11.2. Creating Arguments out of Codes

Reading the work of celebrated sociologists can be intimidating. How did they conceive of such provocative, nuanced, inspiring, and earth-shaking explanations for how the social world operates? Perhaps we are slightly overstating here your personal admiration for sociological theory, but we think you will agree that arriving at compelling theories is a challenging task. The coding process we just described will move you far along this path if you invest sufficient time and effort into it. That said, there are a number of intellectual moves you can take to spark your creativity and guide you to potentially valuable insights.

We’ll discuss three of these strategies: neologisms, typologies, and concept mapping. As you narrow your codes further and further through the open coding process we described earlier, you’ll start to see the limits of existing theories. You’ll encounter phenomena that are distinct from what past theorists have focused on, or are variations on those prior themes, but do not yet have a fitting name. You’ll think of more precise or insightful ways of dissecting and categorizing social reality. And you’ll come up with persuasive new models for the specific processes you’ve observed.

Neologisms

One approach to generating new theory is by creating neologisms—new words or expressions, which in a research context would entail finding a precise, illuminating (and hopefully catchy) label for a particular phenomenon. These neologisms are often italicized in books and articles to emphasize that they represent a theorist’s proposed way of understanding an aspect of reality.

Students are often a bit hesitant to put forward their own neologisms. They tend to want to defer to existing theory, stamping everything they see with their favored theorist’s kosher labels. This isn’t how science advances, however. As a critically minded and independent-thinking sociologist, you will want to question whether the prevailing theory—the prevailing term—actually encapsulates the complexity of social life in an accurate and thoughtful way. Don’t be afraid to weigh into these debates, even if you are a novice researcher.

Now, it is very true that neologisms can be overdone. As you read through the literature on any topic, you’ll inevitably find yourself spotting a writer’s italicized term and rolling your eyes. It may come across as awkward to you. Or you may feel existing terms cover the same ground more than adequately. When creating a neologism, you should keep these concerns in mind. For the first problem, the phrasing of the term, you might want to imagine yourself referencing the term in another writing context. Does it roll easily off the tongue? Does it need an explanation attached every time you use it, or can it potentially stand alone (the latter obviously being better)? Does it come across as gimmicky—a transparent attempt to exploit an issue briefly in the news, to be oh-so-cute in your self-important jargon, or to add hipness to the field (always a tragic endeavor)?

You should also ask yourself whether the literature really needs a new term for this phenomenon. Review any existing terms, comparing and contrasting them to your own and seeing if your new take adds something to the conversation. The answer here will often depend on whether you’ve successfully uncovered a wrinkle in the existing theory—some aspect of the phenomenon that hasn’t been sufficiently examined, and that therefore calls for a new label. One way to spot these shortcomings in current terminology is to diligently code the various dimensions of the topics or themes you observe in the ways we described earlier, which can flag specific areas the literature has neglected.

You should also note that you don’t need to think up entirely new terms to advance the literature in your field. You can also contribute just by identifying elegant and instructive new ways to rework existing labels. A catchier neologism may replace the staid term currently being used. More importantly, if your new wording includes a vivid and accessible metaphor, it can help more people grasp the underlying idea. A classic (if somewhat contentious) example is the Big Bang theory, the now well-established view among physicists that the universe began in a state of unimaginably high density and temperature and then expanded outward. The Belgian priest and theoretical physicist Georges Lemaître proposed the idea of a universe spreading out from a single point, which he called the “primeval atom” (Soter and Tyson 2001). It was actually a critic of this theory, British astronomer Fred Hoyle, who coined the term “big bang.” (Hoyle favored an alternative explanation for the universe’s origins, the so-called steady-state model.) The “big bang theory” spread like wildfire within the scientific community and then the public at large—to the point of having a middling TV show named after it. And you can see why: as Hoyle himself pointed out, the words “big bang” provided a ready visual image that made it easy for even lay audiences to get the gist of the theory. Nevertheless, scientists continue to gripe that the words “big bang” inaccurately imply that the universe’s origins involved an explosion in space (rather than an expansion of space), and that sound somehow travels in space (Kragh 2013; Mathew 2013). For better or worse, the term has stuck.

Another alternative to thinking up whole new terms out of thin air is to borrow the terminology used by your research participants. As part of the aforementioned “folk theories” they cultivate, cultures and subcultures have unique ways of naming practices, beliefs, and other phenomena relevant to their members. You can advance existing theories just by bringing to light these examples of insider knowledge. Through the process of open coding, such terms will naturally stand out, and you should keep an eye for how they might help explain what goes on in the settings you observe.

One danger of this approach, however, is that you may naively adopt the perspective of your participants, who—as with any group—have particular moral and political outlooks that can blind them to other ways of looking at the world. They may not actually do what they say (or what their terms imply), and they may buy into dominant ideologies that could bias the way you interpret their actions. An example of both the strengths and potential pitfalls of adopting folk terms from your participants is the “street” and “decent” terminology that ethnographer Elijah Anderson put forward in his book Code of the Streets (2000). In the Philadelphia neighborhoods that Anderson studied, “street” was regularly used as a label for people more or less involved in criminal activities, whereas “decent” described upstanding citizens who stayed off the gang corners. These terms provided a powerful way of understanding how neighborhood residents navigated their social world. At the same time, they also garnered criticism for uncritically accepting moralistic and simplistic folk categories that “decent” folks appeared to be imposing on an unseemly “street” element—a charge that Anderson vehemently denied (Anderson 2002; Wacquant 2002).

Typologies

As you break down the various dimensions of a concept in the open coding process, you will find yourself creating typologies—classification systems with distinct categories that capture different varieties of a given phenomenon. You can contribute to the scientific understanding of your topic in important ways by either devising a new way of typologizing a phenomenon or revising an existing typology to account for something it missed. A good typology helps us make sense of how a phenomenon varies and what factors might contribute to that variation. We’ve already mentioned one simple typology—Elijah Anderson’s “street” and “decent” dichotomy—which categorized neighborhood residents according to their relationship with illicit activity, and which was drawn from the folk language of the culture Anderson observed. According to Anderson, the two groups could be distinguished by whether they upheld middle-class values (“decent”) or defied them (“street”), conformed to or deviated from social norms, had traditional or nontraditional family structures, and lived private versus public lives. You could imagine how these various moral and cultural dimensions of “street” and “decent” life may have originated from descriptive codes created during the open coding process, which his analysis could have knitted together as dimensions of a broader cultural orientation toward the neighborhood and its criminal activity.

Another classic example of a simple but provocative typology is Annette Lareau’s categorization of parenting strategies in her ethnography Unequal Childhoods (2011). Based on the observations that she and her colleagues conducted in households with children, Lareau argued that middle-class parents engaged in more intensive parenting—what she called “concerted cultivation.” Like a gardener working with a delicate flower, they invested large amounts of time and effort into personally nurturing their children’s academic and social skills. For instance, these more affluent and educated parents would fill children’s schedules with enrichment activities, regularly engage in intellectually challenging, adult-like conversations with them, and model for them how to interact assertively with the authority figures in their lives. In contrast, working-class and poor parents pursued a strategy that Lareau called the “accomplishment of natural growth.” These parents did not love their children any less, Lareau pointed out, but they believed—based on their own upbringing and culture—that it was better to let children roam free, without a parent stepping in to instruct them or intervene on their behalf.

The two parenting styles are summarized in Table 11.1. Let’s consider how a typology like the one in the table might emerge during the coding process. Perhaps you might start this analysis with broad descriptive codes like “language use” and “intervention in institutions.” Breaking down the wide variation in these observed activities could lead you to develop arguments about how parents of different social classes behave in different ways within these areas. Moving in the opposite analytical direction, you could also merge these codes. Recognizing variations across specific dimensions of parenting could prompt you to see them all as part of a larger class divide of more active and passive engagement with children. (We’re only guessing at the analytical process that Lareau undertook, but you can listen to her describe for herself the parenting typology she developed in Video 11.1.)

Typologies can describe dimensions of any particular phenomenon—a range of diverse explanations, consequences, processes, and so on. It helps when typologies draw from an evocative and accessible metaphor, like the gardening imagery that Lareau used in her terms. You should also note that the categories in a typology typically represent ideal types—varieties of a phenomenon with particular characteristics imagined in their purest form. Using ideal types keeps a theory simple and streamlined. The real-world cases you observe may not fit perfectly within a single bucket, and some may span multiple ones. This is fine. In fact, it can inspire you to think of the reasons that certain cases straddle categories.

Another thing to keep in mind when constructing typologies is whether the concepts you’re describing are being treated as independent variables or dependent variables. Are you trying to explain why the variation across dimensions is occurring, or do you want to understand the consequences of that variation? Oftentimes, the categories of your typology can be plausibly viewed as both causes and consequences, but for the sake of generating new theories, it can be helpful to be explicit about the orientation here you’re taking.

Table 11.1. Typology of Parenting Approaches in Lareau (2011)^[1]

Obviously, you don’t have to limit yourself to the sorts of binary or dichotomous typologies we’ve been discussing so far. You can easily identify three or more categories of a given phenomenon during your coding process (but note that adding more and more categories can clutter your theory and make it less easy to apply to other contexts). Furthermore, you can use two-by-two tables of the sort depicted in Figure 11.9. In this analytical approach, you tease out how variation across two dimensions leads to distinct outcomes. This can help you understand how the categories you identified for your typology relate to one another.

In the example, Victor and his collaborators mapped out a typology of entrepreneurship based on the degree of creative independence (the extent to which entrepreneurs could conduct their work activities as they pleased) and the degree of financial independence (the extent to which they were obligated to financial backers) that people who started their own businesses enjoyed (Doody, Chen, and Goldstein 2016). More or less autonomy in these areas could situate an entrepreneur in one of the four categories described in the two-by-two table: a Silicon Valley entrepreneur (e.g., a tech startup founder reliant on venture capital funding), a Main Street entrepreneur (e.g., the owner of a mom-and-pop business that sustains itself on its own income), a corporate entrepreneur (e.g., an individual in a large company incentivized to create and manage new ventures), and a self-employed entrepreneur (e.g., an Uber driver or other gig worker constrained by the rules and oversight of a platform).

With a two-by-two table, you are explicitly contrasting the groupings you observe in your data according to how their characteristics in the two highlighted areas compare. You can imagine how such a comparative approach could naturally evolve from the process of coding similarities and differences across your cases and explaining why cases fall into particular groups (possibly ones that your participants have already identified as being important in their folk theories). Again, you can go beyond the simple categorization scheme in this example—having three or more rows or columns in your table, or locating groups somewhere on a continuum rather than discrete boxes—but this adds complexity that may make your theory unwieldy.

If you want to get really convoluted, you can propose typologies of the sort described in Figure 11.10, which includes four moral perspectives and compares them across several dimensions. One thing to note about this approach is the importance of a clear visual representation of your theory. This goes for distinguishing between categories or dimensions within a typology, but also for presenting theorized relationships between concepts, which we will discuss next.

Concept Maps

In previous chapters, we’ve described how important concept mapping is to understanding existing theories and the relationships you plan to explore with your empirical research. Now we want to talk about how concept mapping can help you create or refine theories at the data analysis stage.

Concept mapping involves visualizing concepts and the relationships between them, generally using boxes and arrows (with the direction of the arrows showing the presumed direction of causality). You can lay out these concepts and relationships in a word processor like Microsoft Word or Google Docs or in a presentation program like Microsoft PowerPoint or Google Slides, which have simple-to-use tools for drawing and linking boxes and arrows. (You can also just start off with a piece of paper or a whiteboard.)

There are several ways to profitably use concept maps to create or improve upon theories. In storylining, for instance, you take the categories and relationships depicted in your concept map and use them to tell a causal story about the observed phenomenon. If sufficiently plausible, this story can be the central to your study’s findings. (As we’ve discussed in previous chapters, you should do your due diligence in elaborating on any alternative explanations and not overstating the degree of certainty in this theory that your data justifies.)

You might also find it useful to compare the concept maps that emerge from your coding with any concept maps you generated for previous studies (see the example in Figure 11.11). What does your study include in its concept map that other studies do not? This difference becomes the basis for theoretical innovation. (You might even want to embark on this process of comparative concept mapping during the early literature review stage of your research, which we discussed in Chapter 5: Research Design.) If your concept map is exactly the same as that for another study, your study may still be contributing in a substantial way because of the specific methods it uses or the specific sample it has gathered. Visualizing your research in this way can help you pinpoint its contribution.