Between-subjects and within-subjects designs
In a between-subjects design, each person does one condition, and you compare groups of people. In a within-subjects design, each person does every condition, and you compare each person with themselves.
A mixed design does both: groups of people, each doing several conditions in balanced orders.
Which to use
| Design | Each person does | Strengths | Watch for |
|---|---|---|---|
| Between-subjects | One condition | No order effects or carryover, and shorter sessions | Needs more people, and differences between people add noise |
| Within-subjects | Every condition | Needs fewer people, and each person is compared with themselves | Order effects and carryover, so counterbalance the order |
| Mixed | Several conditions, in one group | Compares groups and conditions in one study | Both: assign groups evenly and counterbalance the orders |
Between-subjects in ConductExperience
- On the Groups and order tab, add the groups to compare, up to 12, each with a weight and a limit.
- Choose which groups get each section, so each group runs its own condition.
| Assignment | What it does |
|---|---|
| Balanced | Keeps the groups in proportion to their weights, from the people in each group now. |
| Random | Gives each person an equal chance of each group. |
| Weighted | Gives each person a chance of each group in proportion to its weight. |
Within-subjects and mixed designs in ConductExperience
- Put the conditions in one section and set its order to counterbalanced: each person gets one order of a balanced Latin square.
- Or shuffle the parts of the section for each person.
- For a mixed design, use groups and counterbalanced sections together.
- The Participants table and participants.csv show each person's group and order.
Repeated measures
A repeated measures design measures the same people more than once: in several conditions in one session, or at several times. A within-subjects design is a repeated measures design.
- Conditions in one session: put them in one section with a counterbalanced order, as above.
- Visits over time: add visits to the study, and each person gets a private return link. Each session records the visit's name and number and the days since the first visit.
- Check-ins through the day: short questionnaires on the person's own phone, with a row in the export for every check-in.
Related
Counterbalancing
Ways to order conditions, with a balanced Latin square generator.
Study builder
Groups, sections and parts in fixed, shuffled or counterbalanced orders.
Templates
Studies to start from, with their groups and orders set.
Online studies
One link for a panel or a mailing list.
Planning a study?
Start one now, or tell us what you want to measure.