ATLAS.ti networks as a mechanism for establishing trustworthiness

On Tuesday morning I attended a special topics webinar on ATLAS.ti network views conducted by Susanne Friese.

The network view tool in ATLAS.ti enables the researcher to create "interactive mind maps," something I have already written a bit about in a previous post.

Basically, the hermeneutic unit (the ATLAS project file) is the total network. All objects added to or created within ATLAS (documents, artifacts, memos, quotations, codes) instantly become nodes in the network when they are linked to something else (for instance, a code to a quotation, a memo to a code, and so on), and these links may then be visualized.

I sat in on the session in hopes of learning more about this ATLAS function, which I have come to appreciate as a scaffold for both analysis and writing (harder and harder to separate these two processes, I am learning).

As I listened to Susanne Friese and her "vondervul" German accent, I also made a few connections between network views and this week's readings in EP659, which focus on trustworthiness and issues of quality in qualitative inquiry (Anfara, Brown & Mangione, 2002; Tracy, 2010).

According to Anfara et al., "...[A] key part of qualitative research is how we account for ourselves, how we reveal that world of secrets" (p. 29). Neither Anfara and his colleagues nor Tracy specifically mention CAQDAS as a tool for ensuring quality, but both articles discuss transparency of data management, coding, and analysis as a mechanism for strengthening research quality.

Tracy, for example, provides a more than adequate rationale for use of digital tools in her discussion of "rich rigor," one of the eight major criteria in her reconceptualization of excellent qualitative research. She writes,

Rigorous data analysis may be achieved through providing the reader with an explanation about the process by which the raw data are transformed and organized into the research report. Despite the data-analysis approach, rigorous analysis is marked by transparency regarding the process of sorting, choosing, and organizing the data. (p. 841)

And Anfara et al.’s example of “code mapping” (p. 32) is quite simply a picture of one researcher’s analytic process that could easily be depicted using ATLAS.ti’s network view. Although, with charts and tables, one must be critically conscious of not merely showing a hierarchy of codes, which Friese adamantly warns against. (See "Analytic capabilities of network views" below.) The researcher builds a network view in ATLAS based on his or her interpretation of relationships across codes and categories, thus making the network view tool indispensable for conducting constant comparative analyses, building "audit trails," and "documenting the procedures used to generate categories" (Anfara et al., p. 33).

Creswell (2013) is more direct in his endorsement of CAQDAS. In the chapter on validation strategies in the newest edition of his Qualitative Inquiry and Research Design, he mentions the use of computer programs to assist in recording and analyzing data as one of several ways for enhancing the stability and dependability of findings, or, what is known in positivistic terms as "reliability."

Networks are compelling visual diagrams that add transparency to the researcher's process, and I am now considering how I might incorporate them into the write-up of my own research findings or my dissertation appendices.

Here are a few other notes I took from Tuesday’s webinar:

Basic concepts of network views

• Strong and weak links are denoted by solid (strong) and dotted (weak) lines
• Weak links between nodes are unnamed. They exist between memos and quotations, codes and quotations, memos and memos, and between families and their members.
• Named links express relationships between two codes or two quotations
• Named links may be "directed/transitive" ------> and "non-directed/symmetric" <-------> 
• Background colors in network views coordinate to code colors (if you use color)
• Hyperlinks occur on the data level, between two codes or between two quotes

How to link objects

• You can drag and drop from anywhere, bringing objects into the network view manager. You can also import nodes into the network (any object from the HU). For instance, you can select a code and import all its quotations.
• Take time to play with the display options under the Display drop-down menu
• In the relations editor, you can create your own relations with colored lines of different point sizes (See “How to create relations” below.)
• Comments can be added to relations and are denoted with a tilde just as with other comments in other areas of ATLAS
• NEVER DELETE AN OBJECT from the network view, use the "remove from network view" option instead
• Creating links between data opens up different kinds of relations, as opposed to linking codes, which is more conceptual
• You can save a network as a graphic file (png, gif, jpg) and insert into a PowerPoint or MSWord doc

How to create relations

• Open Relations Editor and expand the window until you see the Edit tab
• Decide on the relation you want and crate a unique identifier, the first three letters and then the actual memo text (e.g. REA is the identifier for "is reason for")

Working with hyperlinks

Example of "star" links

• Hyperlinks are "stars" or "chains" of links
• Quotes may be linked within and across documents

Analytic capabilities of network views

• Codes are just topics and areas of interest within the data, they describe
• Networks take it to the conceptual level, so networks are for linking across categories and depicting relationships between data, not for building code hierarchies. DO NOT USE NETWORK VIEW FUNCTION TO REPRESENT YOUR CODE STRUCTURE. IT'S TWO DIFFERENT LEVELS OF ANALYSIS.
• Use network view after coding when you start to see relationships.
• ATLAS.ti version 7 allows you to filter codes that you bring into the view. You can hit Reset Filter to bring back all codes.
• Code families are just grouping mechanisms for codes. Families are filters: create family, then turn on filter.

References

Anfara, V. A., Brown, K. M., & Mangione, T. L. (2002). Qualitative analysis on stage: Making the research process more public. Educational Researcher, 31(7), 28–38.

 Creswell, J. W. (2013). Qualitative inquiry and research design: Choosing among five approaches (3rd ed., Kindle version.). Los Angeles: SAGE Publications, Inc.

 Tracy, S. J. (2010). Qualitative quality: Eight “big-tent” criteria for excellent qualitative research. Qualitative Inquiry, 16(10), 837–851.

ATLAS.ti networks as a mechanism for establishing trustworthiness