Experimental Research Design

What is the main goal of experimental research?

To test causal relationships by manipulating an independent variable (IV) and measuring its effect on a dependent variable (DV).

What is an independent-groups (between-subjects) design?

Different participants take part in each condition; each group receives a different treatment or represents a different group.

What is a posttest-only independent-groups design?

Participants are assigned to different conditions and the DV is measured after the treatment.

What are random groups?

Groups formed by randomly assigning participants to experimental conditions.

What are natural groups?

Groups based on pre-existing participant characteristics rather than researcher assignment.

What are matched groups?

Participants are matched on one or more relevant characteristics, then assigned to different conditions within each matched set.

What is a within-subjects (repeated-measures) design?

The same participants complete every condition, and their DV scores are compared across conditions.

What is a matched-pairs design?

Participants are paired based on relevant characteristics, and members of each pair are assigned to different conditions.

What is a major advantage of between-subjects designs?

They avoid order and carryover effects because each participant experiences only one condition.

What are other advantages of between-subjects designs?

They are straightforward and useful when participants cannot reasonably experience every treatment.

What are disadvantages of between-subjects designs?

They require more participants, may have lower statistical power, and are more vulnerable to pre-existing group differences and extraneous variables.

What is a major advantage of within-subjects designs?

Each participant serves as their own control, reducing variability due to individual differences.

What are other advantages of within-subjects designs?

They need fewer participants and can be more sensitive to treatment effects.

What are disadvantages of within-subjects designs?

They may be impractical and can produce order, practice, fatigue, contrast, or carryover effects.

What is an order effect?

A change in responses caused by the position or sequence of a condition rather than by the condition itself.

What is a practice effect?

Performance improves because participants become more familiar with a task over time.

What is a fatigue effect?

Performance worsens because participants become tired over time.

What is a contrast effect?

A response to one condition is influenced by the condition experienced immediately before it.

What is a carryover effect?

The effect of one treatment persists and influences responses to a later treatment.

How can order effects be controlled in within-subjects designs?

Use counterbalancing: vary the order in which participants experience conditions.

What is random assignment?

Using chance to place participants into conditions; it helps distribute participant characteristics across groups.

How is random assignment different from random sampling?

Random sampling selects people from a population; random assignment places recruited participants into experimental conditions.

What is block randomization?

Assign participants in small blocks so each condition receives participants before the next block begins, helping keep group sizes balanced over time.

What is counterbalancing?

Presenting conditions in different orders across participants to reduce the influence of order effects.

What is complete counterbalancing?

Using all possible treatment orders; with k conditions there are k! possible orders.

What is reverse-order counterbalancing?

Presenting the conditions in one order and then in the reverse order.

What is block-randomized counterbalancing?

Each condition appears once within a block, with the order randomized within each block.

What is incomplete counterbalancing?

Using only a subset of all possible treatment orders across participants.

What is a Latin square?

A counterbalancing arrangement in which each condition appears once in each ordinal position across the selected sequences.

What is the purpose of a Latin square?

To balance the position of each condition without requiring every possible order.

What does manipulating the IV mean?

The researcher deliberately changes the condition, experience, or stimulus participants receive.

What is a straightforward or natural manipulation?

A manipulation using direct instructions or stimulus presentation without constructing an elaborate scenario.

What is a staged manipulation?

A manipulation designed to create a particular psychological state or simulate a real-world situation; it may involve confederates.

What is a confederate?

A person working with the researcher who appears to be another participant or bystander but follows a planned role.

What is manipulation strength?

How strongly the IV is designed to produce a difference between conditions.

What is a potential problem with an overly strong manipulation?

It may reduce external validity if the situation is unrealistic, and it may raise ethical concerns.

What are common types of IV manipulation?

Psychological experiences, stimulus/experimental manipulations, and participant characteristics.

Why are participant characteristics not true manipulated IVs?

They are measured or pre-existing characteristics, not variables the researcher directly assigns or changes.

What is a confounding variable?

A variable that changes systematically along with the IV, making it unclear which variable caused the observed effect.

What are common ways to measure a DV?

Self-report, behavioral observation, and physiological measures.

What can self-report measures assess?

Attitudes, judgments, perceptions, emotional states, and knowledge, often through surveys or tests.

What are behavioral measures?

Observable actions, such as response rates, reaction times, or duration of a behavior.

What are physiological measures?

Biological responses, such as measures collected using EEG, MRI, or fMRI.

What is sensitivity of a measure?

Its ability to detect differences or changes in the DV.

How can a measure be made more sensitive?

A sufficiently broad range of possible scores can help detect differences.

What is a ceiling effect?

Scores cluster at or near the highest possible value, making further improvement difficult to detect.

What is a floor effect?

Scores cluster at or near the lowest possible value, making further decreases difficult to detect.

What is a nuisance variable?

An unwanted variable that increases score variability within groups, making the IV's effect harder to detect.

How is a nuisance variable different from a confound?

A nuisance variable adds unwanted variability but does not systematically vary with the IV; a confound does vary with the IV.

What factors influence participant selection?

The research question, prior research, participant availability, time, finances, and whether the design is within- or between-subjects.