I helped to update the CSQE exam's question bank a while back. There were very few of us engaged with this effort at this time; I believe there was at least one subsequent re-examination since my involvement. I'll have some personal anecdotes about my experience at the end of this post, but the one reference I would specifically purchase and read is Roger Pressman's Software Engineering: A Practitioner's Approach. A familiarity with this text will cover everything of importance in the BoK with the possible exception of specific technologies. The Quality Council of Indiana always has a pretty good sense of the general structure of the ASQ exams, so you should probably pick up their reference as well.
Personal anecdotes about my personal experience working on the question bank follow.
I worked really hard on my parts of the question bank. The hardest part for me was to find published references that spoke to both the specific area of the BoK and the correct "Level of Cognition" (see the last page of the BoK). I couldn't keep referencing the same few resources, no matter how excellent, so I had to dig deep into other libraries. I viewed this as an opportunity to make sure that *I* wasn't writing questions that went into peculiar corners, but it was time consuming.
I did not feel that my peers who were working on other areas of the question bank (at the time) were taking things as seriously. We had to review the work done by others, and I ended up seeing a lot of "Wikipedia" type references. It was nearly always the case that my peers wrote questions at the cognition level of "Remember", and unfortunately the thing the student was being asked to remember was usually incredibly specific trivia. I challenged (and rewrote, or offered alternative) questions that I felt were highly unlikely to be answered correctly.
The last point about "answering correctly" is something I took very seriously; at the time ASQ question banks (supposedly) had the policy along the lines of "Given only the question, a master of the BoK ought to be able to provide the unique answer... without having to see the choices." This is what I mean by "trivia"... some questions were of the flavor "Which member of the 1973 Chicago Cubs led the team in triples?" along with at least two completely non-viable (wrong) distractors answers (in this example, "Marie Antoinette" and "Bilbo Baggins"). So there would be a 50-50 chance of getting it "correct" but this would be a horrible test of the BoK. Obviously for some levels of cognition the wrong answers need to be presented to give context to the (one) correct answer.
Another issue I had with what I was getting to review was that many of the peer-written questions involved negative stems, or non-direct stems. We were plainly directed to avoid such questions... for reasons that were both obvious and explained... yet I kept getting questions to review that looked as if they were written by someone who hadn't understood the assignment.
There was a final rough twist: the person at ASQ HQ who was responsible for the QB was ok, but I don't think that individual had any knowledge of the subject matter... so my feedback got passed back to the peers, who didn't exactly take the criticism well. I felt like I was being fair, but at some point my criticisms simply looked too harsh and I suspect the final QB ended up with a LOT of compromises... which may be why it had to be revisited. I know that many of my (ultimately ignored) criticisms that I felt could have been corrected were along the lines that the questions that were supposed to be at the "Evaluate" or "Analyze" level of cognition were written at the "Remember" level. As I wrote: there were so few of us on the project, I think if "I had gotten my way" I would have been responsible for something like 90% of the question bank, so to avoid that I think they just had to go with what they had.
