Random assignment is often confused with random sampling. Random assignment is how you assign your sample to different groups or treatments in your experiment whereas random sampling refers to how your sample is drawn from a population. When you randomly assign participants you have an experimental design by definition, which is often referred to as a “true” or “controlled” experiment.
You can randomly assign participants by assigning numbers randomly (e.g., computer-generated random numbers) or as participants walk through the door. With any approach you choose, make sure to establish how you plan to randomly assign your sample prior to the intervention or experiment.
You can also randomly assign larger units than participants or students to treatment and control groups such as teachers, sections, or campuses. However, each unit of random assignment has their strengths and weaknesses. Randomly assigning students is generally the most statistically efficient and powerful research design. However, this approach requires that some students in a class receive the intervention while the other students do not (although a wait-list control could help with this limitation), and overall, this is usually not practical. Randomly assigning teachers or sections to treatment and control groups is a powerful approach because it controls for teacher or class-related factors. However, fairness is usually the most often heard concern with this approach (e.g., one section of students receive an instructional intervention while the other section of students do not but both sections are evaluated the same). Randomly assigning campuses or similar departments across universities can eliminate many threats to the study such as generalizability and internal validity. However, cost, issues of fairness, and sample size concerns are some of the major limitations with this approach.