Collaboration in EAP language testing

Something that we've been talking about in our CALL group (last week, to be precise, but I am just getting this post up now) is the issue of using technology in language assessment.  From an administrative (and even pedagogical) sense, there are a lot of good practical reasons for using computer-based testing:

1. Better record of student production
2. More consistent administration (possibly)
3. Possibility of using adaptive instruments
4. For writing-based tasks, using a computer is likely more realistic than using paper and pen.

But there are some challenges that Dooey (2008) outlines as challenges to using computers in testing, primarily that this may provide unequal footing for learners who are not familiar with technology, meaning that they are not just being tested on their language abilities, but also possibly on their technological literacy.  In general, in our class discussion, there was some agreement that this can still be an issues (and there were several examples given by teachers in both ESL and EFL contexts).

On the other hand, there were also several student who put forth the argument in the 4th item above: that having pen and paper writing tests would actually be a less authentic task for EAP students than a computer-based writing task.  This was not as clear when we were talking about speaking assessments, however; while there is value in having a recorded version of the test-takers production for evaluation purposes, it seems that more meaningful and realistic speaking tasks should (ideally) involve interaction with another speaker.  Interaction in speaking assessment integrates both speaking and listening and in part tests the unplanned speaking abilities of the learner.

I wonder if interaction is increasingly going to become something we are interested in for writing as well, as there are many important areas of higher education (especially undergraduate) where writing is collaborative.