Sign in or register
for additional privileges

Using NVivo: An Unofficial and Unauthorized Primer

Shalin Hai-Jew, Author

You appear to be using an older verion of Internet Explorer. For the best experience please upgrade your IE version or switch to a another web browser.

Creating Framework Matrices from Case Nodes and Theme Nodes

A “framework matrix” is a special type of matrix coding query related to case nodes crossed with coded nodes (themes) in a summary table.  The code frame is derived from a codebook and crosses raw data with the coding.   

A "framework method" of qualitative analysis involves the uses of codes that are defined in categories. These categories are co-defined by researchers.  This framework may be (1) pre-defined; (2) partially pre-defined and partially emergent, or (3) wholly emergent (extracted from the coding work).  The framework is comprised of cases in the row headers (stub) with themes in the column headers (banner).  This framework matrix type of query and resulting visualization enable the identification of code categorization and other data patterning (related to case nodes).  

This framework matrix enables understandings of meta-code and organizational structures among the codes.  A framework matrix helps users understand meta-organizational elements among the nodes.   


What are Case Nodes?  

Case nodes are special categories of nodes that belong to a certain category or type.  Case nodes have built-in classification sheets (aka "attribute tables") that describe certain features of the ego (individual) or entity (group) described within that node.  Generally one ego or one entity belongs to one case node.  An entity may be a person (ego) or an organization (entity), for example. The "ego" may be a file of interview responses, focus group responses, online survey responses, an oral history, a video transcript...focused around one person. The "entity" may be a file with collected textual data from a group (including all the prior types mentioned in the "ego" description).  The default types of case nodes are “person” and “organization” / "group" types.  [In the wild, some people are use "case nodes" to code the full contents of formal published academic articles and others, in order to enable their use in a framework matrix, which requires "case nodes" crossed with coded themes.]  






To get a sense of the pre-set attributes in the classifications sheets, go to the Explore Tab in the NVivo ribbon. 

Left-click on the Node Classifications Sheets.  In the Details View, the respective fields show the related attributes. 






The same may be done to view the Person node. 






There may be a myriad of other types of case nodes representing other entity types, such as places, things (like cultural artifacts), types of persons and roles, types of organizations, and so on.  These may be created by the researcher.  


What are Framework Matrices?  How Might these be Insightful for Research? 
How are Framework Matrices Structured?
 

A "framework matrix" is built on the case node structure.  So case nodes have to be created, and data has to be populated in the classification sheets.  (Note that these case nodes may be built off of default “person” or “organization” case nodes.  The classification sheets related to each of these case node types may be revised and edited.  Case nodes and their related classification sheets and related attributes may be created from scratch.)  There also have to be “theme nodes,” or nodes that indicate certain occurrences in the information.  Finally, there has to be coded data to each of the case nodes. For example, if the case nodes are “persons,” then their interviews or survey responses or oral histories have to be coded to their node or their node has to contain some or all of their response data. 

Framework matrices may be insightful by indicating patterns in the data that may not be as easily observable otherwise.  For example, were certain themes more prevalent among certain individuals (or types of individuals) than others?  Were there certain demographic patterns findable in the framework matrix?  What nuances may be found in the data? 

A framework matrix is structured as follows:  The columns are selected theme nodes (a minimum of at least one).  The rows are all of a specific type of case node (whether “person” or “organization” or other) and up to four attributes from each case node’s classification sheet; these cannot contain mixed case nodes.  The intersecting information cells indicate the respective attribute for the particular case nodes to enable a kind of summary view. 





How are Framework Matrices Created in NVivo? 

To create a framework matrix in NVivo, it is important to already have case nodes and the classification sheets for the respective case nodes populated with data.  (This is usually done manually.)  Next, in NVivo, go to the ribbon, and click on the Create tab.  Within the Sources group, click “Framework Matrix.” 

The New Framework Matrix window opens. 






The researcher should give the table a name and a description.  The framework matrix will be stored in the Node Matrices area. 

Next, the Rows and Columns should be selected. 






The framework matrix columns are defined also based on selected theme nodes.






Click “OK.”  The resulting matrix will contain coded text and links to source references resulting from the matrix.  The screenshot does not show this because synthetic data was used to create this as a “demo,” and nothing was coded per se to the case nodes. 


Finding Former Framework Matrices

To find formerly created Framework Matrices, click “Folders” in the Navigation View.  The “Framework Matrices” folder will show in the Folder area. 

New Search Folder

Another way to find the framework matrices is by going into the NVivo ribbon. Click the Create tab.  Go to “Search Folder” in the Collections group. 

Run a general search if the name of the framework matrix is easily recalled.  If more complex searching is desired, go to the “Search Criteria” tab. 





Specific types of sources may be searched for in the dropdown menu at the top. 






One of NVivo’s most powerful attributes (and largely unsung) is its database-like functionalities which enable the calling up and finding of pretty much any resource ingested and all coding and queries and such.  As an ecosystem, NVivo offers a robust and stable space for data exploration. 


Transcoding Files -> Cases 


To transcode a set of files in a folder or a space to cases (in the Nodes area), highlight the files, right-click the list, and Create As Cases.  Once this is done, the cases from the former files may be used to create framework matrices.  




Where possible, it helps to use the automation enabled by the software.  It helps to avoid makework! 


For More Information...

For more information, the QSR International site has a helpful page on this feature





An Afterword about Making Sense of Coding 

A matrix framework offers researchers a way to understand their coding in a table format...in relation to the informing raw files and sources of information (structured as case nodes).  This is a cross-case type of analysis because the themes apply across different cases.   

Other common ways to understand coding are as the following:  

  • codebooks [in an outline format, with main ideas (nodes) and sub-ideas (sub-nodes/child nodes/grandchild nodes/great grandchild nodes), in an alphabetical listing, and others]
  • word clouds of code books 
  • manually created weighted variables (combinations of nodes into larger meta-nodes) 
  • meta-analytic narrative descriptions of codebooks, and others 

Comment on this page
 

Discussion of "Creating Framework Matrices from Case Nodes and Theme Nodes"

Add your voice to this discussion.

Checking your signed in status ...

Previous page on path "Using NVivo" Cover, page 32 of 58 Next page on path