An experimental study to investigate a research question requires experimental and control subjects. Three different designs, namely, within-subjects, between-subjects, and matched participants, can be used to investigate the question; does smiling improve the mood? In a between-participants design, each group of subjects receives a different treatment. In this research, the subjects will be divided into two conditions, namely, partial and full smile (two levels of the IV). They will then be subjected to the same treatment (a funny video) and their mood measured. Since the subjects in each level are different, any effect will result from the treatments.
Therefore, the between-participants design helps eliminate chance differences present in the study, giving a clear cause and effect relationship. However, the variability caused by the difference between partial and full smile conditions reduces statistical power. In a within-subjects design, the study subjects participated in all conditions (IV levels). In this experiment, the participants will each watch the funny video within the two levels (partial and full smile repeated for each subject). This design allows the researcher to control subject differences by comparing the scores of a participant in the two conditions. This implies that each participant can serve as both the experimental and control subject, resulting in a higher statistical power of significance.
However, confounds (not the independent variable) related to the order of the measure can increase or reduce the mood. In a matched-participants design, the researcher matches the participants on relevant variables. In this experiment, the researcher will select individuals with equivalent characteristics and variables such as age, economic status, and career and place them in the two levels. This approach minimizes the carryover effect, resulting in a stronger statistical power. However, subjects matching in relevant variables are difficult to come by, affecting data saturation.