Your question seems to be very specific to your data. Please email to me directly at
spe...@andrew.cmu.edu.
First what do you call a "group"? Is it a group of children within one transcript file? Or do you have one child transcribed per file and a group is some collection of transcript files?
Second what do call "sum up"? Are you talking about combined listing of all answers from all children from multiple transcripts? Or are you talking about combined numerical result of all answers? In your previous email this number would be "7 instances followed by speaker "*CHI", of these". You can also use +b* option if you want to get a listing of all question codes with corresponding children answers on one output. This assumes that you have transcribed one child per one transcript file.
The KEYMAP will not work without you specifying some type of question code with +b option. The +u option will created a combined listing of all answers from all children in a selected group of transcript files that you run KEYMAP on. This assumes that you coded all children and adults across all transcript files with the same code *CHI for all children and the same code, like *ADU, for all adults.
It would be very helpful for me to see an example of your transcript files. Could you please email to me directly any two of your transcript files and may be also the output from "CLAN Output" window that you are getting when you run your KEYMAP command on those two files?