 |
 |
Temple University Center for Research
in Human Development and Education |
|
|
Somewhere between the nonexperimental and true experimental designs lies the quasi-experimental design. This design involves many of the same features as the true experiment (e.g., use of treatment and nontreatment groups to make comparisons) but lacks the full experimental control of a true experiment (meaning random assignment of participants to either group may not be possible) (Cook & Campbell, 1979; Shadish, Cook, & Campbell, 2002). Although the quasi-experimental design is more of a “silver standard” regarding its reliability and validity, it is commonly used in educational settings when conducting a true experiment may not be practical, cost-effective, or ethical due to the complex human factors present in any school. An example of a quasi-experimental design in education would be to test the impact of a math software program on eighth-grade students by selecting one eighth-grade classroom to receive the math software along with regular instruction (the treatment/intervention group) and one eighth-grade classroom that would not receive anything beyond regular instruction (the comparison group). Note that the term “control group” is not used with the quasi-experimental design because unless participants have been randomly assigned to either group, true “control” of the study situation is not possible.
Once the research design has been selected, the second component of scientifically based research comes into
|
|
| |
play—namely, that research must be based on empirical measurement or observation. First, one must define study variables as follows: (a) dependent variables (i.e., what is my outcome of interest?); (b) independent or predictor variables (i.e., what factors—such as race/ethnicity, age, gender, or socioeconomic status—might influence my outcome?); and (c) control/confounding variables (i.e., what variables, such as socioeconomic status, might get in the way of seeing my outcome of interest?) (Darlington, 1990). One must also decide how best to measure the variables of interest. Commonly used methods in education include surveys/questionnaires (which may be biased by their self-report nature, so caution should be exercised), standardized tests, observations, interviews, focus groups, and individualized formats based on the research goals and objectives (e.g., student portfolios, homework assignments).
The third component is that research must be both replicable and applicable. Specifically, one must determine if the research can be reproduced in a straightforward, cost-effective way and if it is worth the cost to do so in terms of time, human resources, and financial resources. Additionally, one must identify other groups that might be affected by the research beyond the actual study participants; in an educational context, such groups could include principals, teachers, students, parents, central office staff, school board members, and community organizations.
Previous Page
1
2 3
4
Next |
