Scribd
Upload
4 views

How To Read Research Papers

Uploaded by

Deepti Sharma
Copyright
© © All Rights Reserved
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
4 views

How To Read Research Papers

Uploaded by

Deepti Sharma
Copyright
© © All Rights Reserved
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/ 3

How to Read and Present Economics Papers

Tarun Jain

June 9, 2024

Reading, analysis and presentation skills are very important for professional success as an
economist (as well as for many other fields). In particular, by presenting and discussing research
papers, we learn how to criticize (in a good way) and evaluate the strengths and weaknesses of the
existing research. An important outcome of this exercise is that we can pursue new research ideas
by emulating the strengths and fixing the weaknesses of previous researchers. After all, there are
relatively few novel ideas and most researchers build on the extant literature.
Listed below are some questions to consider when you read papers and prepare your presenta-
tion.1 While presenting, your perspective should be one of an expert who knows more about the
paper than the other people in the room and is therefore qualified to answer questions. You are not
the author and don’t need to defend every aspect of the paper.
The Introduction tells the reader what to expect. This justifies the topic and the question under
consideration (and therefore why the reader is spending time on the paper). Also, recent papers
have begun dropping the literature section and incorporating relevant sections in appropriate places
in the text, mostly in the introduction. But placing the paper in the context of the literature is
important. Some questions you might want to ask are as follows.

• What is the question that this paper is trying to answer?

• Why is this question important?

• Why is this question important from a scientific perspective?

• Why is this question important from the perspective of policy?

• What does the literature have to say about the question?


1
A related exercise is writing referee reports that evaluate a paper for a journal. I don’t discuss that here.

1
Your second task is to figure out the paper’s precise contribution. For example, the author
might fill a gap in the literature (“previous studies have not examined the case where . . . ”), pro-
pose a new technique (“we introduce a new estimator to examine the effects of education on earn-
ings. . . ”) or fix a literature that’s broken (“estimates of the causal impact of education on earnings
suffer from previously unexamined issues. . . ”). This is the author’s “contract” with the readers. In
the sections following the introduction, you have to evaluate whether the author delivers on that
contract.

• What exactly does the author propose to achieve in the paper?

• How does the author propose answering the question?

Theoretical papers are a bit like classic British mystery novels. All the facts, agents, assump-
tions and timing of events must be laid out, and it must be possible to conclude the results logically
from these. No outside information can be introduced which is not available to the reader at the
beginning. If the model violates these conditions, then it must be immediately discarded. If it does
not, then we can proceed.

• How plausible are the model’s assumptions? A good model captures some essential core
characteristics of the world with minimum assumptions.

• Who are the agents and what information do they have? What is their choice set?

• What is the nature of interaction between agents?

• What is the timing of events?

• Are there any externalities in this situation and how does the model account for these?

• Are the results believable? How robust are they to small changes in assumptions?

The majority of papers we discuss are empirical papers. The most important question here is
whether the empirical strategy matches the research question. For example, a lab experiment is
unlikely to tell you about the impact of historical institutions on modern outcomes, just as archival
data is unlikely to help understand trust in dictator games.

• What is the empirical strategy (RCT, RDD, event study, lab experiment, structural model,
etc.)?

• Does the empirical strategy make sense for what we’re trying to learn?

2
Your next task is to understand data quality. This is perhaps the section where authors spend
the least effort, but is the greatest source of grief for bad papers.

• What is the data source? How was data collected? Is there any missing data?

• Are there any biases that might enter the analysis because of the data collection methodol-
ogy?

• What is the unit of observation?

Closely examine the empirical specification. No statistical technique is perfect, but some
come closer to showing the appropriate relationships between variables than others (in other words,
a better “identification strategy”). Many readers rightly focus on the issue of endogeneity, but also
consider selection bias (especially on unobservable characteristics), omitted variable bias and non-
linear relationships.

• What is the econometric analysis performed on the data?

• What are the robustness checks to ensure that the results are stable?

• What other methodologies that could have been used to answer the question?

If you’re convinced that the data and specification are appropriate, evaluate the results (If
you’re not convinced, there is no point proceeding). Often authors try to make more of their results
than they actually are. So see how well the claims match up to the outcomes. Focus on both the
statistical validity as well as the economic implications of the parameter estimates.

• What is the estimated parameter that corresponds to the main result of the paper?

• What do the additional parameters mean?

• Do you believe the results?

• Do you think the interpretation of the results is correct?

• What other results would you have liked to see?

Finally, the most important job to sit back and dream. Since you are a researcher, this should
be about 50% of your time.

• If you were writing a paper on the same topic, what would you do differently?

• If you were to write a related paper, what would you do?

You might also like