Gender Differences in Returns on Education    I. Introduction    For a society that claims to value equality in the workplace, the gender gap in wages in  America seems awfully persistent. This paper investigates the differences in wages between men  and women at different levels of education using data from a sub sample of the Current  Population Survey (2012). Such analysis will help reveal the nature of the gender gap, and may  help identify the segments in which discrimination in the workforce may exist. Using linear  regressions, I first confirm the wage gap in the data and that returns to education are positive.  Next, I use interaction variables to illuminate gender differences on returns at the different levels  of education (high school, bachelor’s, and master’s). Overall, I find that females see higher  returns than men for completing high school and college, but not for graduate school.   II. Data    The data set consists of 999 observations of working individuals between the ages of 18  and 54:  The average age in the sample is 39.11 years old. On average, individuals made $16.92  an hour with a standard deviation of $9.80. The average highest grade completed, 13.28, shows  that most graduated high school. 88% of the sample have high school diplomas, 24% hold a  bachelor's degree, and 7.4% have completed at least a master’s. ​ ​A majority was white (81.6%).  10% of the individuals were black, 9% were other races. 22.7% of the workers were part­time.  Approximately half of the sample was female. The following histogram shows the distribution of  education level:        Most of the data lies on the milestone years. The 12, 14, 16, and 16 areas represent high school  diplomas, associate's, bachelor’s, and master’s degrees. However there is some ambiguity at the  14th grade level: these observations could be both associate’s degree holders or four year college  dropouts.                 III. Empirical Methodology    To compare gender differences in the returns on wages at different levels of education I  run a linear regression on log wages:    The particular variables of interest are B9, B10, and B11. These interaction variables will show  the additional percentage point increase or decrease in wages that females accrue at the different  levels of education.   Because the distribution of wages is skewed right, I choose to use log wages, which are  more normally distributed and thus may increase the goodness of fit. Based on prior research, I  expect to see positive, though diminishing, returns to age. Thus, one would expect B1 to be  positive and B2 to be negative. Income inequality between whites and blacks is well established  in economic literature, so I expect B3 to be negative. B4 is also likely negative since many of the  higher paying jobs would be full time. I expect a negative coefficient on the female variable,  matching my hypothesis that the wage gap is present in the data. Lastly, the coeffic ...

1. Gender Differences in Returns on Education
I. Introduction
For a society that claims to value equality in the workplace, the
gender gap in wages in
America seems awfully persistent. This paper investigates the di
fferences in wages between men
and women at different levels of education using data from a su
b sample of the Current
Population Survey (2012). Such analysis will help reveal the nat
ure of the gender gap, and may
help identify the segments in which discrimination in the workf
orce may exist. Using linear
regressions, I first confirm the wage gap in the data and that ret
urns to education are positive.
Next, I use interaction variables to illuminate gender difference
s on returns at the different levels
of education (high school, bachelor’s, and master’s). Overall, I f
ind that females see higher
returns than men for completing high school and college, but no
t for graduate school.

2. II. Data
The data set consists of 999 observations of working individuals
between the ages of 18
and 54:
The average age in the sample is 39.11 years old. On average, in
dividuals made $16.92
an hour with a standard deviation of $9.80. The average highest
grade completed, 13.28, shows
that most graduated high school. 88% of the sample have high s
chool diplomas, 24% hold a
bachelor's degree, and 7.4% have completed at least a master’s.
A majority was white (81.6%).
10% of the individuals were black, 9% were other races. 22.7%
of the workers were part-time.
Approximately half of the sample was female. The following his
togram shows the distribution of
education level:
Most of the data lies on the milestone years. The 12, 14, 16, and
16 areas represent high school

3. diplomas, associate's, bachelor’s, and master’s degrees. Howeve
r there is some ambiguity at the
14th grade level: these observations could be both associate’s d
egree holders or four year college
dropouts.
III. Empirical Methodology
To compare gender differences in the returns on wages at differ
ent levels of education I
run a linear regression on log wages:
The particular variables of interest are B9, B10, and B11. These
interaction variables will show
the additional percentage point increase or decrease in wages th
at females accrue at the different
levels of education.
Because the distribution of wages is skewed right, I choose to u
se log wages, which are
more normally distributed and thus may increase the goodness o
f fit. Based on prior research, I
expect to see positive, though diminishing, returns to age. Thus,

4. one would expect B1 to be
positive and B2 to be negative. Income inequality between whit
es and blacks is well established
in economic literature, so I expect B3 to be negative. B4 is also
likely negative since many of the
higher paying jobs would be full time. I expect a negative coeffi
cient on the female variable,
matching my hypothesis that the wage gap is present in the data.
Lastly, the coefficients on the
dummy variables for completion of high school, completion of a
bachelor’s, and completion of a
master’s are expected to be positive because higher education le
vels allow individuals to access
higher wage positions.
There are some potential concerns with this methodology. First,
there is inevitably a
sample selection problem since we are only looking at the data
of employed people. For
example, if it were the case that being female lowered the proba
bility of being employed due to
discrimination, then the sample observations would only represe
nt the females with a relatively
high marginal productivity of labor. Thus, B5 might underestim
ate the true magnitude of the
wage gap. Another possible concern is omitted variable bias. Fo
r example, living in a city is
likely positively correlated with higher wages. Moreover, it may
be the case that having a
master’s degree is correlated with a higher probability of living
in a city. These positive
correlations would cause an upward bias on B8.

5. IV. Results
Column 1 shows the baseline specification which includes all th
e race dummy variables and does
not include the interaction variable for female and education. Al
l the variables are significant at
the 5% level except for the race dummy variables for American
Indian, Asian, Mixed, and
Hispanic. All signs on the significant coefficients are consistent
with the predictions discussed in
the previous section. I conducted an F (4,986) test on the variab
les for American Indian, Asian,
Mixed, and Hispanic and found that none had a statistically sign
ificant impact on wages, all else
constant (p=.18). Thus, I decided to remove them from the regre
ssion in column 2. The adjusted
R^2 for column 1 was .164.
In column 2, I add the interaction variables for females and edu
cation level. The adjusted
R^2 slightly improved in this specification to .168. The interacti
on variables are interpreted as
follows: females see an additional return of 5.6% compared to

6. men for graduating high school,
an additional 24% return compared to men for graduating colleg
e, and a negative 26% return
compared to men for a master’s degree, all else equal. All the in
teraction variables are significant
at the 5% level. The signs on the rest of the coefficients are con
sistent with the initial predictions,
except for the coefficient on bachelor’s degrees. The statistical i
nsignificance of the bachelor’s
variable in this regression can be explained as a point in the edu
cation level where female wages
catch up to the male wages (i.e. the wage gap closes). In fact,
my regression predicts that at the
bachelor’s degree level of education, female wages average abo
ut 8 percentage points higher
than men. However, the gap reappears for graduate level jobs. A
t that point, the model predicts
that men see about a 20% higher return than women, all else con
stant. The following graph
shows how the wage gap is “pinched” for bachelor’s degree wag
es:
V. Conclusion
In theory, it is not surprising that a wage gap persists at lower l
evels of education. Jobs
that do not require degrees tend to involve more manual labor, t
hus have positive returns on
physical strength. According to my results, female wages catch
up to male wages with a

7. bachelor’s degree, but lag behind male wages at the graduate le
vel. This may be evidence of
gender discrimination for senior positions. All in all, these resul
ts lend insight into the nature of
the gender gap: Since the gap closes with a bachelor’s degree, t
here does not seem to be evidence
that women earn less doing the same jobs as men. It seems a mo
re likely explanation for the the
overall wage gap is that a disproportionate amount of men get hi
red for top paying positions.
Further investigation could involve using linear probability mod
els to test gender differences in
the probability of being employed in senior executive positions.
Global Finance Presentation - Content and Format Scoring
Criteria
10 Attributes
Scale
0 = not at all;
4 = very much
Content attributes
Background provided on the topic would enlighten an
uninformed listener.
0 1 2 3 4
Presentation reflected a thorough knowledge of relevant aspects
of the topic covered.
0 1 2 3 4
Presentation made appropriate use of International Financial
Management concepts to manage the risks inherent in global
financial exposures and operations.
0 1 2 3 4
Presentation provided a clear assessment of how and why firms

8. use International Financial Management techniques to create
value.
0 1 2 3 4
Presentation provided clear connections between the functional
topic and the student’s experience in the workplace or
marketplace.
0 1 2 3 4
Format attributes
The presentation was well organized with clear outline given.
0 1 2 3 4
Presenters spoke clearly and at an understandable pace.
0 1 2 3 4
Presenters were not overly dependent on PowerPoint slides (i.e.
avoid talking to the screen).
0 1 2 3 4
Main conclusions were effectively summarized.
0 1 2 3 4
Team responded effectively to questions.
0 1 2 3 4
1
Imprimante S.A.
Cross-Border Valuation and Parity Conditions
On June 23, 2008, a Monday morning, Martin Arnaud arrived at
his office in Imprimante S.A. corporate

9. headquarters in Paris, France. The previous week, Arnaud had
requested additional financial
information about an investment proposal from Imprimante-
Mexico, a wholly owned subsidiary that
operated a manufacturing facility and a regional sales office in
Monterrey, Mexico. The information had
arrived late Friday—too late for Arnaud to analyze—and was
waiting for him Monday morning. As a
financial analyst for a global manufacturer for printing an
imaging equipment, Arnaud examined many
cross-border projects, particularly since Imprimante had
accelerated it move into emerging markets
several years earlier.
The Mexican investment proposal called for the purchase and
installation of new automated machinery
to recycle and remanufacture toner and printer cartridges.
Cartridge recycling had become an
important part of Imprimante’s business in many markets and
promised continued growth. Many office
product retailers operated formal toner cartridge recycling
programs, for both the environmental
benefits of keeping materials out of landfills and demonstrated
cost savings for their customers. Writing

10. in a leading trade journal, one analyst predicted, “We are going
to see more and more refined
approached to recycling and remanufacturing (cartridges) in the
coming months and years…Both
corporate and individual consumers are becoming habituated to
it. They have simply come to expect
recycling as an option, even for smaller cartridges at lower
price points.”
Imprimante’s Monterrey plant began its cartridge recycling
program in 2005. The plant’s recycling
process consisted of a sequence of operations carried out almost
entirely by hand, with the help of hand
tools and a simple machine. The investment proposal called for
replacing this process with new
automated machinery from Germany that cost an estimated
MXP3.5 million (approximately
EUR220,000) fully installed. As described in the project
summary, Imprimante-Mexico expected to
realize substantial savings in labor and materials almost
immediately. Though the proposed expenditure
was relatively small, Imprimante required a discounted cash
flow analysis for all such investments in it
newer foreign markets and a review by corporate headquarters
in Paris. Arnaud was assigned to

11. perform an analysis of the investment proposal and make an “up
or down” recommendation to his
superior by Wednesday morning.
Imprimante S.A.
Imprimante was a global manufacturer of printers, copiers, fax
machines, and other document
production equipment. The company also provided consulting
and document outsourcing services, with
after-sales service contracts constituting about 18% of overall
revenue. Company sales for 2008 were
projected to be EUR3.35billion, down from 2007 due to global
recession. Operating profit was expected
to be EUR61.2million in 2008, and the company projected a
small net loss for the year. Exhibit 1
presents selected consolidated financial data for Imprimante.
2
Imprimante’s low profitability was typical of the industry in
2008; all of its competitors were similarly
affected by the recession. One bright spot in the company’s
outlook, however, was its growth in several
emerging markets, including the so-called BRIC economies of

12. Brazil, Russia, India, and China.
Imprimante had been a global firm for years, but did not move
aggressively into emerging markets until
2003-04. This was later than some of its competitors. On one
hand, this meant Imprimante’s market
hare lagged in some markets. On the other hand, Imprimante
avoided some of its competitors’ earlier
mistakes.
The company’s international operations were conducted
primarily through a large network of
subsidiaries, which operated mostly medium-sized regional
factories in which printers, copiers, and
other products were manufactured to suit local tasted.
Imprimante conducted business in 28 countries
around the world, with operations consisting of manufacturing
facilities, small research labs, as well as
sales and marketing subsidiaries. In 2008, subsidiaries outside
the European Union recorded about half
of Imprimante’s sales and generated slightly less than 40% of
pretax income.
Imprimante competed in a relatively mature market, and its
chief competitors were both established
multinational companies—some of which had developed their
consulting and other after-sales services

13. businesses to a higher level than had Imprimante—as well as
smaller players serving niche markets.
While Imprimante marketed and sold its products across the full
spectrum of industries, it had enjoyed
particular success in financial services, health care, and
government sectors.
Operations in Monterrey: Imprimante-Mexico
According to Imprimante’s CEO Alain Belmont, “We were
attracted to Mexico for the same reason we
built operations in Brazil and other emerging markets. We
wanted to diversify our operations and
believed we needed to establish a strong presence in places
besides Europe and the United States.” He
added: “Certainly there is risk (in these countries), but their
economies are dynamic and Imprimante
must be present. You can see our competitors feel the same
way.”
A key characteristic of Imprimante’s printing and imaging
products was their durability, which
Imprimante’s executives felt conveyed a competitive advantage
in emerging economies where
Imprimante positioned equipment as offering a lower total cost
of ownership. In particular, the

14. company’s marketing material claimed a working life of 10
months longer than its closest competitor,
with 30% lower service costs. CEO Belmont observed: “We
demonstrate to our customers that we have
a local presence and we are the lowest total-cost provider. This
creates loyalty and solid market
positions in Mexico and other of our newer markets.”
The manufacturing facility in Monterrey was located near a
small research and design facility, also
owned by Imprimante. While many product specifications for
Imprimante’s equipment were
formulated at the corporate offices in Paris, France, it was
customary for regional subsidiaries to
conduct fine-tuning research and design activity to tailor the
product more closely to local consumers’
preferences. Thus, it was common for a popular printer or fax
machine whose basic design was
conceived in Paris to be ”localized” for size, color, weight,
and/or range of features by local design staff.
Most of the products produced in the Monterrey plant were sold
in Mexico and were distributed
3

15. through large office-product retailers, department stores, as well
as small specialty shops.
Manufacturing inputs were source locally, and virtually all of
the plant’s employees were Mexican
citizens.
In the summer of 2008 gross output at Imprimante-Mexico was
running at only about 80% of planned
capacity. Nevertheless, plant records indicated that there was a
sizable increase in demand for recycled
printer and toner cartridges.. Imprimante-Mexico’s Programa de
Reciclaje de Cartuchos (Cartridge
Recycling Program) was started in 2005 to provide low-cost
recycling services to all its distributors and
customers. Under the terms of users’ service contracts, when
cartridges reached the end of their useful
lives, the could be returned to the Imprimante facility in
exchange for a significant discount on the
purchase of a like number of new cartridges. Imprimante
pledged to recycle and remanufacture all
returned toner and printer cartridges. Imprimante-Mexico also
had voiced its support for political
efforts to pass legislation that would mandate recycling of
printing cartridges used by most Mexican

16. businesses and government offices. In 2009, the company
planned to launch a pilot program to recycle
selected competitors’ cartridges.
As the number of cartridges returned for recycling increased,
Imprimante-Mexico management needed
to hire and train more employees to carry out the hole-piercing,
drilling, vacuuming, and toner/ink
evacuation required to recycle cartridges. “It’s taking more and
more of my payroll to handle recycling,”
said Beatrice Ernesto, the Monterrey plant manager. “We’re
happy to see the cartridges coming back
in, but the extra volume will become a problem when other
operation return to full capacity.”
Cost Savings from the Proposed New Equipment
The new equipment could process the Monterrey plant’s
projected volume using four employees rather
than ten, resulting in savings of both direct labor and training
costs. Under very favorable
circumstances, only three workers would be required. It would
also eliminate some human error, which
currently resulted in cracked or damaged cartridges which had
to be destroyed rather than reused. The
new equipment would occupy significantly less space in
Monterrey’s over-crowded plant; this space

17. would be freed up for other productive uses. It would also
require only minimal maintenance
expenditures compared to the equipment it replaced, and no
significant change in working capital.
Exhibit 2 compares projected operating data for the existing
recycling process and the proposed
automated process, assuming future Mexican inflation of 7% per
year.
The new equipment would have a useful life of 10 years and
would be depreciated under the straight-
line method for both tax and financial reporting purposes.
Salvage value was likely to equal disposal
costs at the end of the useful life. The manual equipment being
replaced was very simple and, properly
maintained, would last many more years. In June 2008 it had a
book value and tax basis of MXP250,000
and three years of straight-line depreciation remaining.
However, its market value was thought to be
lower, at about MXP 175,000. After considering Imprimante’s
consolidated tax position, Arnaud
determined that his analysis would use Mexico’s federal
corporate tax rate of 35%.
Real GDP growth in Mexico was 4.2% in 2004—the year in
which Imprimante built is Monterrey plant.

18. By 2006, Mexico’s real GDP growth had risen to 5.1%, but
subsequently dropped substantially as global
4
recession arrived. Other macroeconomic data in Mexico,
including bond yields, bank lending rates, and
the consumer price index exhibited similar patterns in recent
years. Exhibit 3 shows selected
macroeconomic and financial market data for Mexico.
Arnaud had yet to decide whether to perform the discounted
cash flow analysis in euros or pesos, or
indeed, whether NPV would be affected by the choice of
currency. Imprimante’s euro hurdle rate for
such a project, if undertaken in France, would be 8%. However,
borrowing costs in France and Mexico
were clearly different: French banks’ prime rate for euro loans
was 4.99%, while the rate in Mexico on
short-term peso loans was about 8.10%. Longer-term peso-
denominated corporate bonds were yielding
9.21%, compared with long-term euro-denominated corporate
issues at 4.75%. The spot exchange rate
on June 23 was MXP15.99/EUR. Many analysts were on record

19. predicting a real depreciation of the peso
against both the USD and the EUR over the next five years. For
example, one international business
publication noted “(Mexico’s) rising external financing
requirement and the fading impact of the US
stimulus package can only increase pressure on Mexico’s
currency.” The article went on to forecast a
rise in the MXP/EUR rate to 20.00 by 2011 and upwards of
25.00 in 2013-18. Selected macroeconomic
and financial market data for France are presented in Exhibit 4.
5
Exhibit 1 - Imprimante SA - Selected Consolidated Financial
Data (millions of
EUR, except as noted)
2008 2007 2006 2005 2004
Sales
3,345.3

20. 3,561.8
3,576.9
3,078.9
3,050.3
Operating income
61.2
189.2
172.9
163.5
149.9
Net income
(0.7)
85.7

21. 61.2
88.2
85.7
Total assets
2,809.3
2,764.9
2,899.6
3,129.0
2,445.5
Total debt
660.6
616.0

22. 613.0
578.4
504.2
Equity
782.6
819.5
829.7
941.0
865.1
Capital expenditures
87.6
100.0
95.1

23. 240.9
234.1
Depreciation
195.0
209.4
214.0
152.9
155.0
R&D expenditures
17.5
20.0
19.0

24. 48.2
46.8
Earnings/share (euros)
(0.0)
1.0
0.7
1.1
1.0
Dividend/share (euros)
0.7
0.7
0.7
0.7

25. 0.7
Return on sales (%) 0.0% 2.4% 1.7% 2.9% 2.8%
Return on equity (%) -0.1% 10.5% 7.4% 9.4% 9.9%
6
Exhibit 2 - Comparison of Projected Operating Data for
Different Recycling Processes (thousands of MXP, except as
noted)
Assumes 7% Inflation in Mexico Tax Rate: 0.35
2009 2010 2011 2012 2013 2014 2015 2016 2017 2018
Projected Operating Costs, Manual Process
Unit volume
(000s) 496 546 600 660 660 660 660 660 660 660
Materials
564,816
664,788

26. 782,456
920,951
985,417
1,054,396
1,128,204
1,207,178
1,291,681
1,382,099
Direct Labor
1,115,184
1,312,572
1,544,897

27. 1,818,343
1,945,627
2,081,821
2,227,549
2,383,477
2,550,321
2,728,843
Overhead
1,680,000
1,797,600
1,923,432
2,058,072
2,202,137

28. 2,356,287
2,521,227
2,697,713
2,886,553
3,088,611
Total
3,360,000
3,774,960
4,250,785
4,797,366
5,133,181
5,492,504

29. 5,876,980
6,288,368
6,728,555
7,199,553
Materials/unit
1.1387
1.2176
1.3041
1.3954
1.4931
1.5976
1.7094

30. 1.8291
1.9571
2.0941
Direct
labor/unit
2.2484
2.4040
2.5748
2.7551
2.9479
3.1543
3.3751
3.6113

31. 3.8641
4.1346
Projected Operating Costs, New Automatic
Process
Unit volume
(000s) 496 546 600 660 660 660 660 660 660 660
Materials
542,223
638,197
751,158
884,113
946,001
1,012,221
1,083,076

32. 1,158,891
1,240,014
1,326,815
Direct Labor
524,136
616,909
726,101
854,621
914,445
978,456
1,046,948
1,120,234
1,198,651

33. 1,282,556
Overhead
1,566,211
1,675,846
1,793,155
1,918,676
2,052,983
2,196,691
2,350,460
2,514,993
2,691,042
2,879,415

34. Total
2,632,570
2,930,952
3,270,414
3,657,410
3,913,429
4,187,368
4,480,484
4,794,118
5,129,707
5,488,786
Materials/unit

35. 1.0932
1.1689
1.2519
1.3396
1.4333
1.5337
1.6410
1.7559
1.8788
2.0103
Direct
labor/unit
1.0567
1.1299

36. 1.2102
1.2949
1.3855
1.4825
1.5863
1.6973
1.8161
1.9433
7
Exhibit 3 - Selected Macroeconomic and Financial Market Data
for Mexico

37. Year
Consumer Price
Inflation (%)
Real GDP
Growth (%)
Year-end Spot
Exchange Rate
(MXP/EUR)
2000 9.5% 6.6% 9.4
2001 6.4% -0.3% 9.5
2002 5.0% 0.9% 10.4
2003 4.3% 1.4% 12.9
2004 4.7% 4.2% 15.3
2005 3.3% 3.2% 13.3
2006 4.1% 5.1% 14.4
2007 3.8% 3.3% 16.2
Source: Mexico Country Reports, EIU
Date
Short-term
Bank Lending

38. Rate
JPMorgan
Mexico 7-10
Year Corporate
Bonds
10-year
Government Bonds
31-Mar-06 7.78% 8.20% 8.47%
30-Jun-06 7.68% 9.35% 9.06%
30-Sep-06 7.50% 8.22% 8.24%
31-Dec-06 7.60% 7.42% 7.42%
31-Mar-07 7.68% 7.50% 7.58%
30-Jun-07 7.82% 7.68% 7.19%
30-Sep-07 7.77% 7.86% 7.82%
31-Dec-07 8.00% 8.17% 8.08%
31-Mar-08 7.94% 7.42% 7.49%
30-Jun-08 8.10% 9.21% 9.12%
Sources: Bank of Mexico, Thomson Datastream, Global
Financial Data

39. 8
Exhibit 4 - Selected Macroeconomic and Financial Market Data
for France
Year
Consumer Price
Inflation (%)
Real GDP
Growth (%)
Year-end Spot
Exchange Rate
(MXP/EUR)
2000 1.7% 4.2% 9.4
2001 1.6% 2.1% 9.5
2002 1.9% 1.1% 10.4
2003 2.1% 0.5% 12.9
2004 2.3% 2.3% 15.3
2005 1.7% 1.9% 13.3

40. 2006 1.7% 2.4% 14.4
2007 1.5% 2.3% 16.2
Source: France Country Reports, EIU
Date
Short-term
Bank Lending
Rate
JPMorgan
France 7-10
Year Corporate
Bonds
10-year
Government Bonds
31-Mar-06 3.08% 3.73% 3.79%
30-Jun-06 3.27% 4.03% 4.08%
30-Sep-06 3.63% 3.69% 3.72%
31-Dec-06 4.07% 3.96% 3.98%
31-Mar-07 4.42% 4.08% 4.11%
30-Jun-07 4.69% 4.60% 4.62%

41. 30-Sep-07 4.91% 4.36% 4.41%
31-Dec-07 5.13% 4.34% 4.42%
31-Mar-08 4.81% 4.00% 4.11%
30-Jun-08 4.99% 4.75% 4.81%
Sources: Thomson Datastream, CEIC, Global Financial Data
Elementary Econometrics Project Spring 2017
Due April 13 by 5:00 pm The paper must be uploaded to
Carmen in a .pdf form.
The turn-it-in feature is enabled, meaning any paper that
plagiarizes from
another person who has ever completed this paper in my class
will be
flagged.
1 Basic Description
Objective:
The purpose of this class is to apply the analytical and
quantitative skills which should
be acquired in this course. The final project should be a
professional looking manuscript
with easily interpreted graphics and charts.
Description:
The project requires using statistical techniques learned in this

42. course to analyze data
from the 1997 National Longitudinal Survey of Youth. The
specific topic that you ex-
amine is up to you.
Formatting:
All aspects of your final paper must be typed. The paper should
be double spaced with
font of at least 11 point. There is no minimum number of pages.
There is a strict
maximum of 10 pages including title page, tables, graphs, and
appendices.
Anything beyond 10 pages will not be graded. The paper should
include the following
sections: introduction, data, empirical methodology, results, and
conclusions.
Data:
You will use the data set nlsy97 small.dta found on Carmen.
The data is a subsample of
the 1997 NLSY. The data set is funded by the BLS. They
describe the data as follows:
The NLSY97 consists of a nationally representative sample of
approximately 9,000
youths who were 12 to 16 years old as of December 31, 1996.
Round 1 of the survey
took place in 1997. In that round, both the eligible youth and
one of that youth’s par-
ents received hour-long personal interviews. In addition, during
the screening process,
an extensive two-part questionnaire was administered that listed
and gathered demo-
graphic information on members of the youth’s household and
on his or her immediate
family members living elsewhere. Youths are interviewed on an
annual basis.
The subset of data that you are given contains only 1199
observations, so that it can

43. be analyzed with small STATA. To avoid the complexity of
working with panel data,
the time aspect is taken out. There are literally thousands of
data points available, but
I have narrowed it to the variables shown in table 1.
2 Variable definitions
Table 1: Variable Definitions
Variable Name Description
grades8 Grades in 8th grade
gradeshs Grades in High School
health 1997 Self Reported Health as an adolescent
icecream 1997 Favorite Ice Cream Flavor as an adolescent
hardtimes 1997 Whether your family experienced ”Hard Times”
as a child”
census region 1997 The census region resided in as an
adolescent
income gross yr 1997 Household income of family when
respondent was an adolescent
msa 1997 MSA of residence as an adolescent
urban rural 1997 Urban or rural residency as an adolescent
race ethnicity Race and Ethnicity
asvab Armed Services Vocational Aptitude Battery percentile
marstat 2013 Marital and cohabitation status
job Have a job?
income Current (2013) income of the respondent

44. weight Weight (pounds)
health 2013 Current (2013) self reported health
momsedu Highest level of education achieved by mother
Many of the variables are categorical. Simply type ”tab
variable” and stata will display
how many observations belong to each category and what those
categories indicate.
3 Commonly missed Points
Grammar Do not simply write your paper in another language
and use google trans-
late to change it to English. Those papers are nearly impossible
to read. Reading the
paper out loud to yourself before you hand it in will help
tremendously.
Following Directions Include everything that is on the attached
grading rubric and
in the correct section.
Potential Problems with your regression All methodologies have
some potential
problems. When you address those problems in the methodology
section do not include
problems that you can fix with your current skills and data,
specifically discuss which
coefficients could be affected and the direction of the affect.
Equation The equation that is written out in the empirical
methodology section must
be the same equation used to generate the tables. If the equation
has perfect multi-
collinearity or treats categorical variables as continuous, there
will be large deductions.

45. Tables and graphs checklist Missing any part of the checklist
will result in significant
deductions.
• Tables and graphs must be understandable without reading the
text. This means
there should be a clear title that explains the contents of the
table, such as ”Sum-
mary Statistics”.
• Do not use variable names in tables and figures.For example,
rather than momedu
my table should say ”Mother’s Education”.
• There should also be a footnote which adds some detail to the
and explains what
is shown. For example ”Each entry is an OLS coefficient with
standard errors in
parentheses. Stars represent p-values as ...”
• If plotting distribution in a figure, use percent as the y-axis.
• All figures and tables should be referenced in the text.
• Figures must show more than just means.
• Cut and pasting from the stata output window will get 0
points. Use the estout
program to create the tables or type them in a word processing
program.
A condensed version of this form will be used to grade papers.
1. Formatting 10pts: The paper conforms to the formatting
standards
outlined in the assignment: Double spaced, font, margins, less

46. than 10 pages,
includes a title, and stapled.
2. Organization 10pts:
• Ideas are presented clearly, free of spelling and grammatical
errors. Any
references are cited. (-1 for each instance1)
• The paper is not cut and pasted from homeworks and
transitions well be-
tween sections. (5) (Not a list of tasks.2)
1Up to 20 points can be deducted
2If your paper is just a list of tasks and does not resemble a
research paper, you will be deducted
20 points.
• Each assigned section is included and labeled. (5)
3. Introduction 5pts: The question to be analyzed is described.
The
importance of the question is discussed. The data source is
mentioned. The
empirical methodology is mentioned. (OLS) A preview of
results is given
4. Data Section 15pts:
• The source of the data is mentioned. There is a brief
discussion of means,
standard deviations, etc. of the wage variable and whatever
other variable(s)

47. on which you choose to focus. (5)
• A summary table with means and standard deviations which is
referenced
in the text. (5)
• At least one graphic (histogram(don’t use density), scatter
plot, etc.) which
describes data and is referenced in the text.(5)
5. Empirical Methodology 15pts
• The question being analyzed is clearly described and the
estimated equation
is written out. (5)
• The inclusion of each variable is supported by theoretical
reasoning and
predicted signs of each variable are discussed. (5)
• Potential concerns with sample selection, omitted variables,
etc. are dis-
cussed along with the consequences of those problems. (5)
6. Results 30pts
• Results table. (7)
• Interpretation of coefficients (8) (You must interpret at least 4
coefficients
including 1 dummy variable and 1 coefficient with a logged
dependent vari-
able.)
• Hypothesis testing: discuss which coefficients are statistically
significant at

48. a 5% significance level; discuss the R2 of the regression;
conduct an F-test
on some aspect of the regression results. (5)
• Use your estimates to predict the earnings of some
hypothetical people. (3)
• Discuss the consistency of your results with your predictions.
(2)
• Estimate an alternative specification, present (in a table) and
discuss. (5)
7. Conclusion 5ptsThe conclusion summarizes results and the
signifi-
cance of those results.
8. Overall Quality 10 pts This is the only subjective portion of
your
grade. The average score will be 6/10.