Lecture 2: Experiment Design
Econ vs Psych experiments:
-No deception
-Incentivized choices
Purposes of experiments:
Theory Experiments Theory
*Testing theory for robustness
Internal Validity:
-Extent to which differences are due to hypothesized mechanism
-Question of sufficient controls
-Must assess plausibility of alternative explanations
External Validity:
-To what extent do experiment results translate into the world
-Extrapolating experiments results
Types of Experiments:
1) Pure Laboratory Experiment
Subjects: Frequently students
Economic Environment: Artificially induced
Physical Environment: Artificial (laboratory) – subjects are aware of experiment
1) Pure Field Experiment
Subjects: Population of interest
Economic Environment: Naturally occurring environment
Physical Environment: Naturally occurring environment - subjects are unaware of experiment
Induced Value Theory:
-How do we induce preferences in the laboratory
-Proper use of a reward medium allows experimenter to induce prespecified characteristics in sub jects
3 Conditions:
-Monotonicity: prefer more reward medium to less
-Salience: reward depends on subject actions
-Dominance: change in subject utility comes “predominantly” from the reward medium
Identification:
Within-subject Design:
-Each subject does all treatments
Worry: Order effects; learning increased salience of inter-treatment differences; easier identification of experimenter demand
Mitigation: Randomly assigned sequences
Between Subject Design:
-Each subject receives one treatment
-Compare treatment groups
-Hope that unobservables independent of treatment assignment
Experimenter-Demand Effects:
1)Hawthorne Effect:
Being Observed Changes in Behavior Not a problem if constant across treatments
2) Intertreatment Differences: Make differences salient
Treatment A: Learn Avg $ given by other givers
Treatment B: