Research Project

Question 1

Research enterprise

Answer 1

Theoretical research - Empirical research - Qualitative/Quantitative

Question 2

Core elements of empirical economic research:

Answer 2

Validity, Replicability and Contribution

Question 3

Validity of empirical research

Answer 3

1) Construct validity: Is my measure right?
2) Internal validity: Is my comparison/conclusion correct for the sample?
3) External validity: Does my conclusion hold for other populations or settings?

Question 4

Replicability of empirical research

Answer 4

1) Data need to be accessible (store and do not change data)
2) Results need to be replicable ( create a do-file)

Question 5

Contribution to empirical research

Answer 5

Empirical economic research aims at either directly testing theories, or measuring relationships to feed theory. The scientific method or research enterprise

Question 6

Stages to empirical research

Answer 6

1. Posting the research question
2. Finding the relevant data
3. Adapting the data to the correct format
4. Analysing the data
5. Writing conclusions in a paper

Question 7

Types of data

Answer 7

Observational and Experimental

Question 8

Dummy variables

Answer 8

Values of 0 or 1 only

Question 9

crwork.do

Answer 9

From raw data to work data

Question 10

analysis.do

Answer 10

From work data to tables (descriptives)

Question 11

graphs.do

Answer 11

From work data to graphs

Question 12

Average Treatment Effect ATE

Answer 12

E(Yi(1)) - E(Yi(0)) = ATE = m1 - m2

Question 13

Sampling uncertainty or Sampling complication

Answer 13

Practical: we can not sample every person
Fundamental: the counterfactual problem - you get JSA or you don’t

Question 14

Random assignment

Answer 14

All individuals have an equal chance to get JSA or not

Question 15

Random sampling

Answer 15

All individuals have an equal chance to appear in the sample

Question 16

The central limit theorem

Answer 16

Y0 - All positive data for Y0 / total number of data for Y0

Y1 - All positive data for Y1 / total number of data for Y1

Question 17

Hypothesis testing

Answer 17

Using the central limit theorem to deal with sampling uncertainty
T-test = (y0 - m0)/SE(y0)
SE = SD(y0)/n^0.5

Question 18

The 95% confidence interval

Answer 18

[y0 - 1.96SE(y0) ; y0 + 1.96SE(y0)]

Question 19

Comparison, Testing the difference of the sample means (m0 =/= m1)

Answer 19

T-stat = y1-y0-(m1-m0)/SE(y1-y0)

SE(y1-y0)=(SE(y1)^2 + SE(y0)^2)^0.5

Question 20

Mean comparison T-stat example, 95% confidence interval

Answer 20

What hypothesises values of m1-m0 would be rejected by the data?
[ATE - 1.96SE(Y1-Y0) ; ATE + 1.96(Y1-Y0)]