Suicide Bomber Instructor Discussion Guide

We want to use this paper as an example for you to follow in your own

work.

Put on board the parts of an empirical paper—just the bolded parts and

have them say what goes in each one.

I.

Introduction

A.

Statement of the topic and question to be analyzed

B.

Rationale for choice of the topic (or why you find this

interesting)

C.

Explanation of the organization of the remainder of the paper

II. Literature Review

Choose some form (e.g., chronological or thematic) to organize the

literature review. Mere listing and summary of several sources is not

acceptable. A good literature review interweaves the various articles

in a seamless way.

III. Theoretical Analysis

Present a brief version of a model or highlight the theoretical source

of the hypothesis to be tested. In many cases, you may wish to combine

the literature review and theoretical analysis into a single section.

For example, a paper you review may contain a version of the model you

wish to adapt for your own analysis.

I

All data and analyses must be completely documented and available for

inspection.

V. Empirical Analysis (the main and longest part of the empirical

paper)

A.

The Data

i.

Provide sources on all variables

ii.

Provide summary statistics on all variables in a

well-organized table

B.

Presentation and Interpretation of Results

V. Conclusion

A.

Restate the topic or question that was analyzed

B.

Provide your answer or conclusion, and compare to previous results

in the literature

C.

Point out the best areas for further research

VI. References

Stylistic Issues

Explain the JEP journal—more talky than usual econ journals

Notice the catchy intro.

How is the paper structured? Get students to list elements:

Intro

Where is the Lit Rev? Included Lit Rev in the Intro

Citation style

Author name (Year) and references at the end

Highlight p. 225 – “Our argument fits within the growing body of lit”

Summary Stats

Fig 1

See SuicideBomber.xls for discussion of correlation

Notice how they discuss specific numbers from figures and tables.

Discuss box model (see Excel file)

Table 1

Excel file reproduces parts of this table

Data Description

How the data were collected?

Augmented with ISA data.

I emailed them and they refused to give me the data because it’s

secret.

Table 2 Target Importance – size of city and military

What are they trying to say with this table?

Big city attacks kill and injure more people so they are more

important

Plus, they are trying to figure out how to measure the importance of a

target and have two competing measures, large/small city and

military/civilian

Content: Bottom Line

What is this paper’s main point?

Suicide bombing involves an optimization problem.

Suicide bombers are educated.

Are older, better educated suicide bombers more productive?

Content: Theory

What is human capital?

People’s education and training. Capital the stock of machinery and

equipment that a firm uses to produce. When it buys capital, it

invests.

Human capital is the stock of skills and knowledge in a person. When

you go to college, you are investing in yourself so you are said to be

acquiring human capital.

What is the optimization problem involved for the head of a terrorist

organization?

Who to send on a mission and how to do it. “Attack assignment”

What is the suicide bomber’s optimization problem? On a bus, what is

the choice variable?

When to blow yourself up. Too soon is bad. Too late is bad.

Content: Metrics

Start with anecdotal evidence from Table 3. What does this show?

That educated people are bombers.

Table 4 is a probit regression. The dependent variable is a dummy

variable, 0/1. We will study this later.

Coefficient on Education is not positive for large city, but they

report it anyway.

Conclusion: older bombers assigned to large cities, less educated to

military targets

Table 5 is the key table. What is the table trying to determine?

If older, more educated bombers are more productive. It’s the title of

the paper.

If you are unsuccessful, what is your value for number of people

killed?

Zero

What does the scatter plot with age look like?

Bunch of zeroes (42 to be exact) sprinkled in for the 148 dots. Y axis

has #killed, X axis has age

This is why they separate out the All and Only successful regressions.

Notice that they report both. You should report different estimates

from different models in your paper. Use this table style to report

the coefficients and other pertinent information.

HUGE QUESTION

If Education coefficient is negative, doesn’t this mean that education

lowers the number of people killed? Doesn’t this disprove the entire

argument of the article?

No, because of the interaction term.

As education goes from 0 to 1, both the Education coefficient and the

interaction term kick in.

How to construct an interaction?

Here are some hypothetical observations. Fill in the table.

Age

Education

Education Dummy

Size of City

Target Dummy

Education x Target

19

In High School

Tamra—26,000

24

Masters Candidate

Haifa—300,000

21

High School

Tira—21,000

29

Law School Graduate

Jerusalem—800,000

How to interpret the coefficients? What is the effect of Education on

the number killed?

See Excel file, Table5 sheet.

You can set up a sheet like this when you are reading an empirical

paper. Reading empirical papers is NOT like reading a novel. You have

to dig deep to figure out what is going on.

Explain how the sheet works: you enter values in the yellow cells and

the sheet computes the Predicted Y (number killed) for that case.

Compare the cases cited in the article and entered in the Excel file.

Explain how the case approach gives the same value as that computed in

the sheet as the change in target for an uneducated, 25 year old.

Predicted Y = b0 + b1Age + b2Education + b3Target + b4Age x Target +

b5Education x Target

They use a formula that can be derived like this.

If an uneducated, 25 year old, small city target (initial value)

changes to an uneducated, 25 year old, big city target (new value), we

would compute

Change in Predicted Y = b0 + b1[25] + b2[0] + + b3[1] + b4[25] x [1] +

b5[0] x [0]

- b0 - b1[25] - b2[0] – b3[0] – b4[25] x [0] – b5[0] x [0]

The b0, b1[25], b2[0], and b5[0] x [0] terms drop out and all you have

left is

Change in Predicted Y = b3[1] + b4[25] x [1] – b3[0] – b4[25] x [0]

This is the formula that they use.

You get the same answer if you compute Predicted Y for an uneducated,

25 year old, small city and then compute Predicted Y for an

uneducated, 25 year old, big city and subtract one from the other.

That is what the Excel file does.

The regression says that more people are killed when a bigger target

is chosen for uneducated, 25 year old bombers.

The effect depends on the age. What happens if an 18 year old bomber

moves from a small to a large city?

Compare the cases. -0.24 is the answer (replicating the 0.2 in the

article).

What about education? (which is what the article is supposed to be

about)

Compare cases

Assigning a 25 year old, uneducated bomber to a big city target gets

you 17.5 people killed. What’s the RMSE? They don’t tell us. The R^2

is pretty low, though, so there’s probably a lot of dispersion.

Assigning a 25 year old, educated bomber to a big city target gets you

19.15 people killed. That’s an increase of 1.65 people killed. More

educated bombers kill more people when the target is a big city.

Does Education improve productivity for a small city?

Compare cases. The answer is no.

Table 6—what is the point?

That successful bombers are older and more educated.

Excel file has the box model for this.

905314.doc Page 6 of 6