Driver's comment is astute, and valid in my opinion, but somewhat kneejerk
regarding some cases he may not have considered.
Allow me to advocate for Drucker's position to test immediately after
training, from my own experience.
One purpose of such testing is to reinforce the key points immediately
after teaching/ training the trainee, and in the case of tests graded in
the session, review of the test gives the instructor yet another chance to
illustrate the concept or action again.
If a large number of participants were to improperly answer, the instructor
can remediate immediately, and then goes away with clues that the session
didn't cover the topic adequately for those students and how the training
can be improved for students with like-abilities in future sessions.
Repetition during the training session, after testing, is a good thing:
that may be the only time the student gets face time with the experts, and
gives students yet another chance to ask about related issues too, which
when answered relevantly will anchor the training soundly.