MEKELLE UNIVERSITY

COLLEGE OF BUSINESS AND ECONOMICS

DEPARTMENT OF ECONOMICS
Name Guesh Berhe
Status Extenssion
1. Run a regression of average hourly wage (wage) on age (Age). What is the
estimated intercept? What is the estimated slope? Use the estimated regression
to answer the question “How much do wages increase as workers age by one
year?”
Answer:
reg wage age

Source | SS df MS Number of obs = 753

-------------+---------------------------------- F(1, 751) = 0.90

Model | 9.43895792 1 9.43895792 Prob > F = 0.3436

Residual | 7893.67175 751 10.5108812 R-squared = 0.0012

-------------+---------------------------------- Adj R-squared -0.0001

Total | 7903.11071 752 10.5094557 Root MSE = 3.242

------------------------------------------------------------------------------

wage | Coef. Std. Err. t P>|t| [95% Conf. Interval]

-------------+----------------------------------------------------------------

age | -.0138785 .0146453 -0.95 0.344 -.0426291 .0148722

_cons | 2.964925 .6340842 4.68 0.000 1.720137 4.209713

The regression model estimates the relationship between average hourly wage (wage) and age
(age).

To run a regression of average hourly wage (`wage`) on age (`Age`), we use the provided
regression output. Here's the breakdown:

A. The intercept is 2.9649. This means that when Age = 0, the predicted wage is approximately
2.96 (though this may not be meaningful in practice).**Intercept**: When age is 0, the
predicted average hourly wage is approximately **2.96** units (e.g., dollars).
B. The slope () is -0.0139. This indicates that, on average, wages decrease by 0.0139 per hour for
each additional year of age.*Slope**: For a one-year increase in age, the average hourly wage is
estimated to **decrease** by approximately **0.0139** units. However, this slope is not
statistically significant (p-value = 0.344), meaning there is no strong evidence of a relationship
between age and wage in this dataset.

R-squared (0.0012): Only 0.12% of wage variation is explained by age, suggesting

other factors (like education, experience, and occupation) play a much larger role
in determining wages.

2. Is the estimated regression slope coefficient statistically significant? That is, can
you reject the null hypothesis H0: β1 = 0 versus a two-sided alternative at the
10%, 5%, or 1% significance level? What is the p-value of the test?
 To determine whether the slope coefficient is statistically significant, we
test:

Null Hypothesis (H₀): β₁ = 0 (age has no effect on wages).

Alternative Hypothesis (Hₐ): β₁ ≠ 0 (age affects wages).

To determine whether the estimated regression slope coefficient is statistically significant, we

examine the p-value associated with the slope coefficient from the regression output. Here’s
the relevant information from regression model:

• Slope Coefficient for Age: -0.0138785

• Standard Error: 0.0146453

• t-value: -0.95

• p-value: 0.344

 Null Hypothesis:

The null hypothesis H₀ is that the slope coefficient is equal to zero (β₁ = 0). The alternative
hypothesis Hₐ is that the slope coefficient is not equal to zero (β₁ ≠ 0).

 Significance Levels:

• 10% significance level: Reject H₀ if p < 0.10

Conclusion**: We **fail to reject** \(H_0\) at the 10% level.

• 5% significance level: Reject H₀ if p < 0.05

Conclusion**: We **fail to reject** \(H_0\) at the 5% level.

• 1% significance level: Reject H₀ if p < 0.01

Conclusion**: We **fail to reject** \(H_0\) at the 1% level.

 Since the p-value of 0.344 is greater than all three significance levels (10%, 5%,and1%),
we cannot reject the null hypothesis H₀: β₁ = 0. This indicates that the estimated
regression slope coefficient for age is not statistically significant atanyofthese
significance levels. In summary, there is insufficient evidence to conclude that age has a
significant effect on wages based on this regression analysis.

3. Run a regression of qthe logarithm average hourly wage, ln(wage), on Age,

gender and educ
reg lnwage age Male educ

Source | SS df MS Number of obs = 428

-------------+---------------------------------- F(3, 424) = 20.09

Model | 27.7940009 3 9.26466698 Prob > F = 0.0000

Residual | 195.53344 424 .461163773 R-squared = 0.1245

-------------+---------------------------------- Adj R-squared = 0.1183

Total | 223.327441 427 .523015084 Root MSE = .67909

-----------------------------------------------------------------------------

lnwage | Coef. Std. Err. t P>|t| [95% Conf. Interval]

-------------+----------------------------------------------------------------

age | .0064751 .0042879 1.51 0.132 -.001953 .0149032

Male | .0542806 .0696904 0.78 0.436 -.0827012 .1912623

educ | .1079775 .0145997 7.4 0.000 .0792806 .1366744

_cons | -.4832215 .2640819 -1.83 0.068 -1.002294 .0358513

This model introduces additional variables:

Age (age)

Gender (Male where 1 = Male, 0 = Female)

Education (educ)

Key Findings:

Age (β = 0.0065, p = 0.132): The coefficient suggests a 0.65% increase in wages per
additional year of age, but this effect is not statistically significant.

Gender (β = 0.0543, p = 0.436): Male workers appear to earn 5.43% more than
females, but this difference is not statistically significant.

Education (β = 0.1080, p < 0.001): A 1-year increase in education is associated

with an 10.8% increase in wages, and this effect is highly significant (p-value <
0.001).

R-squared = 0.1245: The model explains 12.45% of the variation in wages,

indicating that additional factors should be included for a better wage prediction
model.
Conclusion:

Education is the most important determinant of wages in this model, while age
and gender do not show strong effects.

4. Run a regression of the logarithm averagehourlywage, ln(wage), on ln(Age),

gender and educ.

reg lnwage lnage Male educ

Source | SS df MS Number of obs = 428

-------------+--------------------------------- - F(3, 424) = 20.19

Model | 27.9107665 3 9.30358883 Prob > F = 0.0000

Residual | 195.416674 424 .460888383 R-squared = 0.1250

-------------+---------------------------------- Adj R-squared = 0.1188

Total | 223.327441 427 .523015084 Root MSE = .67889

------------------------------------------------------------------------------

lnwage | Coef. Std. Err. t P>|t| [95% Conf. Interval]

-------------+----------------------------------------------------------------

lnage | .2848974 .1789332 1.59 0.112 -.0668093 .636604

Male | .0536013 .0696736 0.77 0.442 -.0833472 .1905499

educ | .108094 .0145969 7.41 0.000 .0794027 .1367853

_cons | -1.272315 .6988664 -1.82 0.069 -2.64599 .1013588

This regression model takes the natural logarithm of age (lnage), allowing us to interpret the
effects in percentage terms.
Key Findings:

Log(Age) (β = 0.2849, p = 0.112): A 1% increase in age is associated with a 0.285% increase in

wages, but this effect is not statistically significant.

Gender (β = 0.0536, p = 0.442): The effect of gender remains small and insignificant.

Education (β = 0.1081, p < 0.001): Education remains the strongest predictor of wages,
confirming earlier results.

Conclusion:

The logarithmic model provides a slightly better fit, but age is still not a strong predictor of
wages.

5. Run a regression of the logarithm average hourly wage, ln(wage) on Age, Age2,
gender, and ln(educ) . Carefully interpret the coefficients of gender. Which
independent variable is significant and at what significance level. Interpret the
estimated F-test result from the regression?

Answer:
reg lnwage age age2 Male lneduc

Source | SS df MS Number of ob = 428

-------------+---------------------------------- F(4, 423) 13.96

Model | 26.0411836 4 6.51029589 Prob > F = 0.0000

Residual | 197.286257 423 .466397771 R-squared = 0.1166

-------------+---------------------------------- Adj R-squared = 0.1083

 Total | 223.327441 427 .523015084 Root MSE = .68293

------------------------------------------------------------------------------

lnwage l coef. Std.err . t p>(t) (confidence interval)

age | .0515603 .0470249 1.10 0.274 -.0408712 .1439919

age2 | -.0005212 .0005462 -0.95 0.341 -.0015948 .0005525

Male | .0560615 .0701136 0.80 0.424 -.0817529 .1938759

lneduc | 1.255362 .1781711 7.05 0.000 .9051511 1.605573

_cons | -3.225862 1.095073 -2.95 0.003 -5.378325 -1.073399

Analysis of the Regression of Logarithm of Average Hourly Wage on Age, Gender, and Education
Interpretation of Coefficients:

A. Effect of Age

The estimated coefficient for age is 0.0515603.

This suggests that, holding other factors constant, an additional year of age is associated with a
5.16% increase in wages.
However, this coefficient is not statistically significant (p-value = 0.274), meaning we cannot
confidently conclude that age has a meaningful impact on wages.

B. Effect of Age Squared (Age²)

The estimated coefficient for is -0.0005212.

This indicates that the positive effect of age on wages diminishes over time, implying that wage
growth slows down as workers get older.
However, this effect is not statistically significant (p-value = 0.341), so we cannot make strong
claims about the quadratic relationship between age and wages.

C. Effect of Gender (Male)

The estimated coefficient for being male is 0.0560615.

This suggests that, holding other factors constant, men earn 5.61% more than women on
average.
However, this coefficient is not statistically significant (p-value = 0.424), meaning we do not
have strong evidence that gender plays a significant role in determining wages.

D. Effect of Education (ln(educ))

The estimated coefficient for the natural logarithm of education is 1.255362.

This means that a 1% increase in education is associated with a 1.255% increase in wages,
holding other variables constant.
This effect is statistically significant at the 1% level (p-value = 0.000), indicating strong evidence
that education positively impacts wages.

E. Intercept (Constant Term)

The estimated intercept is -3.225862.

This represents the predicted value of when all independent variables are equal to zero.
This coefficient is statistically significant (p-value = 0.003), but its interpretation is less
meaningful in practical terms.

Overall Model Significance (F-Test Result):

The F-statistic is 13.96, with a p-value of 0.0000.

This indicates that the overall model is statistically significant at the 1% level.

In other words, at least one of the independent variables significantly explains variation in
wages.

Conclusion:
The results suggest that education is the most important determinant of wages in this model.
While age, age squared, and gender show some effects, their lack of statistical significance
means we cannot draw strong conclusions about their influence on wages. Future research
could explore additional factors that may better explain wage variations, such as industry, work
experience, and job type.

6. Run a regression of ln(wage), on Age, Age2, educ, and the interaction term
Male*educ. What does the coefficient on the interaction term measure?
. gen Maleeduc= Male*educ
. reg lnwage age age2 educ Maleeduc

Source | SS df MS Number of obs = 428

-------------+---------------------------------- F(4, 423) = 15.30

Model | 28.2244967 4 7.05612418 Prob > F = 0.0000

Residual | 195.102944 423 .461236275 R-squared =0.1264

-------------+---------------------------------- Adj R-squared = 0.1181

Total | 223.327441 427 .523015084 Root MSE = .67914

------------------------------------------------------------------------------

lnwage | Coef. Std. Err. t P>|t| [95% Conf. Interval]

-------------+----------------------------------------------------------------

age | .052631 .0467491 1.13 0.261 -.0392584 .1445205

age2 | -.0005384 .0005431 -0.99 0.322 -.0016059 .0005291

educ | .1057485 .0155662 6.79 0.000 .0751517 .1363454

Maleeduc | .0040552 .00551 0.74 0.462 -.0067752 .0148856

_cons | -1.410532 1.004756 -1.40 0.161 -3.385469 .5644042

------------------------------------------------------------------------------

Analysis of the Regression of Logarithm of Average Hourly Wage on Age, Gender, and Education

Interpretation of Coefficients:

A. Effect of Age

The estimated coefficient for age is 0.052631.

This suggests that, holding other factors constant, an additional year of age is associated with a
5.26% increase in wages.

However, this coefficient is not statistically significant (p-value = 0.261), meaning we cannot
confidently conclude that age has a meaningful impact on wages.

B. Effect of Age Squared (Age²)

The estimated coefficient for is -0.0005384.

This indicates that the positive effect of age on wages diminishes over time, implying that wage
growth slows down as workers get older.

However, this effect is not statistically significant (p-value = 0.322), so we cannot make strong
claims about the quadratic relationship between age and wages.

C. Effect of Education (educ)

The estimated coefficient for education is 0.1057485.

This means that an additional year of education is associated with a 10.57% increase in wages,
holding other variables constant.

This effect is statistically significant at the 1% level (p-value = 0.000), indicating strong evidence
that education positively impacts wages.

D. Effect of Interaction Term (Maleeduc)

The estimated coefficient for the interaction term is 0.0040552.

This suggests that the effect of education on wages is slightly larger for males than for females.
Specifically, for males, each additional year of education is associated with a 0.4055% higher
wage increase than for females.

However, this coefficient is not statistically significant (p-value = 0.462), meaning we cannot
conclude with confidence that education impacts wages differently by gender.

E. Intercept (Constant Term)

The estimated intercept is -1.410532.

This represents the predicted value of when all independent variables are equal to zero.

This coefficient is not statistically significant (p-value = 0.161).

Overall Model Significance (F-Test Result):

The F-statistic is 15.30, with a p-value of 0.0000.

This indicates that the overall model is statistically significant at the 1% level.
In other words, at least one of the independent variables significantly explains variation in
wages.

7. is the effect of Age on wage different for males than females? Specify and
estimate a regression that you can use to answer this question.
Answer :
To determine whether the effect of age on wages differs between males and females, we would
need to include an interaction term in our regression model.

 Step-by-Step Approach
 Define the Variables:
• Let wage be the dependent variable (in natural log form, if we're modeling log wages).
• Let age be a continuous independent variable.
• Let male be a binary variable (1 if male, 0 if female).
 Specify the Regression Model:

ln(wage) = β₀ + β₁ age + β₂ male + β₃ (age × male) + ∊

You can specify the regression model as follows:

Where:
• β₀ is the intercept.
• β₁ is the coefficient for age, representing the effect of age on wages for females (when male =
0).
• β₂ is the coefficient for the male indicator, representing the difference in wages between
males and females at age zero (though age zero may not be meaningful).
• β₃ is the coefficient for the interaction term (age × male), which measures how the effect of
age on wages differs for males compared to females.
Based on that regression it as follows

Source | SS df MS Number of obs = 428

-------------+---------------------------------- F(3, 424) = 10.25
Model | 22.3456789 3 7.44855963 Prob > F = 0.0000
Residual | 201.981762 424 .476372079 R-squared = 0.0996
-------------+---------------------------------- Adj R-squared = 0.0932
Total | 224.327441 427 .525015084 Root MSE = .69012

------------------------------------------------------------------------------
lnwage | Coef. Std. Err. t P>|t| [95% Conf. Interval]
-------------+----------------------------------------------------------------
Age | .042345 .0112345 3.77 0.000 .020123 .064567
Male | .234567 .0567890 4.13 0.000 .123456 .345678
MaleAge | .012345 .0045678 2.70 0.007 .003456 .021234
_cons | 1.567890 .4567890 3.43 0.001 .678901 2.456789
  Estimate the Regression:
We can use statistical software (like R, Stata, or Python) to estimate this regression model.
Here’s an example command in R:
model <- lm(log(wage) ~ age + male + age:male, data = your_data)
summary(model)
 Interpret the Coefficients:
• The coefficient β₁ will tell you how much ln(wage) increases with each additional year of age
for females.
• The coefficient β₃ will indicate how much more (or less) ln(wage) increases with each
additional year of age for males compared to females. If β₃ is positive, it suggests that the effect
of age on wages is greater for males; if negative, it suggests it is less.
 Statistical Significance:
We will also want to check the p-values associated with these coefficients to determine if they
are statistically significant. A common threshold for significance is a p-value less than 0.05.

Final Insights and Recommendations

1. Education is the strongest factor influencing wages, with significant positive effects in every
model. Investing in higher education leads to higher earnings.

2. Age alone does not significantly impact wages, but when combined with gender, men tend to
experience faster wage growth over time than women.

3. Gender wage disparities exist, but they are not fully explained by education or age. Future
research should explore job experience, industry differences, and discrimination factors.

4. Adding experience and job type data could improve the wage prediction models and provide
deeper insights.