Observational studies are essential when randomization is unethical or impractical. They may still be human-subjects research if the investigator obtains identifiable data or interacts with participants.