Best Practices

Nick Woolf's third column for the ATLAS.ti newsletter.

How to bring structure to a list of codes – Part 3: Networks

by Nicolas Woolf, Ph.D., Director of Woolf Consulting


Nick Woolf is the founder of an independent training company devoted exclusively to ATLAS.ti and has provided training and consulting in ATLAS.ti since 1997. His workshops (available throughout North America and in refresher web conferencing sessions) emphasize learning ATLAS.ti in the context of designing and conducting a data analysis. Nick conducts his own qualitative research programs and also provides on-going assistance and project management to individuals and research projects. More information is available at www.learnatlas.com.

Throughout this series, I have suggested that ATLAS.ti is best thought of as a concept database, to which you can provide concept names, or codes, for a multitude of qualitative analysis tasks. The program helps you bring order and structure to your growing list of codes in four main ways. In my previous articles I described the first three: code prefixes, families and supercodes. Now, for my final installment, I will introduce the fourth: networks.

Networks are simply graphical views of a part of a project. Anything in ATLAS.ti can be displayed in a network, e.g. quotations, PDs, families and memos. Networks are primarily used for the analytic purpose of moving your thinking and your project forward, but they also support certain housekeeping tasks – merging codes, sorting out a complicated web of coding between a group of quotations and a group of codes, etc. – more effectively than the other tools.

But here we are only talking about displaying codes in networks, i.e. moving a selected group of codes into an appropriate relationship and linking them with “relations.” The overall purpose is to make visible what is already in your head, consciously or unconsciously, and to allow the changing relationships between the codes to be represented in their evolving form.

The network tool is much more than a simple drawing facility: networks of codes are not separate from what is in the HU, as an independent drawing module might be. When codes are dragged into a network to be worked on visually, rather than in the list format of the code manager, almost everything you can do in the code manager can then be done in the network. In other words, it's a working tool, an alternative to working from the list of codes that is helpful in many situations – usually when you are in a conceptual framework-building or theory-building phase of a project.

Most importantly, it's easy to navigate or browse to the underlying quotations linked to each code as you're working in the networks. This prevents you from reifying the codes rather than staying close to the data. To stay on track, I make a habit of never maximizing my network window, because this makes it impossible to navigate to the underlying quotations. Instead, I make my network window large but put it in “rollup mode” so that it rolls away when I navigate to underlying quotations – that way, I can still easily read the underlying text.

All qualitative researchers dream of a magic tool for speeding up their analysis. Alas, this tool doesn't exist, but ATLAS.ti networks are the next best thing. They speed up your analysis by explicating sooner rather than later the conceptual relationships that are at the basis of qualitative data reduction, but which are often buried in our unconscious thinking processes.

Making and using networks is easy. First, select a code (or codes) in the code manager. Then, click on the network button, and a network appears. Move it away from the code manager. Now you can drag additional codes from the code manager into the network whenever needed.



Here is a list of basic things you can do in a network. As you get further acquainted, you'll discover a lot more bells and whistles.

  • Link codes: Select a code, click on the link button, and click the cross over the code you wish to link it to. Then select one of the relations that appear. (See diagrams 2 and 3)




  • Work with codes: Right-click on a code to do all the things you do in the code manager, particularly the last item in the menu: list the quotations attached to the code in order to navigate to them and read them.
  • Work with relations: Right-click over the relation name in order to flip its direction, cut it, write a comment, or change to a different one.
  • Save a network: Go to Network/Save as, and save the network for continuing use and development. (See diagram 4).


  • Change a network's appearance: Click on Display to add or remove various visual features of the codes. (See diagram 5).


  • Copy a network into a Word document or other program: Open Word, return to the ATLAS.ti network and go to Network/copy to clipboard. Return to Word and go to Paste special (not Paste) and select the Picture/enhanced metafile.

We have now covered the four main ways to manage a growing list of codes. Each one serves a very different purpose, and they can all be used together in an evolving analysis design to fulfill your macro- and micro- analysis goals.

I'll leave you with a very important rule of thumb: let your use of these tools be driven by the analytic tasks, and not vice-versa. For example, don't start a project by saying, “What code families can I make?” and then look for ways to use these families. Instead, ask yourself, “What is the analytic task I need to accomplish, can I accomplish it by organizing my codes in some way, and which way would be the most helpful?” This is a skill that will grow with each project that you complete and you'll then be harnessing the full power of ATLAS.ti.

Thanks for all the great advice, Nick! We're sad to see this column come to an end, and we hope to read more of your ATLAS.ti expertise in these pages soon!