Jump to content
[[Template core/global/global/lkeyWarning is throwing an error. This theme may be out of date. Run the support tool in the AdminCP to restore the default theme.]]

Calculating Kripp's Alpha

Recommended Posts

Hello Everyone,


I am working with NVivo 10 on Windows to examine policies in archival government texts. I am interested in

determining and quantifying the focus of the policies. The texts are scanned pdfs (n=40) and each pdf may

contain as few as 8 policies to over 110 policies. In addition, the coding has been largely executed by region.

For my analysis, using 7 Parent nodes and over100 child nodes, I analyze each policy for the presence of any

of the children.


At this point, I need to employ a second User in order to obtain an inter-rater statistic. In my social science field,

Percent Agreement is not frowned on and researchers are moving away from Cohen's alpha. A journal referee

is pointing to Krippendorff's alpha.


Because books are the units of analysis, the ideal solution would be a matrix query with the books as rows,

the users as columns, filter the results by child nodes, then export into excel for the Kripp analysis. Having both

users in the same project would allow for reconciliation of coding disagreements.


Is this possible as I understand the NVivo counts all Users in coding frequency? If not, then how should I proceed

with introducing my second User and obtaining a subsequent Kripp analysis.




Share this post

Link to post
Share on other sites

Hi Darrel,


You could run your matrix coding query in this way and then export the resulting node matrix to run your analysis in Excel. Please keep in mind though that the results of the matrix coding query by default show the number of coding references which may not be ideal when running your analysis.


You can change the cell content in your node matrix to show other information such as words coded which might be more relevant to you. Instructions on how to do this can be found at the following link http://help-nv10.qsrinternational.com/desktop/procedures/work_with_the_content_of_a_node_matrix.htm?


Another approach to getting the information you need might be to run a coding comparison query within NVivo, and then export the results to Excel and continue with your analysis using this Excel spreadsheet. You might find the following spreadsheet helpful. It contains some information on how coding comparison queries are run in NVivo and includes examples of how to use this data to perform further analysis. http://redirect.qsrinternational.com/examples-coding-comparison-nv10-en.htm


Regards, Heather

Share this post

Link to post
Share on other sites

Have found the answer.


After the coding the entire project, select the NVivo Query tab and complete the following Matrix Coding export tasks.

1.      For any single item (node), or a group of items (nodes):

a.      The rows will contain the individual cases,

b.      The columns will contain the item(s)/node(s) in question. However, the column format requires the following steps

                                                    i.     First, select the item(s)/node(s),

                                                   ii.     Second, select the coders by clicking on “By Any Selected Users”,

                                                 iii.     Third, click “Select” to choose the coders by their initials,

                                                 iv.     Lastly, click “Add to List”

1.      This will add each item by each coder as an individual column.

c.      Click Apply to save the above procedures

d.      Click Run to produce the Matrix Coding table,

e.      Export the data into Excel,

f.       For SPSS,

                                                    i.     format the data as per the SPSS procedures

g.      For a web based utility, you will need to save each item in as single CSV file following the procedures as outlined by (Freelon, D. 2010; 2013).




Freelon, D. (2010). ReCal: Intercoder reliability calculation as a web service. International Journal of Internet Science, 5(1), 20-33.

Freelon, D. (2013). ReCal OIR: Ordinal, interval, and ratio intercoder reliability as a web service. International Journal of Internet Science, 8(1), 10-16.


Share this post

Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now