Research and Design 13: Quasi-Experiments and Small N-Designs

True Experiment

Manipulates independent vairable, uses random assingment, strong internal validity, forms strong causation

Quasi-Experiment

Includes the manipulated variable but has no random assingment, the groups exist before the experiment, could be due to confounds

Example: Students chose their groups

Selection Bias in Quasi

Groups differ before the study

Selection and History

Groups change differently overtime

Selection and History

External events affects the groups differently

Non-Equivalent Control Groups Design

Groups are not randomized, there is a pretest and posttest to establish starting point in the groups

Interrupted Time Series Design

Measure repeatedly over time, shows temporal precedence

Small N-Design

A small sample of typically one person, measured repeatedly overtime

Reversal Design

ABA or ABAB

A - Baseline

B - Treatment

ABAB - Measure behavior, apply treatment, remove treatment, introduce treatment again

Proves that the outcome is changing with the treatment

Stable Baseline Design

Observe behaviors until they are stable, introduce treatment after

Small N-Designs Disadvantages

Low External Validity - Not randomized or generalizable

Limited Sample - Sample doesn’t represent population

Subjective

Ethical - You cannot remove treatment

Sample size decreases the external validity

Small N-Designs Strengths

High internal validity, can prove that there are no confounds

Large control

Conditions are specific to the participant