With the passage of the No Child Left Behind Act (NCLB; 2002), attention has focused on the need for evidence-based educational research, particularly educational policies and interventions that rest on what NCLB refers to as “scientifi - cally based research.” In practice, this focus on scientifi cally based educational research has translated into a preference for research studies based on the principles of randomized experimental designs. Specifi cally, Part A, Section 9101 of the No Child Left Behind Act, under the defi nition “Scientifi cally Based Research” states:

Before proceeding, it may be useful to provide defi - nitions of randomized experimental designs, quasi-experimental designs, and observational/non-experimental studies. Following Shadish, Cook, and Campbell (2002), a randomized experiment is “an experiment in which units are assigned to receive the treatment or alternative condition by a random process such as the toss of a coin or a table of random numbers” (p. 12). When a randomized experiment is properly conducted, the groups formed by random assignment are probabilistically equivalent to each other. A quasi-experiment is “an experiment in which units are not assigned to conditions randomly” (Shadish et al., 2002, p. 12). In quasi-experimental studies, units self-select into

treatment. Such self-selection can result in pre-treatment differences that may confound any causal conclusions drawn regarding the effi cacy of the treatment. Careful use of various statistical controls can provide a degree of pre-treatment equivalence that can increase confi dence in causal conclusions. Finally, a correlational study is “usually synonymous with non-experimental or observational study: a study that simply observes the size and direction of a relationship among variables” (Shadish et al., 2002, p. 12).