Agree with you. The execution could've been much better. To start with, it always make sense to do a validated questionnaire experiment than come up with a new survey like we did.
Talking about Milgram and similar experiments, though they are known to help achieve unbiased observations about human behavior, it actually doesn't rule out the validity of a well executed questionnaire experiment. Good questionnaires take care of authenticity using a lie score that tries to detect if the subject is trying to make himself / herself look good with the answers.
