Week 11- Casual research: experimental Design

How to establish causality

Needs to prove 3 things:

• Concomitant variance (aka Correlation)

• Precedence (if X causes Y, then X must occur before that)

• Elimination of other possible factors

  • Z (extraneous factors) do not cause Y

Kinds of Extraneous Factors

• History

• Maturation

• Instrument variation

• Selection Bias

• Mortality

• Testing effects

• Regression to the mean

Types of Experimental Designs

• Pre-experimental

  • Very little control over extraneous factors

• True experimental

  • Introduction of Control Group

  • Random assignment of subjects to treatment

Pre Experimental <

• One-shot design

• One group pre and post test

True experimental

• Before and after with CG

• After only with CG

One shot Design

Do something and observe

One group pre and post test

measure/ record the pre, do something and then measure the post again

Before and After CG

Randomly choose ppl, and assign them into either EG (exposed group) or CG ( control group)

After only with CG

Grab random ppl and assign into EG and CG but only record their responses after exposure

Lab experiments are

Hard to do and lack external validity

Test Markets provide

Alt to Lab experiments

Benefits of Lab tests

• Opportunity to