What is a within-subjects design?

In experiments, a different independent variable treatment or manipulation is used in each condition to assess whether there is a cause-and-effect relationship with a dependent variable.

In a within-subjects design, or a within-groups design, all participants take part in every condition. It’s the opposite of a between-subjects design, where each participant experiences only one condition.

A within-subjects design is also called a dependent groups or repeated measures design because researchers compare related measures from the same participants between different conditions.

All longitudinal studies use within-subjects designs to assess changes within the same individuals over time.

Using a within-subjects design

In a within-subjects design, all participants in the sample are exposed to the same treatments. The goal is to measure changes over time or changes resulting from different treatments for outcomes such as attitudes, learning, or performance.

Example: Within-subjects design for different treatments
You’re studying the effects of different messaging styles (your independent variable) on generosity (your dependent variable). Every participant is presented with 5 different short stories about climate change.

Each story uses a different tone and style. After each story, participants are asked how they feel about it and their willingness to donate to a related cause.

Other unrelated questions are also asked to make sure participants don’t guess the aim of the study. To test the effects of messaging styles on generosity, you compare the willingness to donate across conditions within subjects.

When comparing different treatments within subjects, you should randomize or counterbalance the order in which every condition is presented across the group of participants. This prevents the effects of earlier treatments from spilling over onto later ones.

Randomization means using many different possible sequences for treatments, while counterbalancing means using a limited number of sequences across the group.

Counterbalancing is sometimes more convenient for researchers because an even portion of the sample undergoes each sequence of conditions selected by researchers. Each treatment ideally appears equally often in each position (e.g., third) of the sequence. This helps balance out the effects of treatment sequence on the outcomes.

Example: Counterbalancing vs randomization of treatment order
In your study, you have 5 short stories (labelled A, B, C, D, E) as your treatments.

To counterbalance treatment order, you first decide on a few fixed sequences for these treatments: A-B-C-D-E, B-E-A-C-D, and D-A-B-E-C, etc. Then you divide your participants into sub-groups that each receive only one sequence of treatments.

To randomize treatment order, the order of the short stories is completely randomized between participants using a computer program. Every possible sequence can be presented to participants across the group, but in complete randomization, you can’t control how often each sequence is used in the participant group.

    In longitudinal studies, time is an independent variable. Because researchers can’t prevent the effects of time, longitudinal studies usually study correlations between time and other (dependent) variables.

    Example: Within-subjects design over time
    As a social researcher, you’re studying the effects of time (the independent variable) on perceptions of the pandemic and coronavirus (the dependent variable). You gathered a large sample of participants early in 2020 and have repeatedly sent them an online survey every two to three months since then.

    An important question in the Likert scale survey asks participants to rate their fear of getting COVID-19 on a scale from 1–7.

    To assess changes in perception, you compare differences in survey responses over time within subjects.

    Within-subjects versus between-subjects design

    The opposite of a within-subjects design is a between-subjects design, where each participant only experiences one condition, and different treatment groups are compared.

    Between-subjects designs usually have a control group (e.g., no treatment) and an experimental group, or multiple groups that differ on a variable (e.g., gender, ethnicity, test score etc). Researchers compare the outcomes of different groups with each other.

    In within-subjects designs, participants serve as their own control by providing baseline scores across different conditions.

    The word “within” means you’re comparing different conditions within the same group or individual, while the word “between” means that you’re comparing different conditions between groups.

    Within subjects design

    Between subjects design

    Example: Within-subjects design vs between-subjects design
    You’re planning to study whether the college course learning environment (your independent variable) affects test scores (your dependent variable). You can use either a between-subjects or a within-subjects design.

    If you use a between-subjects design, you would split your sample into two groups of participants:

    • a control group that takes a college course on campus,
    • an experimental group that takes the same college course online.

    You would administer the same test to all participants and compare test scores between the groups.

    If you use a within-subjects design, everyone in your sample would take part in every condition:

    • Half of the college course is administered on campus before a test.
    • Half of the college course is given online before a comparable test.

    You would randomize the order of the learning environment across the participants: some participants would first take the course on campus before switching to online learning, while the others would take the course online first before taking it in person. Then, you compare test scores within subjects between the two conditions.

    In factorial designs, two or more independent variables are tested at the same time. Every level of one independent variable is combined with each level of every other independent variable to create different conditions.

    In a mixed factorial design, one variable is altered between subjects and another is altered within subjects.

    Some longitudinal studies can be experimental when they use a mixed design to study two or more independent variables. If you can directly manipulate one of the independent variables, and participant assignment to conditions, you’re using an experimental approach.

    Example: Experimental longitudinal study
    As an education researcher, you’re investigating whether different teaching methods impact second language learning outcomes. Your participants are students in the 8th grade taking an introductory course of a second language.

    In a mixed design, you experimentally vary teaching methods between subjects, and you measure language knowledge at multiple time points within subjects.

    Each participant is randomly assigned to one of two groups:

    • A control group that receives standard teaching methods,
    • Another group that receives experimental teaching methods.

    All participants are tested before, midway and after taking the course, and their scores are statistically tested for differences across time and between groups.

    What is your plagiarism score?

    Compare your paper with over 60 billion web pages and 30 million publications.

    • Best plagiarism checker of 2020
    • Plagiarism report & percentage
    • Largest plagiarism database

    Scribbr Plagiarism Checker

    Pros and cons of a within-subjects design

    • Smaller sample

    Within-subjects designs help you detect causal or correlational relationships between variables with relatively small samples. It’s easier to recruit a sample for a within-subjects design than a between-subjects design because you need fewer participants. Every participant provides repeated measures, making the study more cost effective.

    • Removes effects of individual differences between conditions

    In a between-subjects design, different participants take part in each condition, so participant characteristics (e.g., intelligence or memory capacity) often vary between groups. This means it’s hard to say whether the outcomes are truly the result of the independent variable or individual differences between groups.

    In contrast, there are no variations in individual differences between conditions in a within-subjects design because the same individuals participate in all conditions. Participant characteristics are controlled for.

    • Statistically powerful

    A within-subjects design is more statistically powerful than a between-subjects design, because individual variation is removed. To achieve the same level of power, a between-subjects design often requires double the number of participants (or more) that a within-subjects design does.

    • Time-related effects

    There are many time-related threats to internal validity that only apply to within-subjects design because it’s hard to control the effects of time on the outcomes of the study.

    Some examples:

    • History: an unrelated event (e.g., a lockdown) may influence the outcomes.
    • Maturation: the natural physical or psychological changes (e.g., growth or aging) in the participants over time may cause the outcomes.
    • Subject attrition: more participants drop out at every subsequent step of the study, leaving you with a potentially biased sample at the end because only participants with strong motivations stay in the study.
    • Carryover effects

    Carryover effects are a broad category of internal validity threats that occur when an earlier treatment alters the outcomes of a later treatment.

    Some examples:

    • Practice effects (learning): familiarity with the study based on earlier conditions leads to better performance in later conditions.
    • Order effects: the placement of a condition in a number of conditions changes the outcomes—for example, participants pay less attention in the last condition because of boredom and fatigue.
    • Sequence effects: the interaction between conditions (based on their sequence) affects the outcomes; for instance, participants in an ad rating survey may compare later ads to earlier ones and base their decisions on the sequence of items.

    Randomization and counterbalancing of the order of conditions can help reduce carryover effects.

    Frequently asked questions about within-subjects designs

    What’s the difference between within-subjects and between-subjects designs?

    In a between-subjects design, every participant experiences only one condition, and researchers assess group differences between participants in various conditions.

    In a within-subjects design, each participant experiences all conditions, and researchers test the same participants repeatedly for differences between conditions.

    The word “between” means that you’re comparing different conditions between groups, while the word “within” means you’re comparing different conditions within the same group.

    What are the pros and cons of a within-subjects design?

    Within-subjects designs have many potential threats to internal validity, but they are also very statistically powerful.

    Advantages:

    • Only requires small samples,
    • Statistically powerful,
    • Removes the effects of individual differences on the outcomes.

    Disadvantages:

    • Internal validity threats reduce the likelihood of establishing a direct relationship between variables,
    • Time-related effects, such as growth, can influence the outcomes,
    • Carryover effects mean that the specific order of different treatments affect the outcomes.
    Can you use a between- and within-subjects design in the same study?

    Yes. Between-subjects and within-subjects designs can be combined in a single study when you have two or more independent variables (a factorial design). In a mixed factorial design, one variable is altered between subjects and another is altered within subjects.

    What is a factorial design?

    In a factorial design, multiple independent variables are tested.

    If you test two variables, each level of one independent variable is combined with each level of the other independent variable to create different conditions.

    Is this article helpful?
    Pritha Bhandari

    Pritha has an academic background in English, psychology and cognitive neuroscience. As an interdisciplinary researcher, she enjoys writing articles explaining tricky research concepts for students and academics.