!Converted with LaTeX2HTML 95 (Thu Jan 19 1995) EI by Nikos Drakos (email@example.com), CBLU, University of Leeds >
Next: Other Requirements
Up: Experimental Design
Previous: Experimental Design
The participants in this experiment will have to have some experience
with reading requirements documents. PBR will be evaluated against
the participants' usual technique.
Any number of participants may be used, though fewer than 12 subjects
may represent too few data points to result in statistically
significant results. We assume 18 participants (IDs will be 01-18)
for our discussion here, though the design may be easily scaled up or
- If the experimental
environment is in industry, the usual technique will be the technique
which is normally applied within the development organization.
- If the
experiment is run in a classroom setting, we assume students should
already have been taught about requirements reviews and will be
applying some technique which they have just been taught. However, in
this case it is crucial that the students get enough training to
minimize any learning effects.
We recommend assigning reviewer IDs to participants in the order they
arrive, using the scheme illustrated in the following chart. This
assures that the reviewers in each group will be randomized, and that
the number of reviewers using each perspective and in each thread will
be as even as possible.