Carl Needs to Empirically Validate Questions on a Measure for Depression

Carl needs to empirically validate questions on a measure for depression in his questionnaire. His best bet would be to develop a series of questions and then administer them to people known to suffer from depression (as diagnosed by a therapist) and to people known not to suffer from depression. He will know that his questions are valid if

A) the questions are not answered differently by the two groups.
B) the participants interpret the questions the same way.
C) the participants answer all the questions.
D) the questions are answered differently by the two groups.