QAT Exam feedback

General I think four questions in 1:30 is too much (Although I did keep them small) and will change the duration for next year.

I was quite tough with marking as the coursework marks were very high.

Many answers started by repeating the question  - this is a waste of time and gains no marks.

The number of people who do not follow the instructions on the exam is staggering.  Just under 50% did not start each question on a new page.  Two people didn’t put the questions answered on the front cover.  One person put the answers in ascending numerical order.  There were a couple of other minor breaking of the instructions.

Question 1.

  1. Poorly done  - should be easy
  2. Tricky, answer expected to be a bit “waffly”, but many completely missed the point.

Question 2.

Straight forward and simple question and answer

  1. for COTS customer is marketing Dept etc
  2. for internal customer is initiating dept (logistics, management, finance etc)
  3. formal has lots of benefits (otherwise would be informal, even with external customers) but might stifle creativity

marks easily obtained, yet most waffled, missed point and got few marks.  Very low spread of marks.

Question 3.

This is a “banker” question – easy to pick up ok marks but hard to get v good ones.  Some misreading of question and lots of repeating question back at me at start of answer.

Question 4.

Lots of good minor stuff about understanding communication, replacing team members etc
Few hit crux  of answer - *quality* risks of non-compliance is an increase in rate of errors end hence cost, then explain why.  V. low average for this question.

Question 5.


Simple knowledge question, you should know whether you got the marks on this one!  Very short answer needed (full marks in 10 mins if you know your stuff?)  Least attempted question with best average mark.

Question 6.

  1. many did not know what “big bang” testing was.   Many ok answers, but didn’t emphasise location and propagation of errors as main problem.
  2.  i) most didn’t explain that path coverage repeats testing of the same lines

ii) many failed to say why large numbers of tests are impractical (cost/resources/time), many didn’t pick on the *most test apps* ie when does it become practical (eg safety critical & modules with highest damage potential)