1. Concepts to be included
Experimental group
Control group
Random assignment
Posttest
Pretest
Treatment
Observation

2. The classical experiment
Henk van der Kolk

3. Aim
Introducing the classical experiment: a research design for testing bivariate causal relationship. Also known as “Randomized experiment”.

4. Example
Does watching television news affect the importance people attach to political topics?

5. Example Political opinions Watching news about a political topic
Importance attached to that political topic

6. (Classical) Experiment (def.)
A research design, in which comparable groups of units are constructed using random assignment and in which the treatment (independent variable) is manipulated differently across groups to see whether the outcome (dependent variable) becomes different across groups.
Deze slide is lastig om overzichtelijk te maken. Er staan nu steeds drie puntjes, maar die kun je gewoon weer weghalen als je het niks vindt.

7. Random assignment = randomization
Experimental group
Units
Control group
Random assignment

8. Randomization created Comparable groups
Experimental group
Pre-test
Dep var
Units =
Control group
Pre-test
Dep var
All other variables: gender, age, etc.
Random assignment

9. Example Experimental group Important? Medium Units Medium
Control group
Important?
All other variables: gender, age, etc.

10. The classical experiment
Experimental group
Pre-test
Dep var
Treatment
Units
Control group
Pre-test
Dep var
Placebo
Random assignment

11. Example Experimental group Important? News Units Medium Medium No news
Control group
Important?
Random assignment

12. The classical experiment
Experimental group
Pre-test
Dep var
Treatment
Post-test
Dep var
Units
Compare
Control group
Pre-test
Dep var
Post-test
Dep var
Placebo
Random assignment

13. Example Experimental group Important? News Important? Units Medium
High
Medium
Medium
No news
Control group
Important?
Important?
Random assignment

14. Three aspects of causality
Time order; X precedes Y in time
Association/correlation; X and Y are correlated
Non spuriousness; there is no other (third) variable accounting for the correlation

15. Time order
Because of the pre-test and the observed similarity between the two groups … Because we focus on the difference between the two groups AFTER giving the treatment … … we are sure the independent variable precedes the dependent variable: no reverse causation.

16. Example Experimental group Important? News Important? Units Medium
High
Medium
Medium
No news
Control group
Important?
Important?
Random assignment
Time
Is it possible that “importance” affects “watching the news”?

17. Association
By treating one group and not the other By comparing the outcomes of the control group and the experimental group We establish (the absence of) an association.

18. Example Experimental group Important? News Important? Units Medium
High
Compare
Medium
Medium
No news
Control group
Important?
Important?
We can check that those watching the news, attached more
importance to a topic after watching the news

19. Non-spurious
Because of random assignment and checked by the pre-test, the only difference between the control group and the experimental group is … … the threatment. We thus make sure, there is no effect of other (third) variables

20. Example Political opinions Watching news about a political topic
Importance attached to that political topic

21. Classical experiments
Two groups
Randomization (= random assignment)
One pre test of the dependent variable
One post test of the dependent variable
One intervention / treatment (the independent variable)

22. This microlecture
Introducing the classical experiment as a research design for testing bivariate causal relationships
Experimental group and Control group
Random assignment
Posttest and Pretest
Treatment

24. Images used:
Slide 5:
Slide 8: