DEVONthink Pro and Pro Office are highly scriptable, and so many procedures can be automated or features extended by scripting. There’s a lively Scripting section on our user forum.

Quantitative Data Analysis (QDA) software is written to provide specific computational tools for quantitative data analysis of document collections.

DEVONthink isn’t designed to do QDA, and for my purposes I would find QDA software not appropriate for my document management/information management/information analysis needs. I don’t do qualitative data analysis, and would find the environment of such special purpose software very weak for my needs. But if I were ever to take on a QDA project (which for me is highly unlikely), I would look for software appropriate to that purpose.

Over the years, I’ve seen comments by some DEVONthink users who do QDA work using QDA software, but also use DEVONthink because it can be used for other purposes with their document collections. DEVONthink can share documents that are stored by different database software by Indexing them, so long as the documents are available via the Finder.

It’s generally good practice to use software tools that are most appropriate to one’s needs.

Please note that I can create a TOC for any group in my databases, and the History view of my Main database is a TOC of the more than 30,000 documents in that DEVONthink Pro Office database.

@korm: No, I did not read or hear any lead to expect that. But that does not dismiss the suggestion I make. And the reason I make the comparison is that I sense a gap in the plethora of academic apps we use. Not just a feeling; I am consciously searching for apps that can bridge the gap between DT and Scrivener, two vital tools of academic research I use. I would love for DT to implement more precise means to organize data at the paragraph level… hence the suggestion, but if that goes against the direction DT is going that’s fine. It was just a suggestion to complement workflow.

@Bill: thank you for your thoughts. There is room for disagreement here, since recently the two companies (Atlas/QRS International) have branched out and and are offering considerably improved versions of their product. For this reason they now claim the ability to conduct qualitative research. Under the main description of their website it is stated clearly that the purpose is for “qualitative research,” (

Hence I am comparing oranges with oranges. Atlas.ti 7 for mac, for example, offers several features that overlap with DT, and some that surpass it. This means DT scripting is replaced by several innovative, inherent features in Atlas. But the Mac version does not have the whole range of features which the Windows version does. Bottom line: because the two apps (DT and Atlas are overlapping in good measure this conversation is, in my view, justified.

Please elaborate a little more on how you enable the TOC for groups, and history view of main database. I appreciate the suggestion.

To create a TOC of a selected group or multiple selected groups, in the Three Panes view document contained in the selection(s) will be listed. Select all of the documents and Control-click. Choose the contextual menu option to create a table of contents.

The History view of a database (Tools > History) is a table of contents of the database.

I’m somewhat familiar with qualitative analysis approaches. QDA is often used as a tool to analyze behavior in focus groups used, e.g., to evaluate marketing potentials of a product. I’ve seen it used in development of a legal case through analysis of responses by a forum group to various presentations considered for use in a trial.

Beyond such uses, QDA has been used in a variety of fields including psychology, sociology, philosophy, ethnic studies, literary analysis &c. There are a number of methodologies, some of which, as I studied under Karl Popper, seem non-falsifiable to me. I’m a bit intrigued by the emphasis on the researcher’s philosophy in approaching data, which reminds me of the observational problem of determining the fate of Schrödinger’s cat. Perhaps some form of QDA may be useful in the strange universe of quantum physics. :slight_smile:

I don’t do QDA. There’s a variety of QDA software which hard-codes the researcher’s philosophy of his relation to the data.

If I were to use DEVONthink for QDA, perhaps the annotation script approach by korm and Frederiko could be adapted as such a tool. See Export an index of annotations to a numbers spreadsheet

Just to point out, that these scripts are just the start of what you can do once you start to annotate, cross reference and index the tag cloud and documents. The virtue (and for people starting out, the hinderance) of DT is that you have to design your own workflows, rather than being tied into the workflows of software like Nvivo. As an example I have started experimenting with exporting the DT tag cloud to a graph database ( as I look for other ways to find connections between documents and sources.


I’m an occasional, and rather amateur, user of QDA for market research purposes. I’ve looked at Atlas and Nvivo in the past because my friends in academia tell me they’re the gold standard, but then they don’t have to pay the huge (for me) corporate price!

I’ve managed with two products, one free, the other I had for other purposes:

  • TAMS Analyser is free QDA software. It takes you back to 1990s (or earlier) levels of user-unfriendliness, but if you can get your head round it, it offers some pretty powerful tools (e.g. audio transcription and tagging), and once you learn its approach and shortcuts etc. its actually quite fast.
  • Tinderbox can be used as a surprisingly powerful tool for QDA. A couple of people have blogged about this, e.g. … 969453/6#6)

Good luck!

Hold the horses - you’re exporting a DT tag-cloud? :open_mouth:

Where? What? How?

This is a bit off topic but here goes anyway. A lot of my annotations have natural nodal connection points - people, issues, places, times etc - so it seems there should be better ways to visualise and associate important information that easily gets lost in a sea of text. In practice though producing a useful timeline or knowledge graph proves to be a much harder task than it seems.

Timelines are one way of mapping tags and software like Aeontimeline with its abilities to create different arcs over the same timeline and link participants to events are a natural fit for DT. I am going to post some scripts shortly which do a richer job of exporting DT tags and metadata to Aeon than is currently possible.

Graph databases are another way to tackle the informational mapping problem. An interesting recent use of graph database techniques was to link Paul Revere to his revolutionary contemporaries ( Its easy to see how this kind of analysis could have come originated out of an analysis from documents in a DT database.

If one has been tagging carefully and consistently this kind graphical result seems an obvious and next step. (In fact I could see how DT artificial intelligence could build this web itself but that is the future hopefully.) There are some interesting takes on making this informational mapping useful in the legal field -

Right now I am a long way from having a reliable export from DT to Graphviz, hence the experimental tag.


So far this thread has exceeded the expectations of my initial post. I appreciate the useful tips and ideas from DT, and especially the promising intersection with qualitative analysis (CAQDAS) now in three mac versions: Atlas.ti, MaxQDA, and NVivo. Thanks to Fredriko for expanding my imagination and putting me in touch with the research of Florix Bex.

Wow, some fascinating information in this thread - I like the idea of a data graph (I’ve dropped the link into my sorter to look at later) - I may look at using it to simplify an operations manual that I am currently putting together - so many different pieces of information that need to be tied together to produce a usable document. I’d love to use DT, but unfortunately we use Windows at work - mind you I might still use my Macbook Pro :slight_smile:

Still learning here, making my way through the manual and keeping an eye on these forums (fora?)


Another entrant in the QDA toolkit – pricey, still.

You might also consider Dedoose, a less expensive QDA tool.

This is a revealing discussion. I am new to DT. Prior to buying DT I had also been contemplating Atlas.ti. The new update for Mac makes Atlas.ti look very appealing.

At no point did I ever assume that DT was intended for Qualitative Data Analysis. Atlas.ti and DT are different tools. There is some overlap (inasmuch as there is overlap between Devonthink and Finder), but they are not the same, nor are the supposed to be. I ended up going for Devonthink anyway because it became apparent that DT was going to serve me in a large variety of ways indefinitely. It will help me with all of my research projects, it will help me with all my administrative obligations, and it will help me with my personal affairs. Atlas.ti might serve me for the duration of a particular project where QDA is needed, and only for the duration of that project. It is not clear yet whether the current project requires QDA, that will depend on the amount and type of data that end up being generated. There’s also a lot of disagreement in the methods literature about whether QDA can actually meaningfully improve a qualitative study. I’ve yet to spend enough time looking at this methods literature to make the decision for myself. I think DEVONThink could facilitate manual coding and analysis a great deal.

So as I see it, Atlas.ti looks very appealing if I decide, based on the literature and my intuition, that it would improve the quality of my upcoming study, and if the amount of data generated justifies it.

DEVONThink, on the other hand, will be an indispensable tool regardless of the type and quantity of data generated, as well as a fantastic tool for a number of other aspects of my work and personal life.

I am very new to DEVONThink and I look forward to pushing its limits. Atlas.ti also looks very appealing and I look forward to trying that out in conjunction with DEVONthink, should QDA be justified.

Korm, thanks for pointing out the newcomer Quirkos. More user friendly than NVivo and Atlas.ti. A simplified version, let’s see how quickly it can develop and improve. The price of 65 for students is not bad.

Scottlougheed, good explanation of the complementary differences between DT and QDA.