Concepts

1
Concepts
2

• Assertions
• Multiple working hypotheses
• multiple assertions
• social preferences
• auxiliary unobserved uncontrolled test conditions
• market institutions
• False dichotomy
• unrecognized hypotheses / explanations to
  behavior
• Confounds
• uncontrolled for influences assertions or test
  conditions (opportunities or choice options)
• Rationality
• Intentions vs. actions

3
Assertion

• When you want to explain an observed behavior
  some aspects of the explanation are not
  observable
• Objectives of the agent
• Preferences of the agent
• You therefore make assertions about these
• These assertions, combined with observable facts
  about the reality of the agent (test conditions)
• make it possible to make theory predictions
• to compare to your behavioral observations

4

• Comparing theory predictions to behavioral
  observations provides a test of the assertion
• but watch out
• Does Ceteris Paribus really hold in the data?
• Your inferences may be confounded by uncontrolled
  changes in other non-observed test conditions
  (opportunities)

5
Multiple Working Hypotheses

• Many possible assertions may seem reasonable
• A
• maximize utility over my own income/consumption
• B
• maximize utility including considerations for
  rewarding people who are nice to me (reciprocity)
• C
• maximize utility including considerations for
  those who end up with less (equity income
  inequality)

6
False dichotomy

• when several assertions make sense but you only
  test them pair wise
• A vs B neglecting the possibility of C

7
Confounds

• If you test one hypothesis against another, not
  recognizing the possibility of a third, this
  mistake can have a confounding influence on your
  inferences
• A causal explanation is confounded if there is
  more than one possible cause and the test does
  not control for that

8
Other possible assertions

• All optimization models assume some decision
  process that relies on cognition
• Agents making choices may not be maximizing
  absolutely
• cognition may be costly
• or there may be special technological
  considerations in the cognitive process

9
Rationality

• If you offer a higher reward for a task you would
  expect to see more effort / productivity
• Consistent with transitivity if in other
  situations the agent prefers more money to less
• Rejecting this hypothesis should not be
  interpreted as rejecting rationality
• other hypotheses such as detrimental rewards
  high extrinsic rewards can lower intrinsic
  rewards or high rewards cause distraction (a
  negative externality)

10
What counts as an observation?

• Stated intention
• Action

11
Keep this in mind

• throughout all of your research career
• and dont get fooled by the ceteris paribus
  assumption embodied in theory
• think outside the box