Experimental Designs: Within-Subjects Design

These flashcards cover key concepts and facts from the lecture on within-subjects experimental designs in behavioral sciences.

What are the general characteristics of a within-subjects design?

Uses a single group of participants who are tested in all treatment conditions; also referred to as a repeated-measures design.

What are time-related factors that can threaten internal validity in within-subjects experiments?

History, maturation, instrumentation, regression toward the mean, and order effects (practice, fatigue, carry-over effects).

How does time delay between treatments affect internal validity in a within-subjects experiment?

Time delay can increase the influence of time-related threats; switching to a between-subjects design may be advisable.

What is counterbalancing?

Changing the order of treatment conditions across participants to minimize time-related confounding.

What are the limitations of counterbalancing?

It does not eliminate order effects and may not sufficiently balance the effects across treatments.

What are the advantages of within-subjects designs?

Requires fewer participants, eliminates individual differences, and reduces variance, increasing the chance of detecting a treatment effect.

What are the disadvantages of within-subjects designs?

Time-related factors can influence scores, and participant attrition can occur.

Define a matched-subject design. How does it attempt to avoid disadvantages of other designs?

Uses separate groups for each treatment, matching individuals one-to-one across groups on relevant variables to combine benefits of within- and between-subjects designs.

What are order effects in the context of within-subjects designs?

Changes in scores resulting from the order in which treatments are administered, which can confound study results.

What statistical techniques are used for analyzing data in two-treatment designs within-subjects?

Repeated-measures t-test or single-factor ANOVA (repeated measures); Wilcoxon signed-ranks test for ordinal data.