Answer to Question 1
Experiment can't be done, because the independent variable, parents' divorce, can't be manipulated. Survey ask a sample of parents and/or students a variety of questions, mostly closed-ended, about the family (including marital status) and school achievement (grades, test scores). Compare academic achievement of students from divorced or intact families. Observation not practical. How would you observe whether or not parents are divorced, and how would you observe grades? Interviews ask questions similar to those in surveys, but conduct the research face-to-face with open-ended questions.
Answer to Question 2
research conducted in the real world in which the independent variable is manipulated. Example: send couples or people of different races, ethnicities, or gender to stores or to answer a job ad to see if they are treated differently.