Creating the Links: An Exploration of Element Relations Within ATLAS.ti

March 16, 2014

B. Jane Scales

In this brief paper, I discuss relations between elements within an ATLAS.ti project or HU, and the development of these relations as one progresses further into the coding.  We know that as the researcher works with ATLAS.ti, he or she observes relationships between quotations and codes, and links them. Quotations, codes, and memos can be linked, and studied in terms of how they relate to one another.  At the beginning of one’s analysis especially, many of these elemental relationships may be associative in nature, rather like a labeling.  However, as one develops past the initial coding, more complex relationships may emerge.

To facilitate this discussion of this topic, I draw from a personal project in which I explore the lyrics of Bob Dylan, his use of syntax, meaning, and patterns of expression.  My purpose in using Dylan’s lyrics is not without some forethought.  I have often found it useful to engage familiar material when exploring a new realm of research or technology.  And while I have used ATLAS.ti for several years, in many ways I feel that I’ve only scratched the surface of its potential.  (By contrast, I am very familiar with Dylan’s lyrics.)  My sense is that under these circumstances, I can better reflect on the process of using ATLAS.ti more effectively.  Then, when more formal projects present themselves, I will be able to use these new skills and insights.

Jane 2 figure 1 Figure 1.  Bob Dylan

The Simple Link

Let’s start by reviewing what linking within ATLAS.ti constitutes.  Dey (1993) aptly explains this idea of “linking data.”  Two separate pieces of data that share some quality or have some relationship to one another can be connected physically or visually.  Within ATLAS.ti, you can see an example of this by simply linking a code to a quotation.  In my study of Dylan lyrics, I looked for songs which mention the word “dream,” or any of its forms (dreams, dreaming, etc.) and created a code for the purpose of identifying and labeling those song titles.  You can see examples of this in Figure 2, which illustrates a set of “simple links.”  The top-most box in the graphic is a code created to label the song titles.  Between the code and the each song title, you see line connecting them.

Jane 2 figure 2

Figure 2. Simple Code-Quotation Link in Network View

Figure 2 also shows that I have completed some of the “primary-cycle coding,” that occurs in the initial phase of the research when one begins simply coding the data. (Tracy 2012:189) and Friese (2012;108) similarly refers to “first-stage” and “second-stage” coding to describe the process in which one eventually exhausts the data present in the primary documents (PDs) being examined.  Identifying which Dylan songs contain some form of the word dream is the first step in setting up my research project.

In Figure 2, one can also see examples of “second class relations.”  This phrase, “second class relation,” describes the type of relationship (or link) between the code and quotation. If you look at any of the lines connecting the “Dream:: Song Title” code, and the song titles, you will notice that they have no labels, but only arrows at either end.  The link between the code and quotation is one of simple association.  While it was important for me to collect the song titles and code them, there is nothing in particular the link tells me beyond this association.  Establishing these associations can be a significant part of one’s project, to be sure.  However, in order to look at meaning, syntax, and usage patterns, I would need to keep exploring.  

During this first-stage of coding, I created many such simple or second-class links.  After identifying songs which contained the term “dream,” I began to code quotations within the lyrics, and ended up with fourteen dream codes.  Among them were:  bad dream, day dreamin’, dream, dreamed, dreamt, dream-lover, dreamless sleep, and others.

Code to Code Relations

The project had achieved some level of complexity. At this point, I had identified 110 uses of the word “dream” in Dylan’s lyrics from seventy different songs, spanning twenty-five albums.  Still, this was just the background work – the initial stages of the project.  I had really not begun to address any of the project goals outlined at the beginning of this article – to examine the “syntax, meaning, and patterns of expression” of Dylan’s use of dreams, but I was anxious to do so.  Because the term “dream” and it’s variations can refer to so many different things, I decided to get a better handle on the term by using definitions from the Oxford English Dictionary (OED).  The OED is a very rich resource of information containing not only definitions, but quotations, etymology, and examples that track the change of meanings over time.  To keep things as simple as possible, I created a primary document (PD) within ATLAS.ti and copied and pasted definitions from the four main OED entries for “dream.”

This made the project richer.  I now had to start thinking in terms of what these instances of “dream” in the lyrics really were.  When Dylan talked about a “dream” was it a dream while he was sleeping?  Was it about a vision? An ideal or aspiration?  These were all possibilities outlined in the OED.  Consequently, I generated more codes to match these varying shades of meaning that Dylan could have used.  One of these OED-based codes I created was:

 OED:: n2a2 asleep involuntary vision

It occurred to me that there was one particular code I had created that clearly had some relationship to the OED code above.  One of the quotations from the song Million Miles (“…..I’m driftin in and out of a dreamless sleep.”) was coded:

dreamless sleep

Clearly, the “dreamless sleep” Dylan wrote of had some sort of connection to the OED:: n2a2 asleep involuntary vision code I created.  A “dreamless sleep” would be one without the “involuntary visions” that happens when one is asleep.  The question was then, how to label the link between these two codes?  What label would succinctly document the relationship between them?  Generally, one sees either a transitive verb, or an intransitive verb and preposition used to label these relationships. After looking in the Code-Code Relations Editor, I did not see a relation that was sufficient.  For that reason, I created a new relation entitled “LACKS,” because one could say that “dreamless sleep” LACKS involuntary visions that occur while one is asleep.

JAne 2 figure 3

Figure 3. Code-Code Relations Editor with edited areas highlighted

 Figure 3 shows how I created this new Code-Code relation.  In setting up this new relation, I established the following:

  •  LACKS is the Internal ID
  • The layout direction is right to left.
  • The Formal Property is Asymmetric.

Jane 2 figure 4

Figure 4. Network view of new code-to-code relation

After creating the LACKS relation, I linked the two codes under discussion together as seen in Figure 4.  Earlier in this text, I discussed “second class” relations.  As one might guess, since there are “second class” relations, it follows that there are also “first class” relations in ATLAS.ti.  Figure 4 illustrates an example of two codes with a “first class” relation with the label LACKS.


In this article, I have taken a taken a small, but significant step in building my project beyond first-cycle coding.  However, it’s also clear my primary-cycle code structure will have to be looked at before I can advance the analysis.  I have coded every instance in which Dylan uses the word dream, but in a way that does not sufficiently support the next phase of my research.  The key to continuing this project is to examine why I was able to understand the relationship between the two codes in Figure 4, and what about my work so far prevents me from identifying additional first class relations between codes.

I will need to look more closely at the context of how the term is used, which means returning to the text and perhaps coding the lines, sentences, or phrases which help me determine exactly how Dylan uses dreams.  No doubt my answer lies somewhere in applying concepts of basic linguistics a bit more skillfully within ATLAS.ti.

Jane 2 Figure 5Figure 5. Code family

My plan is to examine the text around the instances of “dream” I have coded so far and find a more effective strategy to code meaning, part of speech, and tense.  I should then be able to continue discovering code-code relations.

References Cited

Dey, Ian. (1993) Qualitative Data Analysis : A User-friendly Guide for Social Scientists.  London: Routledge.

Friese, Susanne. (2012).  Qualitative Data Analysis with ATLAS.ti.  London: Sage Publications Ltd. pp. 184-6.

Tracy, Sarah J. (2012) Qualitative Research Methods Collecting Evidence, Crafting Analysis, Communicating Impact. Hoboken: Wiley.


Chenail, Ron. “Categorization.” The SAGE Encyclopedia of Qualitative Research Methods. Ed. Lisa M. Given. Thousand Oaks, CA: SAGE Publications, Inc., 2008. 73-74.

Benaquisto, Lucia. “Axial Coding.” The SAGE Encyclopedia of Qualitative Research Methods. Ed. Lisa M. Given. Thousand Oaks, CA: SAGE Publications, Inc., 2008. 52-53.

About the Author

Jane article 1 PhotoB. Jane Scales is the Reference Team Leader, and E-Projects Librarian at the Washington State University Libraries.  She holds a bachelor’s degree in Russian Language from Indiana University, a master’s in German Language and Literature from Ohio State University, and a master’s in information science (MLIS) from the University of Kentucky.  Her research focus includes information literacy, online learning theories, and academic reference services.

Because of the COVID-19 pandemic, we offer temporary easy home-use licenses for institutional license holders, 25% price-reductions on courses and 10% on all licenses. We have new courses both free and premium at our ATLAS.ti ACADEMY and free semester licenses. See here for more -->