"in academia, the issue of attribution and citation is very important" - true, but strictly speaking it's the attribution and citation of other research, not the research tools. After all, you don't find many research papers citing that they were typed on Microsoft Word with tables created in Microsoft Excel.

However, you've tried to convince them of this and they still want to go ahead, and it's your job to sort it out to their specifications. So...

The ideal solution would be to make the software so damn good that citing the use of your software in their publications is a good way for the external researchers to prove the integrity of their data analysis.

I might even suggest a reverse-psychology strategy with something along the lines of:

"You may only cite the use of {SOFTWARE} to guarantee the integrity of your data analysis if all data analysis has been done with {SOFTWARE}."

...and then provide a few suitable icons like this one:

http://validator.w3.org/images/valid_icons/valid-xhtml10