Research Design and Sampling for Comparing Student Groups

Quantitative Algorithm:

According to Paul D. Leedy & Jeanne Ormrod, Practical Research: Planning and Design, Pearson Education, 10th Edition (2015) [invalid URL removed], one algorithm used to identify research designs considers the level of manipulation of the independent variable and the presence of a control group.

Research Design:

The appropriate research design for comparing two groups of students (school team vs. no school team) is a quasi-experimental design specifically a non-equivalent groups design.

Explanation:

• In a true experiment, the researcher randomly assigns participants to groups (intervention and control). Here, we cannot randomly assign students to be on a school team or not. This lack of random assignment removes the ability to establish causation between being on a school team and the outcome variable (e.g., academic performance).
• A non-equivalent groups design acknowledges the pre-existing groups and compares them as they are.

Sampling Method:

To achieve the highest level of research design possible within the constraints (non-equivalent groups), we should strive for a sampling method that minimizes selection bias.

Selected Method: Propensity Score Matching

Propensity score matching is a statistical technique used in non-experimental research to create comparable groups. Here’s why it’s a good fit:

• It attempts to statistically balance the two pre-existing groups (school team vs. no school team) on factors that might influence the outcome variable (e.g., prior academic performance, socioeconomic background).
• By statistically matching students on these characteristics, we can create groups that are more similar, allowing for a more fair comparison of the impact of being on a school team.

Conclusion:

While a true experiment with random assignment would be ideal, due to the nature of pre-existing student groups, a non-equivalent groups design with propensity score matching offers a more robust approach to minimize bias and strengthen the internal validity of the research.