a type of analytical observational stuys. see the outcome then give treatment. Two groups: one with identified outcomes and one without. Start with the outcome of interest. Then get two groups of people within that outcome. Ex: people who tried to quit smoking, one who did successfull,y and one who didn't. see what treatment worked on the group that quit. Sometimes in these studie,s it is unethical to not give treatment and sometimes it is unethical to give "treatment". Assess whether there were differences in treatment exposure retrospectively using chart reviews (medical records), self-reports (ask what they did), and interviews (asking more in-depth questions). Good for rare outcomes and to save time and money since there are no long-term follow-ups. Limitations: difficult to select appropriate control groups because you want the groups to match, recall bias (make people think back on what they did), and you can not see the prevalence of the outcome, only the odds of experiencing both because ther'es no control group.