 |
 |
Temple University Center for Research
in Human Development and Education |
|
|
Application of appropriate research design, methods, and techniques is the important first step. To begin, one must decide whether to conduct quantitative or qualitative research. Quantitative research is more objective in nature, with emphasis on empirical measurement, large sample groups, and establishing correlations or cause-and-effect relationships through such means as standardized tests and surveys/questionnaires. Qualitative research, in contrast, is much more subjective, with a focus on obtaining richly detailed anecdotal information through such techniques as case study interviews and observations, even if the sample group consists of just one individual (Onwuegbuzie & Daniel, 2003). For the purposes of scientifically based research (and, by extension, NCLB), one should begin with quantitative research and provide support through qualitative methods.
Next, one must identify which quantitative research design is most appropriate: nonexperimental, true experiment (randomized field trial), or quasi-experimental. The nonexperimental design is the simplest type in that the researcher just observes what occurs naturally without intervening in any way. An example of a nonexperimental design in education would be to correlate students’
homework assignment completion with their achievement scores.
|
|
| |
At the other end of the spectrum lies the true experiment (or randomized field trial) in which participants are randomly assigned to one of two conditions: treatment (intervention) or control (no intervention). The goal is to determine whether the treatment has a measurable impact on the outcome of interest, which is best determined through the process of random assignment. Random assignment is a powerful research tool because it gives participants an equal, unbiased chance of being put into either the treatment or control group, and because it reduces the likelihood that participants’ personal characteristics and other “background noise” will interfere with the ability to detect the true impact of the treatment/intervention. Indeed, random assignment is what makes the true experiment the “gold standard” of scientifically based research and therefore the standard all research studies should strive to meet whenever possible (Darlington, 1990; Shadish, Cook, & Campbell, 2002).
Random assignment is sometimes confused with random sampling, so clarification is in order. Random sampling requires access to the entire population of interest in order to select participants at random from that population for a particular study (Darlington, 1990). To illustrate, if a school district wants to conduct a reading intervention program study of its entire fourth-grade student population, it could put the names of every fourth-grade student into a large hat, then draw out 100 names to participate in the study—hence, a random sample. That random sample of 100 fourth-grade students would then be put into another hat, with 50 names drawn out for the treatment/intervention group and 50 names drawn out for the control/nonintervention group—hence, random assignment.
|
| |
Previous Page
1 2
3
4
Next
|
|
|
Copyright 2001 © MARTEC
|
|
