Content coding is a method to systematically analyze pieces of information. In this entry, we will focus on analyzing written documents such as interview transcripts or court cases, but you can also use content coding to analyze other communicative expressions such as speeches, videos and media content.

You can code your content in a qualitative or a quantitative manner. For some research questions you might be interested to learn about numbers, because you already know what variables you are investigating: for example, how many of your interview participants experience a certain phenomenon or how many court cases use the same legal arguments.

Or you might be interested in relationships (correlations) between certain factors – for example whether a certain legal argument from a defendant correlates with a certain outcome of that court case. In that case you can consider using quantitative content coding. When coding your content in a quantitative manner, it is important to operationalize your variables beforehand. When you know what variables you are interested in, you can see whether that variable occurs in the piece of data (e.g., an interview or a court case). You can use an excel file and state ‘1’ if the variable occurs and ‘0’ if the variable does not occur in the piece of data. When you have coded all your data, you can use SPSS to see whether there are certain relationships between the variables.

If you are interested in more qualitative information (why/how-questions), you need to code your content in a different manner. This is usually the case if you do not know the specific variables yet, but want to learn more about how something works or why something works that way –­ for example why your interview participants feel a certain way or how courts motivate certain decisions. In that case you can code your content in a qualitative manner. There are different ways to code your data, but the most common way of coding in a qualitative manner consists of the following three stages:

• Open Coding. You start by describing useful pieces of data with labels (in vivo coding), where you derive codes from the data itself. With every new piece of data, you consider whether you could code that new piece of data with your existing codes (constant comparison). If not, you create a new code. In this stage, you will group your codes into categories. This coding process will tell you when you have reached the point of “saturation”, the point when new data will not give you any new codes (information). When you are still creating new codes when analysing your data, you can use theoretical sampling to gather new data (e.g. cases, participants, documents) on certain underexposed concepts.
  
  
• Axial Coding. In this stage you connect the categories to find patterns and connections between codes and categories.
  
  
• Selective coding. You will develop a storyline around your core category and the relation to other categories. After identifying the core category, you go back to your data to validate axial relations and to see if you can further code the data in relation to the main storyline. If needed, you can use theoretical sampling to gather new data until saturation is reached. Your final storyline will be the theory you created on the data (see the entry on grounded theory).

You can use Computer-Assisted Qualitative Data Analysis Software (such as Atlas.ti, MAXQDA or NVivo) for your coding process, but you can also – depending on the size of your data set – doing it by hand using markers.

Which coding method to use depends on the results you’re looking for. If you are looking for numbers or relationships between predefined variables, use quantitative coding. If you are looking for more in-depth knowledge on a certain topic, without exactly knowing what variables to look for, use qualitative coding.