Ask yourself if your central concern is establishing a causal linkage between the program, intervention or treatment under examination and the outcome or effect you are measuring. If the answer to this question is yes, you should select a research design that uses random assignment of study subjects to treatment and control groups. For example, a clinical trial of a new blood pressure medication would randomly assign study subjects to the treatment group, whose members would receive the new medication, and the control group. Rhonda BeLue of Cornell University's Department of Policy Analysis and Management, points out that the main benefit of random assignment is probabilistic equivalence between the two groups. That is, any differences between the two groups are due to chance. Consequently, any differences between the two groups after the experiment can be attributed to the treatment itself rather than pre-test differences. If establishing causation is not your main concern, you should select a quasi-experimental research design. These research designs to do not use random assignment of subjects.
Examine the budget and setting of your study to determine if a randomized experiment is feasible. Random assignment may not be feasible in some education or social science studies. For example, randomization may not be feasible in a study of a new reading curriculum that compares two classrooms because students have already been assigned to classrooms.
Assess the types of quasi-experimental research designs, considering such factors as your research questions, the subjects of your study and the types of data you will collect. These factors will help you select the optimal quasi-experimental research design, each of which offers varying levels of scientific rigor. For example, if you are studying a new tutoring program and comparing two existing groups of students, you should choose a quasi-experimental research design that compares a non-randomized treatment and control group, conducts a pre-test to control for existing differences between the two groups, and conducts post-intervention measures to gauge the impact of the tutoring program. If your study is more descriptive in nature and focuses on only a single group of subjects, you could employ a case study or single-group pre- and post-test design that measures the group on the outcome of interest before and after the intervention or treatment.