Lecture01a Slides

Uploaded by

edes.balhes

We take content rights seriously. If you suspect this is your content, claim it here.

Available Formats

Download as PDF, TXT or read online on Scribd

0% found this document useful (0 votes)

14 views8 pages

Lecture01a Slides

Uploaded by

edes.balhes

We take content rights seriously. If you suspect this is your content, claim it here.

Available Formats

Download as PDF, TXT or read online on Scribd

You are on page 1/ 8

Métodos Quantitativos em Contabilidade I

Programa de Pós-Graduação em Controladoria e Contabilidade

Lucas Barros
[email protected]
Henrique Castro
[email protected]

Universidade de São Paulo

1 of 8
Lecture 1a
Questions about Questions

2 of 8
Successful research project
Questions about Questions

• A research agenda can be organized around four questions (FAQs):

1. The relationship of interest,

2. The ideal experiment,
3. The identification strategy,
4. The mode of inference.

3 of 8
First FAQ: in the beginning . . .
Questions about Questions

What is the causal relationship of interest?

• Much of the most interesting research in social science is about
cause and effect. Examples:
◦ The effect of class size on children’s test scores.
◦ The effect of accounting practices on firm performance.
◦ The effect of external auditing on the likelihood of accounting fraud.

4 of 8
Second FAQ
Questions about Questions

What is the ideal experiment?

• The second research FAQ is concerned with the experiment that
could ideally be used to capture the causal effect of interest.
◦ Usually, that would be a randomized controlled trial (RCT).
• Ideal experiments are most often hypothetical:
◦ Randomly assign students to different class sizes and measure
performance.
◦ Randomly assign different accounting practices to firms and measure
performance.
◦ Randomly assign external auditing to firms (leaving some firms
without external auditing for a few years) and measure the likelihood
of accounting fraud.

5 of 8
Second FAQ
Questions about Questions

What is the ideal experiment?

• We usually won’t be able to implement the ideal experiment.
• Still, they are worth contemplating because they help us find
fruitful research topics and also design and implement the actual
research strategy.
◦ In other words, they provide a useful benchmark.
• If your research question cannot be satisfactorily answered with the
ideal experiment, your chances of answering it in real life will be
very small.
◦ You may have a fundamentally unidentified question (FUQ).

6 of 8
Third FAQ
Questions about Questions

What is your identification strategy?

• The manner in which a researcher uses observational data (i.e.,
data not generated by a perfect randomized trial) to approximate
the ideal experiment. For example:
◦ Comparing firms that are required by law to hire external auditors to
similar firms that are not.
◦ Using a natural experiment, such as a change in accounting rules
that affected different groups of firms in different ways.
◦ Using control variables or panel data strategies to isolate relevant
confounding factors.

7 of 8
Fourth FAQ
Questions about Questions