Two Group Designs

Two-group experiment

An experiment that separates participants into two groups, exposes them to different levels of an independent variable, and compares their outcomes.

Independent variable (IV)

The variable manipulated by the researcher to examine its effect on an outcome.

Dependent variable (DV)

The outcome measured to assess the effect of the independent variable.

Level

A specific value or amount of an independent variable; a two-group design has two levels, such as painkiller versus placebo.

Condition

A label identifying the level of the IV participants receive, or the group receiving that level.

Condition labels as nominal data

Labels such as experimental and control identify categories; the labels themselves do not represent numerical amounts.

Control condition

The group exposed to a neutral level of the IV, often its absence, such as a placebo.

Experimental condition

The group exposed to the manipulation or level of the IV being investigated.

Assignment

The process of placing participants into study conditions.

Random assignment

Assigning participants to conditions by chance to reduce systematic differences between groups.

Purpose of random assignment

To make groups comparable on individual differences so that the main systematic difference is the IV; larger samples help differences balance out.

Purpose of two-group designs

To test hypotheses involving two levels of an IV, such as presence versus absence, more versus less, or Treatment A versus Treatment B.

Ethical concern about experiments

Intentionally manipulating people's experiences, emotions, or behavior can raise ethical concerns.

Manipulation

The procedure used to change participants' exposure to or experience of the independent variable.

Pilot test

A small preliminary study used to assess whether a manipulation works as intended.

Prompt-based or instruction-based manipulation

Instructions that direct participants to think about or engage with different topics to influence a psychological construct.

Vignette

A story or passage with features that can be varied between conditions to manipulate participants' experiences.

Writing task

A manipulation in which participants write about a topic to engage with it; its primary purpose is manipulation, even if the writing can also be measured.

Priming

Presenting a stimulus to bring associated concepts to mind.

Confederate

A research team member who poses as another participant or someone unrelated to the study.

Manipulation check

A measure used to determine whether the intended construct was actually manipulated.

Experimental realism

The extent to which a study engages participants and psychologically affects them.

Demand characteristic

A cue that reveals the study's purpose and may change participants' behavior.

Reactance

Responding against the perceived intention of a manipulation, sometimes by doing the opposite of what the researcher expects.

Longitudinal experiment

An experiment involving manipulation or outcome measurement over time; useful when effects develop later or treatments require repeated sessions.

Challenges of longitudinal experiments

They require more time and introduce additional opportunities for error.

Between-groups or between-subjects design

A design comparing outcomes from different groups of participants.

Independence of observations

Observations do not depend on or share information with other observations; between-subjects comparisons generally use different people in each group.

Dependent observations

Observations that share information, such as repeated scores from the same person; ordinary independent-samples analyses do not account for this dependence.

Why not compare group means alone?

Group means usually differ somewhat because of sampling and assignment error; statistical tests assess the evidence beyond that expected variation.

Within-groups differences

Variation among individuals within the same condition.

Independent-samples t-test

A statistical test comparing two independent groups on an interval- or ratio-level outcome.

Logic of the t-test

Assesses the difference between group means relative to the variability and uncertainty in the data.

Social-pressure experiment example

Participants doing math while peers watch are in the experimental condition; participants doing math alone are in the control condition.