One way researchers analyze data is by using tags to code qualitative research. Once data is collected, the process of naming, describing, and classifying data is referred to as analysis.Tagging data is a significant and influential step in the analysis phase of research.
A tag is a label used to assign meaning, classify, or “code” data. Tags code pieces of meaningful qualitative data in order to:
An example of a common tag used in research is to label or tag data by personas. For example, if we set out to learn about the challenges of remote learning, then we might start by interviewing parents, teachers, and students. As we collect that data, we’d use methods to label and group that data with a tag for teacher, student, or parent. This method gives us new data points to pull apart our data and find patterns.
The primary objective of using tags to code qualitative research is to add a layer of meaning to raw data. With these new layers of meaning, data evolves into information, which is more valuable and rich with potential for learning than raw data alone.
Using tags to code qualitative research requires thought, intention, and some actions that can sometimes feel tedious for the researchers involved.
So is the juice worth the squeeze? Is it worth the effort and time to code qualitative research? The answer, like most topics of research methods, is it depends. The value derived from using tags to code qualitative research depends on the dataset and the tags you choose to use. Not all tags are created equal.
A good strategy with tags can identify tension and discover rich patterns. A poor strategy with tags requires the same time and effort, but won’t yield as valuable takeaways or perspectives.
As a whole, the group of tags you choose to use is referred to either as your framework or, in a slightly more technical way, as your tag taxonomy.
In biology, taxonomy is the science of naming, describing, and classifying. Anthropologists use taxonomy to identify and enumerate different species, including new ones. As the naming and classification continue, anthropologists create a repository of knowledge underpinning the management of biological diversity.
The emergence of this naming system establishes a defined methodology of analysis. The naming system also comes with an accompanying description that describes the naming process so that other people can name a similar finding by themselves.
In short, taxonomy identifies a difference, helps other researchers contribute to a data repository, and helps people make sure they're talking about the same thing. These are the principles of creating a repository taxonomy through tagging.
There are several ‘tried and true’ methods and frameworks for tagging qualitative data. Usually we call upon those frameworks in the beginning of a new research initiative and incorporate our tag taxonomy to the research documentation. Other times we might realize new tags we want to include to our taxonomy once we’ve started analysis and realize a new classification is needed.
One thing to consider about a tag taxonomy is that depending on the data practices influencing your research as a whole, you might only be concerned with a tag taxonomy for a specific project. Alternatively you might also contend with how your body of research work contributes to a larger repository or collection of research projects. In that case you might need to include some broader tags into your analysis phase and consider how those tags will compliment or work together with your project-specific tags.
In either case, there are a few ways to break down and think about your tag taxonomy which are outlined below.
Global tags are excellent for zooming out of data and tracking data sets across different projects on a repository. Generally, global tags are agnostic to the set of data and can be used over and over again for a variety of research initiatives. A few examples of global tags are:
Here are some examples of specific global tags to get started with:
Project tags are good for zooming in on a set of data. These types of tags are specific to data and research goals. Examples of tags are:
An introduction to using tags to code qualitative research would be incomplete without exploring a few common challenges associated with tags.
It’s always a good exercise to question or poke at your tag taxonomy to find potential bias. Creating a tagging system to name and classify what you already know makes it easy to establish a taxonomy that focuses or indexes on existing assumptions.
One way to navigate confirmation bias is to conduct assumption mapping. Gather your research team, discuss your assumptions, measure your confidence for each assumption, and then nudge yourself toward using tags that index on the unknown.
It's not obvious what methodology researchers should use to tag data. What's more, taxonomy is a learned methodology, which means researchers cannot apply what they do not know. Consequently, it's up to lead researchers and design teams to introduce tagging as a methodology to researchers and for those decisions to be made, then enforced through good data management practices.
It takes time to analyze data, create a suitable taxonomy, and apply it to existing data. Without appropriate research support within an organization or research team, researchers may not find the time to create and use tags. In the face of time constraints, researchers have to do cost-benefit analysis to help prioritize a minimally viable taxonomy or get by with more anecdotal analysis over a rigorous approach. Fortunately more tags does not necessarily correlate with better research outcomes and often narrowing the scope of a taxonomy is a better alternative than a bloated and more time-consuming approach.
There are some instances where automated tagging can be a worthwhile investment. For example, in very large datasets or when faced with time constraints, it can be worth it to invest in tools that automatically classify data. For example, in Notably we use AI to help researchers automatically analyze sentiment and more.
Creating a tagging taxonomy as a team is a great way to create a reliable coding system, reduce bias, and discover new analysis opportunities.
Using tags to code qualitative data solves an analysis problem. To create the right tags, researchers could consider questions such as: