Code Relations Browser

One of the critical tasks in identifying patterns in your data is looking for relationships between codes. This can be done in many ways, but one tool that is particularly useful is the code relations browser. This is a visual tool which "creates a visualization of the intersections of codes in either a single document, a group of activated documents, or all of the documents in the Document System. This allows you to find connections or “relationships” between your codes.

You can find it in Visual Tools:

Code Relations Browser - appears in Visual Tools

To use it, select the icon or menu item, and then you have several options, including selecting from all codes or activated codes. The code relation browser creates a spreadsheet-like display, so you can choose which codes to have in row, and which ones to have in the column.

You can also select whether you want to identify codes that overlap (where text is coded by multiple codes) or where codes are "near" (and here you can select how close they are to each other in paragraphs). A separate dialog box appears after you select NEAR and run the tool.

In this example, I have activated a group of codes and I am considering how they relate to the full code dataset. I selected "Near" and specified 3 paragraphs.

The browser has a lot of options you can tweak. By default it displays the relations graphically with circles sized in proportion to the number of coded segments. I usually switch to "Display nodes as values" so it looks like this:

How do you interpret this browser?

The easiest way is to select a row (which then gets highlighted) and read across the columns (you can expand the column names as needed either by three icon choices, or by manually expanding the size of the column, like in a spreadsheet).

In this example I selected the main code for "expert witnesses" and can see how often that code appears nearby a variety of theories related to the dog sniff. For example, the dataset includes 11 intersections of expert witnesses and false positives. It includes 13 examples of "cueing" with expert witnesses. This result is intuitive, as expert witnesses are often called by defendants to raise criticisms of a dog's reliability, and handler "cueing" the dog's behavior and the dog alerting to contraband where none is found ("false positives") are two of the major criticisms levied by experts.

You can do more with the code relations browser however. Double click on a data point (e.g., the 11 that is highlighted) and then minimize or move the code relations browser, so you can view retrieved segments from the main MAXQDA window.

This retrieves all 11 of the coded segments. You can focus in on individual documents. It displays the first code and then the code that is nearby. Notice that in document 31, false positives occurs in location 39-39, and expert witnesses appears two paragraphs later in location 41-41.

There are many ways you can use the code relations browser. The best way to learn it is through experimentation. Play with it and see how it helps you identify patterns and relationships, and how you can explore the underlying data that it reveals.

For more information visit the MAXQDA page on the code relations browser