3. Experimental Design

Question 1

What are the four classic causal hurdles (questions to ask yourself before saying there’s a cause-and-effect relationship

Answer 1

1. Is there a credible causal mechanism ?
2. Can we rule out reverse causality?
3. Is there a covariation between the cause and effect?
4. Have we accounted for confounding variables?

Question 2

Ways to cross these hurdles ?

Answer 2

• Multivariate regression analysis: statistical adjustment for all
• Research design with randomization

Question 3

What is randomization and why is it so important ?

Answer 3

Randomization is the process of assigning participants, treatments, or experimental conditions in a study by chance rather than by choice. (treatment groups and control)

Gold standard in experimental research because it enhances fairness, reduces bias, and strengthens causal conclusions.

Question 4

What is a factorial design

Answer 4

It’s a a design experiment that uses more than one independent variable.

Question 5

What are the 5 first step to set up an experiment

Answer 5

1. Problem
2. Choice architecture (statu quo): what is the default (control)
3. Potential behavioral factors or biases leading the problem
4. Ideas to re-design choice architecture
5. Experimental design