Lecture 1 - Methodology Flashcards
Behavioural Economics is the study of:
how people make decisions under scarcity
Normative (perspective) theory
sets a benchmark for behaviour if certain assumptions are met
e.g., a rational consumer will choose the optimal bundle given their preferences and budget constraint.
- captures how people should make decisions
Positive (descriptive) theory
aims to describe behaviour
e.g. describe how social identities influence our decision making and behaviour.
- captures how people actually make decisions
What are “As-if” models not concerned with?
not concerned with the process
e.g.. Model of utility maximization
What do Process models do?
- predict choices
- model the processes that produce them
What are the roles of economic theories?
- make predictions
- describe and explain human behaviour
What are the criteria for “good” theory?
- congruence with reality: explain + predict correctly
- Generality: applicable to different phenomena
- Tractability: complex theories may have more parameters(by making fewer assumptions) - difficult to represent as models
Parsimony: having fewer parameters (by making more assumptions) - potential trade-off with descriptive accuracy - Precision (giving exact numerical predictions about behaviour
What are the types of empirical evidence?
. Observational data/naturally occurring: (e.g. data on wages or unemployment rates)
. Experimental data:
- Randomized control trials (e.g. effect of 4-day working week on productivity)
- Field experiments (e.g. sending resumes with African-American or white-sounding names)
- Lab experiments (e.g. framing effects on individual choice)
- Lab-in-the-field experiments (e.g. dictator games and ultimatum games in a field context)
. Neuroscientific evidence: (e.g. MRI scans)
. Biological markers: (e.g. stress markers)
. Surveys: (e.g. who do you intend to vote for)
Observational Data Pros
- Ecological validity (relevant for the population studied)
- Often cheap to obtain
- Does not involve interference
- Often large datasets available
Observational Data Cons
- Internal Validity is not clear (there may be confounding factors that do not allow a causal interpretation of the findings)
- External Validity not clear - to what extent do the findings apply to other populations, places, time, environments
- Researcher has no control over data collection process
What is an experiment?
A method to collect data in a controlled environment
Why is randomization crucial?
It’s crucial for establishing a causal effect of treatment on outcome
(Lab) experimental studies pros
- Researcher can design the data generating process
- Internal validity - if study is well-done it allows for a causal interpretation
- Often less costly than other methods
- Can systematically vary features of decision situation and are therefore suitable for testing theories (keeping all else equal)
(Lab) experimental studies cons
- Need to think about external validity in other contexts, with other populations
Neuroscientific evidences pros
may learn more about “process” of decision making
Neuroscientific evidence cons
- small sample sizes
- Non-neuroscientists have difficulty interpreting an MRI scan
Biological markers pros
Learn more about physiology of decision making
Biological markers cons
- difficult to interpret by non-specialist
- difficult to collect
Surveys pro
learn more about what people think (beliefs that are not revealed through choices)
Surveys cons
- people may not be honest when answering questions
- people may behave differently than they say they would
What can theory do for experiments?
- provide models to be tested
guide for identifying key variables, potential confounds, competing explanations - Derive new explanations
What can experiments do for theory?
- Experiments can test predictions of a theory
- Test which of several competing theories best explains the data
Help define the scope of a given theory
Refine theories & suggest/construct new theories
Interpreting empirical evidence
Theories can be proven false but they can never be shown to be correct
Internal Validity
- statistically significant
- statistical credibility of the empirical analysis
- refers to the extent to which a study accurately establishes a causal relationship between the treatment and the outcome
Something is internally valid if it’s
- internally valid
- statistically significant
- theory
External Validity
- Relevance of empirical analysis for example to policy
- Refers to the extent to which the results of a study can be generalized to other contexts, populations, and times
Causal Impact
Introduction of a program brings an outcome that would not have been achieved in the absence of the policy