5.0 quasi-experiment

5.0 quasi-experiment

TRUE EXPERIMENTS

• leads to an unambiguous outcome regarding what caused an even

• - some interventions are implemented

• high degree of control → high internal validity

• Appropriate comparison

• random assignment → group start equivalent

• high degree of control of true experiments:

1. Over the arrangement of experimental conditions

2. Assignment of participants

3. Systematic manipulation of IV

4. Choice of DV

experiments in natural settings:

• low degree of control

• → Lower internal validity, but higher external validity because results apply to real-world settings.

experiments in lab

• high degree of control of conditions & participant assignment → High internal validity (can identify cause + effect).

Why do we conduct experiments in natural settings? (Applied Research)

• Goal is to solve real-world problems & improve real situations.

• Applied experiments can have immediate, practical consequences (e.g., education, health, public policies).

• They are often used to test whether lab findings apply in real life → boosts external validity.

Why do we use Quasi-Experimental Designs in Applied Research?

• random assignment is not always fair, practical, or possible in real-world settings → therefore researchers study naturally occurring groups instead of randomly assigning participants.

Main Challenges in Applied (Quasi-Experimental) Research

• Hard to get permission from organizations or authority figures.

• Limited access to participants (researcher cannot choose who participates).

• Lower internal validity because groups may differ before the intervention.

QUASI-EXPERIMENTS

• Quasi- = resembling

• Includes some intervention / treatment types

• Provide comparison

• Lack degree of control because groups are pre-existing

• lack of randomization = no random assignments

Key Difference Between True Experiments and Quasi-Experiments

• True experiments: use random assignment → high internal validity (can infer cause + effect).

• Quasi-experiments: no random assignment → groups may differ before the intervention → lower internal validity.