Qualitative Data Analysis with ATLAS.ti: Author Interview

Qualitative data analysis is a MethodSpace focus for April. One topic is the use of software to aid the process.  See a post on Qual Data Analysis with Software, and the entire series of posts about qualitative data analysis.

You found, downloaded, or collected qualitative data... now what? NVivo and Atlas ti are popular software options for analyzing diverse types of data for qualitative research. Two important books that guide researchers who want to use these CAQDAS are newly updated. In this post Susanne Friese discusses the new edition of Qualitative Data Analysis with ATLAS.ti. In a previous post Kristi Jackson, co-author with Pat Bazeley, discussed Qualitative Data Analysis with NVivo.

JS. The new edition of your book, Qualitative Data Analysis with ATLAS.ti, is now available. Can you tell us briefly what you've updated in the third edition?

SF. The book has been updated for version 8 Windows and Mac. Mac users will find instructions for all skills trainings on the companion website. The changes are quite extensive, as version 8 of ATLAS.ti is a completely new software. In addition, 9 years have passed since the writing of the first edition and I have developed my thinking about computer-assisted qualitative data analysis further. This is also reflected in the book.

JS.  What features of ATLAS.ti are difficult to learn and use, and how will your book help researchers to move forward with their data analysis? 

SF. It starts with setting up a project. Knowing what you can do with coded data, already helps making the right decisions when creating a project. If you are unfamiliar with the software, you don't know this yet. Here I give the reader first guidance so that they don't run into problems later.

The next step is learning how to code.Technically, that's easy. The construction of a meaningful code system,however, is less self-explanatory. Chapter 5 therefore deals exclusively withthe methodological aspects of building a coding system. I show typical mistakesand how to avoid them, such as the "code swamp".

Teamwork and Intercoder Agreement (ICA), more use cases than functions, are two further topics that need a bit more explanation. I describe different scenarios, explain the methodical prerequisite for ICA and guide the reader step by step through the individual work phases.

JS. There are a number of CAQDAS options on the market. What do you feel is the benefit for choosing ATLAS.ti? What types of qualitative research are best served by analysis with ATLAS.ti?

SF. To give a sales pitch is not really my cup of tea. There are several features that all programs have, but each of these packages also has unique features. Those familiar with ATLAS.it know that these are the quotation level and the networks. Although other packages have something similar that they call a model or map, ATLAS.ti's networking capability has never been surpassed. In terms of analytical approaches, ATLAS.ti supports interpretive approaches better than others due to the quotation level. However, at the other end of the spectrum, ATLAS.ti also gives numbers like absolute and relative frequencies, allows for data normalization and export to statistics programs, etc. that serve those working with a mix-method approach.

Another aspect I like about ATLAS.ti isthat the development of software is driven by an innovative team focused on theneeds of qualitative researchers. An example of this is that ATLAS.ti hasalways offered a non-proprietary XML-based project export format that othersoftware companies used to provide an import function without offering anythingof equal value in return. See this article written by Thomas Muhr already publishedin the year 2000.

Thanks to the initiative and relentlessefforts of Jeanine Evers, this vision of Thomas Muhr has become true. ATLAS.tiis one of the founding members of the Rotterdam Exchange Format Initiative (REFI), the consortium that governs and designsthe interoperabilitystandard QDPX, andwas amongst the first who released this new export format that allows projectsto be exchanged between different CAQDAS packages when it was launched on March18th this year.

JS. Can ATLAS.ti help qualitative researchers who are using Big Data? What about visual data, media, physical artifacts?

SF. If someone claims that one of the CAQDAS packages is suitable for analyzing Big Data, it would be wrong and misleading. ATLAS.ti and other CAQDAS can not handle millions of records. Some are better than others, but most are very slow or crash if you want to work with 2000 or more documents. This does not mean that the AI ​​functions developed in connection with Big Data in recent years, such as sentiment analysis or entity recognition, could not be useful. I think such methods are a good and useful addition to manual coding but will not replace it. As CAQDAS software offers more cloud connectivity, larger online data archives can be accessed as a benchmark. This will significantly improve the results of AI applications, even for smaller data sets.

ATLAS.ti was already able to process visual data (pictures and videos) in 1996 with the first version of Windows. This was one of the reasons why I chose ATLAS.ti. Social media data is an important data source today, and I am sure that ATLAS.ti will offer more social media import options in the future.

Physical artifacts would need to be digitized as either image or video. Then you can also analyze them with ATLAS.ti.

JS. What research experience prepared you to develop this comprehensive guide?

SF. My not so straightforward career path I would say 😊. This is how "my points" connect:

I met John Seidel, the developer of theCAQDAS software The Ethnograph in 1992 while studying at Oregon State Universityand worked as a research associate for him for two years after completing mymaster's degree. He is a trained sociologist and not a software developer. Ilearned a lot from him and am grateful that I had him as a mentor. For example,the N-C-T model that I present in the book is based on his ideas and I havedeveloped it further over the years. After returning to Europe and starting apost as a research fellow at the University of Sussex in 1994, John put me intouch with Nigel Fielding, Ray Lee and Ann Lewins at the nearby University ofSurrey at the launch of the CAQDAS project. As part of the project, I startedteaching CAQDAS programs such as The Ethnograph and Nud*ist. In 1996 I returnedto Germany and it was John who encouraged me to start my own companyQualitative Research & Consulting, QuaRC for short. He also came up withthe name. If you want to know more about it, here is the story on my website.

When someone asks me about mydiscipline, I say, I'm a social scientist. I have never studied a classicalsubject such as sociology, psychology or economics. I started with nutritionand home economics in Bonn, Gemany. At Oregon State University, I studiedFamily Research Management and Marketing (Consumer Behavior). Then as Imentioned, I worked for John, my entry into computer-aided qualitative dataanalysis. The research project in the UK was anchored at the Institute ofSocial Psychology and Experimental Psychology. My dissertation was in the fieldof consumer economics and this resulted in a position as assistant professor atthe Copenhagen Business School, not a match made in heaven. Between 2006 and20011, I headed a Methodology and Media Center in the Department of Sociologyat the University of Hanover. Most recently, I was a researcher at the MaxPlanck Institute as part of a larger research network in Digital Humanities.The all-connecting topic since my master's degree was computer-aided qualitativedata analysis. This is my area of ​​expertise and this has prepared me over the years to write this book.

JS. You offer lots of "skills training" exercises in your book. How do you suggest readers use them?

SF. Readers should click through the individual exercises 😊. You cannot learn software just by reading something about it. The book is a hands-on book. I know from the readers of previous editions that they have the book on their desks next to them and work through the chapters in parallel with the advances they are making in their research projects.

Workshop participants can use the bookto re-read what they learned and heard in the workshop. Teachers can use thebook as a textbook and perform the exercises in seminars. There are severalsample projects on the companion website that can be used in combination withthe skill trainings.

JS. Is there anything else you would like to add about ATLAS.ti or the new edition of your book?

SF. I think this issue is the best version I've written so far. With each issue you learn something new. The feedback that I got from the reviewers in designing the 3rd edition and during the writing process was tremendously useful. Regarding ATLAS.ti, I would like to thank the wonderful, multi-faceted team. It's great to work with you.

Also see: Learn from Cases: Using ATLAS.ti