Experimental designs

Experimental design

The way participants are allocated to the conditions in an experiment.

Independent groups design

Participants take part in only one condition of the experiment.

Strength of independent groups

No order effects such as practice or fatigue since each participant only completes one condition.

Limitation of independent groups

Participant variables may affect results because groups differ in abilities or characteristics.

Reducing participant variables in independent groups

Random allocation is used to distribute differences evenly.

Repeated measures design

Participants take part in all conditions of the experiment.

Strength of repeated measures

No participant variables because the same individuals take part in each condition.

Limitation of repeated measures

Order effects such as practice, boredom or fatigue may impact performance.

Order effects

Changes in performance caused by the order in which conditions are completed.

Practice effect

Improvement in performance due to experience from previous condition.

Fatigue effect

Worse performance due to tiredness when performing later conditions.

Counterbalancing

A method to control order effects by varying the order in which participants complete conditions.

ABBA counterbalancing

Participants do conditions in one order then reverse order to cancel out practice/fatigue.

Latin square

A more complex counterbalancing technique ensuring each condition appears in each position equally.

Matched pairs design

Participants are matched on key variables relevant to the experiment (e.g. IQ, age).

Strength of matched pairs

Reduces participant variables because groups are similar on important characteristics.

Limitation of matched pairs

Time-consuming and difficult to find closely matched participants.

Matched pairs allocation

Members of each matched pair are placed in different conditions.

When to use independent groups

When avoiding order effects is important or repeated measures is impractical.

When to use repeated measures

When participant variables must be controlled and fewer participants are available.

When to use matched pairs

When individual differences could confound results but repeated measures is not possible.

Participant variables

Differences between participants such as motivation, intelligence or personality.

Controlling participant variables

Random allocation or matched pairs design can reduce their impact.

Demand characteristics in repeated measures

Increased risk because participants may guess the aim after completing both conditions.

Demand characteristics in independent groups

Less likely because participants only do one task.

Design suitability

Choice depends on study aims, practical constraints and control of variables.

Independent groups disadvantage

Requires more participants; harder to recruit large samples.

Repeated measures advantage

Economical design as fewer participants are needed.

Repeated measures disadvantage

High likelihood of demand characteristics influencing behaviour.

Matched pairs advantage

No order effects and fewer participant differences than independent groups.

Matched pairs disadvantage

Matching process may be subjective and imperfect.

Random allocation

A method of assigning participants to conditions using chance to reduce bias.

Random allocation purpose

Ensures participant variables are evenly distributed across conditions.

Counterbalancing purpose

Ensures practice or fatigue effects are balanced across conditions, not favouring one condition.

Types of counterbalancing

ABBA, Latin square, or randomised order.

Order effects control

Only needed in repeated measures design.

Participant variables control

Indirectly controlled in repeated measures; directly controlled in matched pairs.

Control group

A group not receiving the IV manipulation, used for comparison.

Experimental group

A group receiving the IV manipulation.

Extraneous variables and design

Design must minimise variables that could influence the DV.

Experimental design validity

Validity increases when confounding variables are minimised through appropriate design choice.

Internal validity and design

Design enhances internal validity when it controls participant and order effects effectively.

Choosing the right design

Depends on balancing control, practicality, and minimising bias.

Strength of using designs strategically

Different designs allow researchers to tailor the study to reduce confounds.

Design-specific ethical issues

Repeated measures may cause boredom or fatigue; independent groups may risk unfair comparisons.