Experimental and Quasi-Experimental Research Designs
Pre-experiments
Pre-experiments offer minimal control.
- Case Study with a Single Measurement: A stimulus or treatment is applied to a group, followed by measurement of one or more variables. This design lacks a true experiment’s rigor, with no independent variable manipulation or comparison group, hindering causality establishment and internal validity control.
- Pretest-Posttest Design with a Single Group: A pretest is administered, followed by the stimulus/treatment, and finally a posttest. While offering a baseline reference, it lacks manipulation and a comparison group, limiting scientific validity and causal inferences.
These designs are vulnerable in terms of control and internal validity, serving primarily as exploratory studies or trials for more controlled experiments. Results should be interpreted cautiously, requiring further investigation.
True Experiments
True experiments fulfill the requirements for control and internal validity, incorporating comparison groups (independent variable manipulation) and group equivalence.
- Posttest-Only Design with Control Group: Two groups are involved, one receiving the experimental treatment and the other serving as a control. The independent variable has two levels: presence and absence. Random assignment ensures group differences are solely due to the treatment. A posttest measures the dependent variable. The t-test is commonly used for analysis.
- Pretest-Posttest Design with Control Group: Pretests are administered before the treatment and posttests afterward. This allows for control and analysis of group changes. Random assignment and the control group mitigate internal validity threats.
Quasi-experiments
Quasi-experiments manipulate at least one independent variable but differ from true experiments in the initial group equivalence. Intact, pre-existing groups are used instead of random assignment.
Problems of Quasi-Experimental Designs: The lack of randomization can compromise internal and external validity. Researchers must establish group similarity by considering relevant variables.
Types of Quasi-Experimental Designs:
- Posttest Design with Intact Groups: Two pre-existing groups are used, one receiving treatment and the other serving as a control. Posttest comparisons assess treatment effects. Comparability issues can confound results.
- Pretest-Posttest Design with Intact Groups: Similar to the previous design, but with pretests to assess initial group equivalence and track changes.
Steps in an Experiment or Quasi-experiment
- Determine the number of independent and dependent variables.
- Choose the independent variable manipulation levels and translate them into treatments.
- Develop instruments to measure the dependent variable(s).
- Select a representative sample.
- Recruit participants and provide necessary information.
- Select the appropriate design based on the research question and objectives.
- Plan participant handling throughout the experiment.
- Randomly assign participants in true experiments or analyze intact groups in quasi-experiments.
- Administer pretests (if applicable), treatments, and posttests. Document the experiment’s progress to aid in result interpretation.
Hypotheses
A hypothesis is a proposed explanation for a problem, subject to empirical testing. It suggests relationships between elements or potential solutions. Validation involves confirmation (for universal hypotheses) or verification (for existential hypotheses).
Statistical Significance
Statistical significance (p) indicates the probability of observing a difference as extreme or more by chance. A p-value < 0.05 is traditionally accepted, meaning a 5% chance of the observed difference being due to chance. Lower p-values (< 0.01) are preferred for high-risk procedures or critical decisions.
Criteria for Hypothesis Formulation
- Use clear and specific terms for operational definitions.
- Ensure testability through empirical verification.
- Maintain objectivity, avoiding value judgments.
- Be specific about the problem and measurement indicators.
- Align with available resources and techniques.