It said I chose the wrong answer. In explanation, it doesn’t matter the extra data are added into both TRAIN and DEV sets. But the question only said the extra data are added to the DEV set. Am I correct or wrong?
{moderator edit - image of quiz question removed}
I think you are misinterpreting the explanation they are giving you. The point is that it is ok to have the training set distribution be different from the dev/test distribution. What is not ok is having the dev and test sets be from different distributions. So adding the new data to the training set would be the right thing to do. In other words, the correct answer is “True”.