Staying Open: The Use of Theoretical Codes in GT

Barney G. Glaser, PhD, Hon. PhD

Theoretical codes (TC’s) are the abstract models that emerge during the sorting of mature memos in to a potential substantive theory. They conceptualize the integration of substantive codes into hypotheses of a substantive theory. The researcher is challenged to staying open to their emergence and earned relevance rather than their preconceived forcing, which is very strong. They not only bring in their framework, but also their theoretical perspective, which can easily force the data beyond emergence. For example, using a basic social process TC requires at least two stages and there may not be a process in the sorting emergence. It may be all just be dimensions or conditional. In this chapter I discuss the skill of staying open to the emergent TC. As the reader knows, there are many TC’s and each has its requirement for use and perspective. In chapter 3 of the book Getting out of the Data: I will discuss more fully all the TC effects that originate preconceptions.


The full power of GT comes with staying open to the emergence of codes that fit with relevance when generating a GT. This power emerges especially with sorting mature memos into theoretical codes for writing up. Substantive coding comes comparatively easy and is exciting giving the researcher the exhilarating feeling of discovery. Putting the theory together with theoretical coding seems for many not so easy. It can have a beguiling mystique, with forcing implications for preconceptions. As Cutliffe says from his experience: “TC usage places the most demand on researcher’s creativity.”

TC’s are frequently left out of otherwise quite good GT papers, monographs and dissertations although they are always implicit, such as range, dimension, or process. The novice GT researcher finds them hard to assimilate into his/her theory, which has to be finished for external requirements. Here I will consider several sources of difficulty with preconceptive consequences in using TC’s. Staying open to the non forced, non preconceived discovery and use of TC’s is the focus of this chapter. For a more extensive discussion on the emergence of TC’s during sorting see chapters 3,4,and 5 of my book: The Grounded Theory Perspective III: Theoretical Coding (Sociology Press, 2005). I hope to add new insights in using TC’s.

Readers who are challenged in staying on the substantively abstract of conceptualization may find this chapter even more challenging. Keeping researchers on an abstract conceptual level is hard for those trained in immediate accurate description, such as medicine, nursing, business, management, social work etc., many of whom are attracted to GT research. Practical considerations of work easily take over. Staying open to the emergent conceptualization will actually increase their power of description, they soon learn. Getting on the TC level of abstraction even more so. TC’s are abstract models that integrate categories and their properties into a theory. They emerge and put a theory together when sorting mature memos. They are easily forced. Thus, staying open to their emergence is not easy for novices. Their use comes with experiencing many research studies as part of the experiential growth of doing GT and learning earned relevance with theoretical sensitivity. Remember they are always implicit in a substantive GT, the simplest being dimensional, if one doesn’t emerge for the novice. TC’s are not to be preconceived forced by a discipline, supervisor or a pet code. Pet codes happen with grab easily, such as basic social process or networking. They spread like wild fire like the fashionable, “self sustaining systems.” They preconceive the theory model easily.

Staying Open

For the researcher, staying open to the earned relevance of a TC means being open to the fullest possible array of TC’s. The researcher learns and masters sensitivity to as many TC’s as possible. The more TC’s the researcher learns, the more this requirement becomes exciting. Substantive theories use them in their name such as “survivalizing of homeless: a basic social process,” or “fluctuating support networks.” There are lists of of TC’s in my books Theoretical Sensitivity, Doing GT and Theoretical Coding: the Grounded Perspective III. TC’s can come from all fields like revolving amplifying causal systems from economics.

Most GT researchers I have read to date get the staying open point easily for substantive coding but not for TC’s. For failure to study TC’s they are not sensitive to what could emerge. Rather, they use the TC of the perspective of their field of origins, hence are likely to be preconceived. They even can describe their GT research by the preconceived TC. For example “I am studying a basic social process,” which is a pervasive, popular TC with grab. I say maybe yes, maybe no. TC perspectives become assumed by the preframed researcher being wedded to say symbolic interaction or social structure categories.

There is nothing wrong with using social structure or symbolic interaction perspective TC’s if they earn relevance through emergence. My effort here is to put a stop to the assumed default relevance, caused by routine forcing irrespective of fit or relevance and remolding GT to just another a QDA method. One reads of Strauss’s conditional matrix everywhere in the literature as if always relevant, which assertion is not grounded. GT is a general methodology that can be used with any type of data and therefore any TC. Therefore, it has no special theoretical perspective or epistemology. The point is always to figure out what the data is, not what it is not. GT just searches for latent patterns in whatever data is being used. GT is a concept indicator method. Not all data is socially constructed, but if the researcher has some, it too will have its patterns. It is up to the GT researcher to tell the reader exactly what data he has and how the concepts and TC’s emerged from it. “All is data” and all data can yield substantive and theoretical codes. GT does not need a “grande” epistomology with a favorite TC to justify its use. Those researchers wedded to a perspective, an epistemology will preconceive their TC use. Staying open to the TC’s that come from all fields and types of data is enriching for the GT research. For example, the random walk model for biochemistry or amplifying causal looping from economics or conjunctional causation in political science are all enriching to watch for. If the researcher stays open.

Staying open to what can emerge can be tuned in on itself, however. It is as if to stay open and suspend preconception cannot be based on the simple ability to suspend knowledge. It has to be based on expected or expert knowledge. Experts in a field find it easy to say a category emerged or a TC emerged, which is really just a product of their advanced training. They will claim preconceptively that their exquisitely tuned capacity guided them where to look to get the best categories and TC’s. It is claimed as an undeniable asset that makes they open to learned and experienced preconceptions. In sum, highly trained people well formed in their field find it hard to transcend their experienced view. They see it everywhere rather that staying open. However much they pretend to be open. They can easily spot preconceptions of categories and TC’s in others. Staying in control wins over staying open. They must be aware of more subtle forcing based on experience when doing their own research. The novice GT research can suspend preconception based on lack of knowledge and training in an area. The experienced GT researcher is not so fortunate in this regard: subtle forcing abounds.

By now the reader may be discouraged and feeling that he/she cannot stay open to TC’s. That is, it is just too hard to leave the comfort, safety of cherished learned and trusted TC’s of their field and colleagues. I say “not so.” They are not to be given up. They are to be added to by learning more TC’s, by becoming sensitive to the additions and then letting learned relevance dictate their use. Staying open to emergent TC’s progresses as the researcher learns more and more. The researcher should study beyond the boundary of his field. Of course this can yield endless possibilities, As one progresses sensitivity to them increases and it is easy to pick up on the model putting together any theory. They are exciting to learn as they give an abstract view of data and grasping them is not hard, once the researcher gets the sensitivity. The wider the array of TC’s that one learns the less the tendency is to preconceive with a pet or discipline TC. I have listed many in my books: Theoretical Sensitivity, Doing Grounded Theory and GT Perspective II, Theoretical coding. Start with reading them to learn staying open to be sensitive to emergent TC’s; Remember they model a substantive theory excitingly well, but the researcher does not have to have one. It is important not to force one on the theory just because one has not emerged.

The excitement of learning TC’s is well put by Walter Fernandez when he says, quite rightly, ”Theoretical coding conceptualizes hypotheses that are integrated into a theory. The grounded integration of concepts is a flexibility activity implying theoretical sensitivity to a number a number of possible TC’s that provides new perspectives that remain grounded in the data.” Fernandez then provides his reader with a two page chart of 26 TC/s. This list helps the researcher stay open to which, if any, TC may emerge. The more TC’s a researcher learns, the less is the tendency to derail a substantive theory into a routine discipline TC. Earned relevance of emergent TC/s that emerge in theoretical sorting of memos is a must, if the researcher uses one. They emerged with the skill development of the researcher. It is part of developing theoretical sensitivity about how they model and how to let their use emerge. Skimming and dipping in of papers from other fields can be fun, quick and easy. Let me give an example on how a TC can pop up.
In perusing a biochemistry paper I came upon the “random walk” model. This means that all variables are in unorganized flux until one variable is introduced and then, all of a sudden, all the variables in flux immediately fall into a stable organized patter. This model is highly applicable to social life and action. People mixing at random and visiting each other in all directions before a meeting suddenly come to order when the host or lecturer appears. It happens in fancy seminars, courts, staff meetings, children’s play yards etc. We can see it everywhere. A “come to order” is announced, in many ways, and the order of the occasion is produced almost immediately.

Another TC that jumped off an economic article I was perusing is “amplifying casual looping.” This TC is part of the interaction of effects family. As consequences become continually causes and the cause become continually consequences one sees either worsening or bettering progressions. This model can apply everywhere in continuing relationships that improve or worsen. It applies to abusive relationships or love relationships. I am sure the reader can see its generality and application booth positive and negative, say in growing spousal abuse. It is an easy TC to preconceive and force, so careful, it must grounded.

Another TC comes from political science and is harder to spot: “conjunctural causation.” It means that a set of causes have to occur in some connection to generate a consequence. The connection between the cases requires a complex set of conditions. Different sets of the same causes have different consequences. Some of the causes are very relevant and some are just triggers. It fits political science on a large national scale, for example organizing various subgroups to vote on measure or when do conditions reorganize to turn a peaceful demonstration into a brutal riot.

These three examples show how complex causal models that emerge can provide integration substantive codes that go far beyond simple causation that is forced “as appropriate“ by local authoritative disciplines. The reader will find it fun to skim theories from other fields to pick up their TC’s and thereby open them selves up to many TC’s. The more this is done, the more the researcher will have the realization that the number of TC’s is endless and therefore staying open and sensitive to what ever TC emerges is the way to do GT. To focus on only one TC preferred by a field is a pure preconceived shut down of GT methodology.

Bear in mind, do not worry if your substnative GT theory has no TC. It will be implicit in the theory. The important ideas is to not preconceive one, just to have one to point to. Also, studying theories in many fields to find TC’s may be a task, exciting for some, but not for others. It is optional. Learn the ones in my books.

TC’s are Slippery

TC’s are the least understood aspect of generation GT When GT methodology is used merely as a legitimating jargon for a QDA then of course understanding TC is moot and one just preconceives the field favorite TC. But when the researcher is genuinely trying to do GT research, the first confusion is theoretically substantively coding the data. Later in the research, when sorting mature memos, starts the confusion is sorting for theoretical codes. Both types of coding emerge from the data and are recorded in memos. They occur in mixes. When a TC integrates the substantive codes with fit and relevance, GT is being generated. For example, a causal model can easily be mixed with a zone of tolerance and two outside sitting points. Learning TC’s emphasizes the earned relevance of these mixes as they model substantive codes. The possibilities are there and grounded. Unlike substantive coding, the underlying groundedness is less clear since they are abstract models of integration based on sorting mature memos for a best fit. Their fit, therefore, is not as underlying tight with the data as a substantive code is. Their organization of a theory is not right or wrong so much as variable on an abstract level. There can be alternatives as the researcher generates and maintains the fit, work and relevance of his substantive codes.

The variability can be slippery and often result in confusion, depression and anxiety over the emergence of the best fit model of integration. Of course, best fit grounding is required in the TC emergence, but given the ready modification of a GT the TC model can easily get adjusted as it emerges. This, of course, can add to its confusion and then forcing of a preconceived TC on the codes as a way out of the confusion that comes with waiting and sorting for the TC of earned relevance. Forcing with familiar field concepts can easily lead to irrelevancies. For example, every GT is not a BSP (basic social process) and rich as this TC is, forcing it on a theory with stages of a process that does not exist can dilute fit and relevance. One goal of a GT researcher is to develop a repertoire of as many TC’s as possible. This maximizes the emergent fitting of the substantive theory into a well generated integrated emergent model.

Some researchers get confusion between the meanings of substantive coding and theoretical coding. Needless to say, substantive codes are the categories and their properties that emerge in conceptualizing the data from the substantive area being researched. They are used to build the substantive theory, but are not theoretical codes. TC’s are used to provide the abstract model that emerges when sorting mature substantive coded memos. TC’s must also pattern out when sorting memos to provide grounded integration of the substantive codes. Preconceived TC’s can easily force the integrative model.

Without substantive codes, theoretical codes are empty abstractions. But substantive codes can be related without explicit TC’s. Without a TC the results is easily somewhat confused and theoretically unclear as to integrative connections. The implicit TC is typically dimensions of a core concept. It is the integrative interaction between substantive and theoretical codes that fully characterize the generating of GT. This is simple to say but leads to confusion since TC’s exist on a higher abstract level of modeling substantive code relationships or hypotheses. However, substantive codes are often called theoretical codes and thus mixing the two usually meaning confusion or just missing TC’s all together.

Everyone after an initial try loves and understands the constant comparative method for generating substantive categories and their properties. As one GT researcher wrote me, “your phrase ‘fluctuating support networks’ has really grabbed my attention” (Holton email 6/9/03). But this joy and grab is not so for TC’s, except perhaps for discovery of a BSP. TC’s are often ignored, left implicit or just plain missed with no understanding. Researchers generate categories naming latent patterns social action all the time. And these names easily grab others. The same researchers often do not systematically sort memos to generate TC’s, except maybe to mumble causes or processes.

Substantive categories grab by denoting recognizable patterns, where TC’s seldom have this grab since they denote abstract models that are usually implicit in the theory, and seldom explicitly mentioned. And it is even more confusing if the TC and the core substantive code have the same name, such as process. Thus it is clear that substantive codes are on a different conceptual level of abstraction. And TC’s are more abstract since they model the integration of the substantive categories which name grounded patterns. Mixing the two types of codes is typical and hard to figure out at times.

Confusions occur like this. A core category may be a TC named such as becoming or cultivating or routing and they are BSP’s, but the BSP is not the core, it just models the core substantive code. Thus, in one dissertation the core category was survivalizing, which was a basic social process. The abstract level relationship of both types of coding is always the same. TC’s are more abstract than substantive categories, no matter what level they start at and GT is readily modifiable, unlike the accuracy of description. In sum, the researcher constantly compares to generate substantive codes from and sorts mature theoretical memos into a what ever TC he best thinks articulates his theory, or he leaves the TC implicit. On the abstract level it is easy enough to distinguish between substantive codes and theory codes. But on the descriptive level they get muddled. This muddling occurs in the writings of many QDA theorists such as Ian Dey or Jan Morse.

Are TC’s necessary? As I have said, the answer is “no,” but a substantive theory is best when a TC’s is explicitly used. TC’s are always implicitly in the theory, even when not consciously used. But a GT will appear more plausible, more relevant and more enhanced when integrated and modeled by an emergent TC. The hypotheses will be clearer and stand in relief from the superficiality of a conceptual QDA. Using a TC from sorting mature memos makes generating substantive categories and their properties easier and the resulting theory more complex and multivariate. As long as there is no preconception, a TC helps more theoretical sampling, theoretical saturation, and delimiting the theory to reach theoretical completeness because it provides an emergent guiding framework. Analyzing a theoretical framework without an emergent TC is harder, but happens. But be cautious, as this is when preconception of a TC occurs and is forced for its guidance in integration. But the TC should emerge however tempting forcing may be. It is easy by prior training to force one on forming the theory as a framing tendency. Resist this preconceived temptation, however strong. Staying open to emergent TC’s is important and totally necessary. They provide the full rich understanding of the substantive theory being generated. They also place the most demand on the generating sensitivity skills of the researcher. Experience in generating theories increases this skill.