Experimental Designs: Structures of Research in Education

Research designs are the structure of research. It is the foundation of the study. Research design consists of observations, measures, treatments, programs, groups, assignment to groups, and time. Among the designs to consider are the True experimental design, Quasi-Experimental design, and the Non-Experimental design. Each design offers its own advantage and disadvantage in regards to feasibility, validity, and reliability. This will be the basis for this document. This report will explain each research design and show its vulnerabilities.

Quasi-Experimental design probably has the most disadvantages. The reason why it has the most disadvantages is because it lacks a very important element in research: random assignment. This type of research requires the use of a pretest and posttest. This could be used to exempt the use of standardized tests in children. The design could test the level of anxiety prior to testing and after testing. The results could be classified as on equivalent groups design. The disadvantage of this research design is the analysis of statistical data among the groups. Random assignment might offer better results. It might offer more validity. The advantage of this type of research design is that it explores three of the four basic elements in most research designs. For example, the use of Time to determine a causal relationship between testing and anxiety would benefit from a pretest and posttest. Additionally, the observations of groups or individuals during the testing process could serve as vital research.

A True-experimental design is different from a Quasi-Experimental design. It is different because random assignment is not the same as random selection. When random assignment is used, it can be considered a true experiment. Once the random issue is established, it is then questioned whether there is a control group or multiplicity of measures. If the answer is yes, then the previous mentioned experiment takes precedent (Quasi-experiment) and if no, then it is a non-experiment research design. A True-experiment can reduce threats by examining its use of argument, measurement, design, analysis, and preventative action. The disadvantage of this type of research design is the researchers willingness to sacrifice external validity in order to achieve better internal validity. The researcher develops and procreates a surreal situation in order to find a causal relationship. This relationship would justify high internal validity. However, the data received through research becomes very “general” and “basic” and warrants further study.

For the research needed in this study of test anxiety, the use of Non-Experimental design would probably offer the least internal validity and causal assessment. The reason behind this notion is that the design is very one sided. The design consists of only one observation. Although it is a common form and mode for research, it would not benefit the research of standardized tests and anxiety. If the research question were a very descriptive one then the Non-Experimental design would serve its purpose well. However, using its design to assess the cause-effect relationship between standardized tests and anxiety would not be a good method. Of the three research designs mentioned in this document, the Non-Experimental design would be the least effective.

In all of the research designs there are ways to decrease the amount of errors in the research. Expanding research can provide better data. The data would represent internal validity and would enable the researcher to provide evidence for his methodology. Expanding research over time by including additional observations of children taking standardized tests is just one way of developing better research strategies. Expanding across programs would also serve the same purpose in the research. The researcher could possibly observe two different types of standardized tests and see if the children experience the similar attributes of anxiety. Another way is to expand among observations. For example, the researcher could observe children who did not sleep well the night before a standardized test to see if the lack of sleep increased the anxiety? Different measurement approaches would definitely provide different data, but it would require the researcher to rely on multiple measures as opposed to the single strategy. Lastly, another way to decrease the threat to validity in these research designs is to expand research across groups. The use of expanded groups would lessen the discussion among students who do experience test anxiety to those that do not. The purpose is attempting not to identify through inter-group communication, who is part of the nonequivalent group, and who is part of the research design.

Leave a Reply

Your email address will not be published. Required fields are marked *


8 − = five