Domain Testing: Divide and Conquer - Testing Education

points. The material presented in training was mostly about 1-D spaces. So,
in many respects, this problem was a good test in that it required
performance at level 6 of Bloom's taxonomy. Unfortunately, none of the
subjects made this conceptual leap. They tried to treat the problem as if it
were two independent 1-D spaces. This led them to propose tests that I
thought were not very powerful. (Appendix Z: Performance Tests’
Evaluation Report by Pat McGee, pp. 1-2)
7.01.03 Final Remarks
Kaner (2003) made some final comments about my experiment:
Padmanabhan provided her students with detailed procedural
documentation and checklists. From the consistency of the students’ results,
I infer that students followed the procedural documentation rather than
relying on their own analyses. It’s interesting to note that, assuming they are
following the procedures most applicable to the task at hand, the students
not only include what the procedures appear to tell them to include, but they
also miss the issues not addressed by the procedures. In other words, the
students aren’t doing a risk analysis or a dimensional analysis; they are
following a detailed set of instructions for how to produce something that
looks like a risk analysis or dimensional analysis.
I think we’ll find Padmanabhan’s materials (especially the practice
exercises) useful for bringing students to a common procedural baseline.
However, we have more work to do to bring students’ understanding of
domain testing up to Bloom’s Level 6 (Evaluation) or Gagne’s (1996)
development of cognitive strategies. (p. 4-5)
Bach made the following final comments:
I expect more insight, product knowledge, and imagination from a serious
tester who had more than a few months of experience working for a
company that cared about doing good testing. So, I would not say that this
116