The purpose of this article is to present an alternative way to use comments and memos in ATLAS.ti, as an example of using the program creatively. Every research project has its own needs according to its methodology and objectives, and every researcher has their own preferences, and so it is helpful to know that the program offers great flexibility in how it can be harnessed.
A major part of qualitative data analysis is writing – composing analytic memos, jotting down unexpected insights and ideas, providing formal definitions of codes, and so on. When using ATLAS.ti the question arises: where is the best place to capture all this writing? One way is to seek out a feature of the program that most closely corresponds to the kind of writing. So an analytic memo would be written as an ATLAS.ti-memo. An unexpected insight that does not yet relate to anything in particular might be written in the HU-comment, a general purpose place that refers to the project as whole. And a definition of a code would be written as an ATLAS.ti-code-comment. This is a reasonable approach, but is not always the most powerful way to use the program. Here is one example where I almost always take a different approach: when representing and writing about larger themes or concepts.
Different methodologies use different terms for the bigger-picture concepts in qualitative research, such as categories (Glaser & Strauss, 1967) or theoretical codes (Charmaz, 2006). A common term is theme, even though the word has very different meanings in different methodologies (e.g. cf. Chang, 2008; Guest, MacQueen, & Namey, 2012; Thorne, 2008). For our purposes I will use theme for higher-level concepts in general.
Perhaps because writing about themes is often called memo writing, or writing an analytic memo, it is natural to go to ATLAS.ti’s memo feature as presumably the right place to do this writing. You can create a new ATLAS.ti-memo, give it a name, and begin writing. I want to make the case that a better place to write about the emerging themes in a data analysis is the comment areas of special-purpose ATLAS.ti-codes, in conjunction with using ATLAS.ti-networks. In other words, the choice of what the program’s features are used for should not just be based on what the program calls them.
Codes and analytic memos in qualitative research are obviously completely different things with different purposes. (When I write “code” I am referring to a named concept you create in the research study – nothing to do with ATLAS.ti. When I write “ATLAS.ti-code” I am referring to a thing in the program. The same goes for memos). In contrast to codes and memos in the research project, ATLAS.ti-codes and ATLAS.ti-memos in the program have a great deal in common, with only a few things that are different. It might be better if the names of these two software features were different from the names of things in the research process, because sometimes coding is best done without the use of ATLAS.ti-codes (see my blog post “Using quotation names for coding: An illustration from grounded theory”); and as I am about to describe, analytic memos are often best written using special-purpose ATLAS.ti-codes. Perhaps ATLAS.ti-memos could be called Writings in the program, indicating they can be used for writing of any kind – for example, I use them for keeping a project journal, writing notes on conversations with co-researchers, pasting in reference materials of one kind or another, communication among team members, etc. And ATLAS.ti-codes could be given a more general name that covers all their uses, including their usual use for naming groups of quotations, as well as other uses such as the one I am suggesting in this article. My preference would be to call ATLAS.ti-codes Concepts, in the sense that a concept is simply a name given to any grouping of things.
An ATLAS.ti-code and an ATLAS.ti-memo are essentially variants of the same ATLAS.ti design: a list of things that have a name, a comment or writing area, and the ability to link to other things in the program. But some of their features are different. Here’s a table of the similarities and differences that are relevant to my argument.
Based on these similarities and differences, I have come to think of an ATLAS.ti-memo as a different kind of ATLAS.ti-code – like a second set of ATLAS.ti-codes that have less functionality. Rather than using both of these features for different conceptual tasks, why not just manage a single set of ATLAS.ti-codes in a single Manager for all the conceptual tasks in a project – not only for codes attached to data, but also for representing themes and their associated analytic memos?
Here is an illustration of a project with a clear-cut distinction between the ATLAS.ti-codes that represent “regular” codes attached to data (indicated by the first number after the codename), and those that represent emerging themes that are not attached to data (these have a zero after the codename). It does not have to be that way – data could be attached to the themes as well if that serves the needs of the data analysis.
In this network the five themes are at the top of the visual display, and the other “regular” ATLAS.ti-codes are beneath them. They do not have to be kept spatially separate, it depends on what the visual display is representing, and on personal preference.
The ATLAS.ti-codes representing themes start with a “#” to distinguish them, and also to have them cluster together as a separate group of ATLAS.ti-codes in the top portion of the Code Manager.
There are several advantages to this style of using ATLAS.ti. The first advantage is that all the concepts in the project – whether lower-level codes or higher-level themes – are available for view in a single manager window. This is efficient – only one manager has to be opened and take up space on the screen, and you do not have to remember which concepts are in which manager. And the writing tools – formatting, editing, and so on – are identical in the comment areas of ATLAS.ti-codes as they are in ATLAS.ti-memos.
A second advantage is greater flexibility. In many methodologies there is a clear distinction between the lower-level codes attached to data, and the higher-level themes used for writing analytic memos. However, if the themes are represented with ATLAS.ti-codes (rather than ATLAS.ti-memos) they can also be used to code data in the normal way with the full features of ATLAS.ti-codes available, if some items of data are considered particularly relevant to illustrate a theme. It is true that an ATLAS.ti-memo can also be linked to a quotation, but the features for retrieving and outputting them are not as efficient or comprehensive, and I find it more convenient if everything is working in the same way. Using ATLAS.ti-codes for the themes also better fits those methodologies in which the role of “codes” and “themes” starts to blur as the analysis proceeds, becoming part of a single continuum of concepts with varying degrees of abstraction.
A third advantage is that all the concepts in a project, whether considered codes or themes, are all represented by ATLAS.ti-codes that can be linked together in ATLAS.ti-networks. All the concepts in a project can then be viewed together, as a whole, and analytic memos about themes (written as comments on their ATLAS.ti-codes) can be written in an ATLAS.ti-network while viewing their relationship to underlying or supporting lower-level codes, with easy navigation available in the ATLAS.ti-network to the underlying quotations. The process of analytic memo writing is therefore kept close to the data by seeing all or many of the project’s concepts together, in context.
Again, it is true that an ATLAS.ti-memo can also be linked to ATLAS.ti-codes and viewed in a network – but not linked with “relations”, the more powerful named links that are used to link ATLAS.ti-codes and that have much greater functionality than the simple link between an ATLAS.ti-memo and an ATLAS.ti-code. You can then use all ATLAS.ti’s tools for furthering an analysis with all the concepts in the project – higher-level themes as well as lower-level codes – by creating supercodes, exploring co-occurrences, querying the coded data, and outputting coding frequencies, because all of these tools work only on ATLAS.ti-codes, not ATLAS.ti-memos. ATLAS.ti-memos, off by themselves in their own manager, are then available for other non-analytical writing purposes, such as those mentioned earlier.
This article is a polemic – it challenges the usual way of working with ATLAS.ti-memos without presenting the other side of the argument. In my workshops I long ago gave up trying to authentically present any reasons for using ATLAS.ti-memos for analytic purposes. I won’t start trying now, but will leave that to others who have their own good reasons for using the program in that way. Perhaps some methodologies lend themselves to having the coding scheme kept separate from the higher-level themes and analytic writing. Or maybe it is a personal preference for using ATLAS.ti-memos in that way. At the least, I hope I have got you thinking about the value of harnessing ATLAS.ti’s powerful features in creative ways.
Chang, H. (2008). Autoethnography as method. Walnut Creek, CA: Left Coast Press.
Charmaz, K. (2006). Constructing grounded theory. Thousand Oaks, CA: Sage.
Glaser, B. G., & Strauss, A. L. (1967). Discovery of grounded theory. Chicago: Aldine.
Guest, G., MacQueen, K. M., & Namey, E. E. (2012). Applied thematic analysis. Los Angeles: Sage.
Thorne, S. (2008). Interpetive description. Walnut Creek, CA: Left Coast Press.