Within-subjects design can be useful in experimental research, especially when there is a limited sample size or individual differences between participants must be controlled.
Definition: Within-subjects design
A within-subjects design, also known as one dependent group, is a research design in which each participant serves as their own control and is exposed to all levels of the independent variable. This means that participants are tested in all study conditions rather than randomly assigned to only one condition.
The within-subjects design is the equivalent of one between-subjects design (individual participants are offered a singular condition). The word “within” in within-subjects design refers to the fact that the comparisons made are within the same subject or participant rather than between different groups of participants.
It is also referred to as repeated measures design or crossover design since researchers compare aspects of the same participants in varied conditions.1
Within-subjects design: How it works
In a within-subjects design, all participants are drawn from a single sample and are exposed to identical conditions to measure changes over time or due to different conditions (also known as “treatments”). This design can be used to examine a variety of variables, such as opinions or performance.
When studying various conditions using the same group of participants, it is important to randomize the order in which the conditions are presented (known as counterbalancing).
This helps to eliminate the possibility that the effects of previously presented conditions may impact the results of the later conditions. In other words, the order in which the conditions are presented is not a factor that affects the results. Randomization involves using a wide range of different sequences for presenting the conditions, whereas counterbalancing involves only a few sequences.3
Within-subjects design vs. between-subjects design
A between-subjects design is the opposite of a within-subjects design, as each participant is subjected to only one condition, and the group means are then compared.
|Within-subjects design||Between-subjects design|
|participants are tested in all study conditions||each participant is subjected to only one condition|
|In contrast, within-subjects designs simultaneously involve the participants in a control group as a guideline measurement is taken. This baseline measurement is then observed against other conditions.||Typically, between-subjects designs comprise a control group, which does not receive any manipulation, and singular or multiple experimental groups that vary in one particular variable. These may include gender, ethnicity, or key scores. Researchers can collate the results of the different groups.|
“Within” in this context refers to comparing different conditions in the same group or individual, while “between” involves comparing conditions across several individuals or groups.
You can apply both within-subject and between-subject designs in some cases, for instance, if you’re investigating whether e-commerce sites (independent variable) influence levels of consumer spending (dependent variable).
You would shop in both ways with all participants and compare the spending of the control group with that of the experimental group.
To compare the spending habits of participants between online and in-person shopping experiences, it’s important to randomize the order in which participants shop. This means some participants would start shopping on an e-commerce site, while others would start at a physical store.
Comparing the spending of the same participants in both conditions makes it possible to assess whether the shopping experience affects spending habits.
Counterbalancing can be a more convenient option for researchers as every possible order in which the conditions are presented occurs equally often. For instance, if four participants are presented with the order XYZ, four in the order ZXY, and four who are offered XZY.4
Within-subjects design: Pros and Cons
The within-subjects design has advantages and disadvantages, which can affect the validity and reliability of the results.
Within-subjects designs are suitable for establishing correlations or causality between variables, even with small sample sizes. However, it could also compromise internal validity.
A within-subjects design typically requires fewer participants than a between-subjects design, making it easier to find enough participants. Each participant is subjected to multiple tests under different conditions, making the design more efficient.
Identifying participant differences
In a between-subjects design, the participants are only exposed to one condition, resulting in variations in individual properties such as income level and race. Determining whether the differences found were caused by the alteration or participant characteristics is challenging.
Greater statistical power
A within-subjects design has higher statistical power due to the absence of individual variation. It requires half the number of participants of a between-subjects design to attain the same level of statistical power.
Where the order in which participants experience the conditions affects their performance. For example, participants may improve their performance over time due to practice effects or become fatigued due to the length of the study, which can affect the results.
Where the effects of one condition carry over into subsequent conditions. This can happen when a participant’s performance in one condition is influenced by their experience in the previous condition, which may confound the results.
A within-subjects design is a research design in which each participant is exposed to all levels of the independent variable, allowing for a direct comparison of the effects of each level.
In a within-subjects design, the same participants are tested under all conditions, allowing researchers to control for participant properties.
The within-subjects design allows for higher statistical power and requires fewer participants, but it is vulnerable to carryover and order effects, which can compromise the internal validity of the study.
Researchers can control for order effects by counterbalancing the order in which the different levels of the independent variable are presented to participants.
1 Kelly, Rose. “What is Repeated Measures Design? – Definition & Example.” Study.com. Accessed February 24, 2023. https://study.com/academy/lesson/what-is-repeated-measures-design-definition-example.html.
2 Cherry, Kendra. “Within Subjects Design and Between Subjects Design.” Verywell Mind. Accessed February 24, 2023. https://www.verywellmind.com/what-is-a-within-subjects-design-2796014.
3 Alferes, Valentim R. “Within-Subjects Designs Randomization.” SageResearchMethods. Accessed February 22, 2023. https://methods.sagepub.com/book/methods-of-randomization-in-experimental-design/n3.xml.
4 Cutler, C. “Experimental Design.” March 5, 2020. https://opentext.wsu.edu/carriecuttler/chapter/experimental-design/.