DATA Step, Macro, Functions and more

A Macro to Calculate Kappa Statistics for Categorizations by Multiple Raters

Reply
Contributor
Posts: 23

A Macro to Calculate Kappa Statistics for Categorizations by Multiple Raters

Greetings,

 

I would like to use A Macro to Calculate Kappa Statistics for Categorizations by Multiple Raters for a Delphi study. I need to assess the agreement on multi-category ratings by multiple raters.

 

I'm trying to use the MKAPPA Macro (http://www2.sas.com/proceedings/sugi30/155-30.pdf) but without success. I'm using SAS9.4, and i'm getting no output from the Macro.

 

Could anyone suggest me on what to do? Is there other Kappa Statistics Macros that I can use? 

 

Thanks!

 

D

Super User
Posts: 17,784

Re: A Macro to Calculate Kappa Statistics for Categorizations by Multiple Raters

When you copy and paste from a PDF you likely introduced some invisible characters in your code. Sometimes typing it out or using a editor to remove these symbols is enough to get it working. 

 

Why isn't it working for you? 

Contributor
Posts: 23

Re: A Macro to Calculate Kappa Statistics for Categorizations by Multiple Raters

Hi Reeza,

 

1. Thanks for your prompt answer. I did copy-pasted from PDF. I will try to write it down and then to run it again.

2. I've also tried to use now this macro (http://support.sas.com/kb/25/006.html). It worked for the example with 5 raters and 10 subjects, but did'nt work for 35 raters and 35 subjects. I've got the following error:

 

WARNING: Apparent symbolic reference M not resolved.
ERROR: A character operand was found in the %EVAL function or %IF condition where a numeric operand is required. The condition was:
&m=1 or &n=1
WARNING: Apparent symbolic reference N not resolved.
ERROR: The macro MAGREE will stop executing.

 

3. Do you have an idea what can be the problem?

 

Thanks!

 

Super User
Posts: 17,784

Re: A Macro to Calculate Kappa Statistics for Categorizations by Multiple Raters

I can't run anything now, but I suspect with 35 raters and subjects you don't have enough information to calculate a kappa - degrees of freedom and inability to invert a matrix idea. You need more subjects than taters. 

Ask a Question
Discussion stats
  • 3 replies
  • 114 views
  • 0 likes
  • 2 in conversation