ETC2410:

Assignment 1
Group Members: Abhinav Mathena, Afnan Nahyan, Parth Mahajan and
Yu Ooi

Group 12
Question 1 a)
i)
Estimated Equation:
lbwght = β + β 𝑐𝑖𝑔𝑠 + 𝑢

𝑙𝑏𝑤𝑔ℎ𝑡 = 4.769 − 0.004𝑐𝑖𝑔𝑠

(0.005) (0.001)

ii)
Null Hypothesis Testing:
𝐻 :β = 0
𝑯𝟏 : β < 𝟎
The 𝑡 can then be calculated as follows:

β −0.00491
t = = = −5.7
se β 0.000849

We will use the 𝑡 value calculated from EViews as it is to a greater accuracy, the above is
only to 6 decimal places. Hence 𝑡 = −5.2898

We know that the degree of freedom is given by:

~𝑡(𝑛 − 𝑘 − 1) ⇒ 𝑡(1388 − 1 − 1) ⇒ 𝑡(1386)
Using Excel =T.INV(0.05,1386) we get
𝑡 = −1.64595
Decision Rule:
𝑡 <𝑡
−5.7 < −1.646
Decision:
As the 𝑡 is smaller than the 𝑡 value, we can be 95% confident when rejecting the Null
Hypothesis and can conclude that there is not enough evidence to support that the regressor
cigs has no impact on the weight of the newly born infant.
iii)
Predicted average birth weight of infants born to women who smoke 0 cigs per day is given
by:

𝑙𝑏𝑤𝑔ℎ𝑡 = 4.769 − 0.004 ∗ 0

𝑙𝑏𝑤𝑔ℎ𝑡 = 4.769
𝑏𝑤𝑔ℎ𝑡 = 𝑒 . = 117.801
The predicted average birth weight 117.801 ounces

iv)
Predicted average birth weight of infants born to women who smoke 10 cigs per day is given
by:

𝑙𝑏𝑤𝑔ℎ𝑡 = 4.769 − 0.004 ∗ 10

𝑙𝑏𝑤𝑔ℎ𝑡 = 4.769 − 0.04
𝑙𝑏𝑤𝑔ℎ𝑡 = 4.729
𝑏𝑤𝑔ℎ𝑡 = 𝑒 . = 113.182

The predicted average birth weight 113.812 ounces

v)
We have conclusive evidence that smoking by the mother during pregnancy will REDUCE
the average birth weight of a newly born infant. This can be explained by looking at the
predicted values of a mother who does not smoke and a mother who smokes 10 cigarettes a
day. It is predicted that the mother who smokes 10 cigs a day will result on average the infant
weighting 4.619 ounces less than an infant for a mother who does not smoke.
b)
I)
𝑙𝑏𝑤𝑔ℎ𝑡 = 4.688 + 0.053𝑤ℎ𝑖𝑡𝑒 + 0.026𝑚𝑎𝑙𝑒 − 0.004𝑐𝑖𝑔𝑠 + 0.008𝑙𝑓𝑎𝑚𝑖𝑛𝑐
(0.019) (0.013) (0.010) (0.001) (0.006)

II)
Based on the output obtained we see that the sample value of the F statistic for testing the
joint significance of the regressions is 15.025, with a p-value of 0.000. Since the p-value is
zero, we reject the null hypothesis that the regressions are jointly insignificant at any
significance level.
III)
From appendix b, we see that the regression family income is individually insignificant, with
a p-value of 0.1597 at the 10% significance level.
IV)
From appendix b(I), we see that the regressions family income is individually insignificant,
with a p-value of 0.1597 at the 10% significance level.
V)
–Controlling for gender, the average number of cigs smoked per day by the mother during
pregnancy and family income, we have no strong reason to believe that whether the mother is
a person of colour will have any effect on the weight of her newly born baby than that of a
white mother. Therefore, we have no strong prior belief regarding the sign of 𝛽 .
– Controlling for the mother’s race, the average number of cigarettes smoked per day by the
mother during pregnancy and family income, it seems plausible that male babies have higher
birthweights as males have greater bone density in comparison to females. Therefore, our
prior expectation is that 𝛽 is positive which aligns with the regression output.
– Controlling for the mother’s race, gender, and family income it seems plausible that the
number of cigarettes smoked by the pregnant mother during pregnancy will have an adverse
effect on the baby's health due to the known negative effects of cigarettes. Therefore, our
prior expectation is that 𝛽 is negative which aligns with the regression output.
– Controlling for the mother’s race, the gender of the newly born male and the average
number of cigs smoked per day by the mother during pregnancy, it seems plausible that
birthweight will be higher for mothers with greater family incomes as they have greater
access to food and healthcare. Therefore, our prior expectation is that 𝛽 is positive which
aligns with the regression output.
VI)
From appendix b:
𝑅 = 0.04
Therefore, the explanatory variables explain about 4% of the sample variation in lbwght
across newly born babies.
VII)

𝛽 gives the estimated average percentage change in baby birth weight of 0.008%, if there is
a 1% change in family income, holding all else constant.
VIII)
According to the regression output the most significant influence on the weight of a newborn
baby, is the number of cigarettes smoked by the mother during pregnancy. Since this has a p
value of 0 it will pass significant tests at all levels.
c)
i)
Original Equation:

𝑙𝑏𝑤𝑔ℎ𝑡 = β + β 𝑤ℎ𝑖𝑡𝑒 + β 𝑚𝑎𝑙𝑒 + β 𝑐𝑖𝑔𝑠 + β 𝑙𝑓𝑎𝑚𝑖𝑛𝑐

𝑙𝑏𝑤𝑔ℎ𝑡 = 4.688 + 0.053𝑤ℎ𝑖𝑡𝑒
(0.013)
+ 0.026𝑚𝑎𝑙𝑒 − 0.004𝑐𝑖𝑔𝑠
(0.001)
+ 0.008𝑙𝑓𝑎𝑚𝑖𝑛𝑐
(0.019) (0.010) (0.006)

𝐻 :β =β =0
Restricted Equation:

𝑙𝑏𝑤𝑔ℎ𝑡 = β + β 𝑚𝑎𝑙𝑒 + β 𝑐𝑖𝑔𝑠

𝑙𝑏𝑤𝑔ℎ𝑡 = 4.757 + 0.024𝑚𝑎𝑙𝑒 − 0.004𝑐𝑖𝑔𝑠
(0.008) (0.010) (0.001)
Null Hypothesis Testing
𝐻 :β =β =0
𝐻 : β ≠ 0𝑎𝑛𝑑/𝑜𝑟β ≠ 0
𝑠𝑖𝑔𝑛𝑖𝑓𝑖𝑐𝑎𝑛𝑐𝑒 𝑙𝑒𝑣𝑒𝑙 = α = 0.05
Calculating 𝐹
𝑆𝑆𝑅 − 𝑆𝑆𝑅 𝑛 − 𝑘 − 1
𝐹 =
𝑆𝑆𝑅 𝑞
49.216 − 48.320 1388 − 4 − 1
𝐹 = = 12.583
49.216 2
Calculating 𝐹
𝐹 = 𝐹(𝑞, 𝑛 − 𝑘 − 1) = 𝐹(2,1383)
Using excel F.INV.RT(sig,num_DOF,den_DOF) => F.INV.RT(0.05,2,1383)
𝐹 = 3.00231
Decision Rule:
𝑅𝑒𝑗𝑒𝑐𝑡 𝐻 𝑖𝑓: 𝐹 >𝐹
Decision:
12.583 > 3.00231
Because the 𝐹 value is greater than 𝐹 , we are 95% confident that there is sufficient
evidence in the sample to reject the Null Hypothesis which states that the variables “white”
and “lfaminc” are jointly insignificant, hence the variables “white” and “lfaminc” should not
be removed from the regression model.
ii)

We can see that the 𝑙𝑏𝑤𝑔ℎ𝑡 of a female baby with a white mother who does not smoke, and
has a family income of $100,000 would result in the baby weighing:

𝑙𝑏𝑤𝑔ℎ𝑡 = 4.688 + 0.053 ∗ (1) ∗ 0.026 ∗ (0) − 0.004 ∗ (0) + 0.008 ∗ (ln 100)
𝑙𝑏𝑤𝑔ℎ𝑡 = 4.779
Now holding all coefficients but “white” constant, will allow us to find the weight of the
baby coming from a coloured family.

𝑙𝑏𝑤𝑔ℎ𝑡 = 4.688 + 0.053 ∗ (0) ∗ 0.026 ∗ (0) − 0.004 ∗ (0) + 0.008 ∗ (ln 100)
𝑙𝑏𝑤𝑔ℎ𝑡 = 4.726
By holding the gender, the number of cigs consumed, as well as the family income constant,
we can see that the birth of a while baby is 0.053 ounces greater than if the baby was to be
coloured. This goes in favour of our findings in part I, where we found that β ≠
0 𝑎𝑛𝑑/𝑜𝑟β ≠ 0 as the null hypothesis was rejected. By holding all variables constant, we
can see that the race of the baby has some impact on the weight of the baby.
d)

Answer (d)-(i): 𝛽 gives the percentage change in average baby weight, depending on the
average number of cigarettes smoked per day by the mother during pregnancy.

Answer (d)-(ii): Holding all other factors constant, 𝛽 gives a decrease of 0.4% in average
birth weight, for each additional cigarette smoked per day by the mother on average during
pregnancy. For every 1 cigarette smoked, we can predict the weight of the baby to decrease by
𝑒 . = 1.004 ounces.

Answer (d)-(iii): The coefficient 𝛽 is -0.004, which means that while there is a decrease in
birth weight for every additional cigarette smoked, the magnitude of this decrease might be
considered small in practical terms. However, when considering the potential cumulative effect
of smoking multiple cigarettes daily over the course of a pregnancy, the overall impact could
be substantial.

Answer (d)-(iv):

𝑙𝑏𝑤𝑔ℎ𝑡 = 4.688 + 0.053𝑤ℎ𝑖𝑡𝑒 + 0.026𝑚𝑎𝑙𝑒 − 0.004𝑐𝑖𝑔𝑠 + 0.008𝑙𝑓𝑎𝑚𝑖𝑛𝑐

(0.019) (0.013) (0.010) (0.001) (0.006)
Therefore,
𝑙𝑏𝑤𝑔ℎ𝑡 = 4.688 + 0.053(1) + 0.026(1) − 0.004(20) + 0.008 log(50)
𝑙𝑏𝑤𝑔ℎ𝑡 = 4.72
𝑏𝑤𝑔ℎ𝑡 = 112.168
the predicted average birth weight of a male child, born to a white mother, who smokes 20
cigarettes per day and who has a family income of $50,000 will be 112.168 ounces.

Answer (d)-(v): The t-value that we get from the table for 𝑡( ) degrees of freedom is
1.96. Now, deriving a 95% confidence interval for 𝛽 using the standard formula is given
below:
CI = coefficient ± (1.98 × standard error)

For 𝛽 :
Lower Limit: −0.004 − (1.98 × 0.001) = −0.00598
Upper Limit: −0.004 + (1.98× 0.001) = −0.00202

Therefore, 95% CI for 𝛽 is (-0.00598, -0.00202)

Answer (d)-(vi): As the standard of error of 𝛽 is 0.001, which indicates how much 𝛽 would
differ from the actual population parameter 𝛽 if we were to take multiple samples from the
population. Since a smaller standard of error typically indicates that the estimator is more
precise, we can presume that 𝛽 is a precise estimate of 𝛽 .
e)
Answer (e)-(i):
The Population Model for Quadratic Function:
𝑙𝑏𝑤𝑔ℎ𝑡 = 𝛽 + 𝛽 𝑤ℎ𝑖𝑡𝑒 + 𝛽 𝑚𝑎𝑙𝑒 + 𝛽 𝑐𝑖𝑔𝑠 + 𝛽 𝑐𝑖𝑔𝑠 + 𝛽 𝑙𝑓𝑎𝑚𝑖𝑛𝑐 + 𝑢
The Estimated Equation:

𝑙𝑏𝑤𝑔ℎ𝑡 = 𝛽 + 𝛽 𝑤ℎ𝑖𝑡𝑒 + 𝛽 𝑚𝑎𝑙𝑒 + 𝛽 𝑐𝑖𝑔𝑠 + 𝛽 𝑐𝑖𝑔𝑠 + 𝛽 𝑙𝑓𝑎𝑚𝑖𝑛𝑐

𝑙𝑏𝑤𝑔ℎ𝑡 = 4.691
(0.019)
+ 0.055𝑤ℎ𝑖𝑡𝑒
(0.013)
+ 0.025𝑚𝑎𝑙𝑒
(0.010)
− 0.007𝑐𝑖𝑔𝑠 + 0.000122𝑐𝑖𝑔𝑠
(0.002) (6.85E-05)
+ 0.008𝑙𝑓𝑎𝑚𝑖𝑛𝑐
(0.006)
Null Hypothesis Testing:
𝐻 :𝛽 = 0
𝐻 :𝛽 ≠ 0
𝑆𝑖𝑔𝑛𝑖𝑓𝑖𝑐𝑎𝑛𝑐𝑒 𝐿𝑒𝑣𝑒𝑙: 𝛼 = 0.10
Now, to calculate tcalc we can use the following formula:

𝛽 0.000122
𝑡 = = = 1.781
𝑠𝑒(𝛽 ) 0.0000685

The tcalc given by the EViews is 1.788 ≈ 1.781

Additionally, to calculate tcrit we need to know the degrees of freedom which is,
𝑡 ⟹ 𝑡 ⟹𝑡
To determine the critical value of test statistic, we used conservative approach and took tcrit at
the 10% significance level for 120 degrees of freedom so that we can have more confidence
on our decision since higher tcrit makes it harder to reject the null. Therefore, the critical value
of test statistic with 1382 degrees of freedom that we get from the table is,
tcrit = 1.658
Decision Rule:
Since,
𝑡 > tcrit
We reject the null, in favour of the alternate hypothesis.
Given that the p-value by EViews for 𝛽 is 0.0739.
Also, we know that if
𝛼 > p-value
then, we reject the null in favour of the alternate hypothesis.
Hence, we can conclude with 90% confidence that there's sufficient evidence in the sample
data to indicate that the relationship between the number of cigarettes smoked by the mother
during pregnancy (𝑐𝑖𝑔𝑠) and the natural logarithm of the birth weight of the newborn (𝑙𝑏𝑤𝑔ℎ𝑡)
might possess a quadratic form. Nevertheless, we can also emphasize that the coefficient for
the quadratic term, though statistically significant, is very close to zero. Therefore, the impact
of the negative term in the equation will always be stronger than the effect of the positive
quadratic term. This denotes that while the quadratic effect exists, it's practically very weak.
Also, if we do the hypothesis testing at less than 10% significance level, then we have 𝛼 < p-
value and the quadratic relationship becomes insignificant, and we accept the null hypothesis
that the 𝑐𝑖𝑔𝑠 has no quadratic effect on 𝑙𝑏𝑤𝑔ℎ𝑡.
Answer (e)-(ii):
The Estimated Equation:

𝑙𝑏𝑤𝑔ℎ𝑡 = (0.019)
4.691 + 0.055𝑤ℎ𝑖𝑡𝑒
(0.013)
+ (0.010)
0.025𝑚𝑎𝑙𝑒 − 0.007𝑐𝑖𝑔𝑠 + 0.000122𝑐𝑖𝑔𝑠
(0.002) (6.85E-05)
+ 0.008𝑙𝑓𝑎𝑚𝑖𝑛𝑐
(0.006)
Now, holding all else constant the predicted effect on lbwght from increasing cigs from 10 to
15 is:
Δ𝑙𝑏𝑤𝑔ℎ𝑡 = 𝛽 (𝑐𝑖𝑔𝑠 − 𝑐𝑖𝑔𝑠 ) + 𝛽 (𝑐𝑖𝑔𝑠 − 𝑐𝑖𝑔𝑠 )
Δ𝑙𝑏𝑤𝑔ℎ𝑡 = − 0.007(15 − 10) + 0.000122(15 − 10 )
Δ𝑙𝑏𝑤𝑔ℎ𝑡 = −0.01975
|Δ𝑙𝑏𝑤𝑔ℎ𝑡| = 0.01975
Δ𝑏𝑤𝑔ℎ𝑡 = 0.980
Considering the model controls for gender, race, and income, a mother increasing her cigarette
consumption from 10 to 15 cigarettes per day during pregnancy is predicted to decrease the
birth weight of their child by 0.980 ounces. Therefore, it can be concluded that keeping other
things constant if a mother increases her average cigarette consumption per day from 10 to 15,
the average birth weight will decrease by 0.980 ounces.
Answer (e)-(iii):
When controlling for all other variables, we must first derive the partial derivative of cigs and
𝑐𝑖𝑔𝑠 from the estimated model:

𝑙𝑏𝑤𝑔ℎ𝑡 = 4.691
(0.019)
+ 0.055𝑤ℎ𝑖𝑡𝑒
(0.013)
+ 0.025𝑚𝑎𝑙𝑒
(0.010)
− 0.007𝑐𝑖𝑔𝑠
(0.002)
+ 0.000122𝑐𝑖𝑔𝑠
(6.85E-05)
+ 0.008𝑙𝑓𝑎𝑚𝑖𝑛𝑐
(0.006)
𝜕
= −0.007𝑐𝑖𝑔𝑠 + 0.000122𝑐𝑖𝑔𝑠
𝜕𝑐𝑖𝑔𝑠
𝜕
= −0.007 + 0.000244𝑐𝑖𝑔𝑠
𝜕𝑐𝑖𝑔𝑠
Now solve for cigs, where cigs has 0 marginal effect on birthweight:
0 = −0.007 + 0.000244𝑐𝑖𝑔𝑠
0.000244𝑐𝑖𝑔𝑠 = 0.007
𝑐𝑖𝑔𝑠 = 28.689
From this result, we can infer that taking 28 or less cigarettes will result in a negative
marginal effect on birthweight.
f)i)

𝑙𝑏𝑤𝑔ℎ𝑡 = 𝛽 + 𝛽 𝑤ℎ𝑖𝑡𝑒 + 𝛽 𝑚𝑎𝑙𝑒 + 𝛽 𝑐𝑖𝑔𝑠 + 𝛽 𝑙𝑓𝑎𝑚𝑛𝑖𝑐 + 𝛽 𝑐𝑖𝑔𝑠 ∗ 𝑤ℎ𝑖𝑡𝑒 + 𝑢

𝑙𝑏𝑤𝑔ℎ𝑡 = 4.684 + 0.061𝑤ℎ𝑖𝑡𝑒
(0.014)
+ 0.026𝑚𝑎𝑙𝑒 − 0.002𝑐𝑖𝑔𝑠
(0.002)
+ 0.008𝑙𝑓𝑎𝑚𝑛𝑖𝑐 − 0.003𝑐𝑖𝑔𝑠
(0.002)
(0.020) (0.010) (0.006)
∗ 𝑤ℎ𝑖𝑡𝑒
f)ii)
𝐻 : 𝛽 = 05
𝐻 :𝛽 ≠ 0
𝑆𝑖𝑔𝑛𝑖𝑓𝑖𝑐𝑎𝑛𝑐𝑒 𝐿𝑒𝑣𝑒𝑙: 𝛼 = 0.10
p-value = 0.0604
Decision Rule: Reject 𝐻 if p-value < 𝛼
Decision: Since 0.0604 < 0.10, there is enough evidence to reject 𝐻 in favour of 𝐻 .
Conclusion: Thus, it can be said with 90% confidence that the average marginal effect of
cigarette smoking on birth weight varies by weight, holding all else equal
f)iii)
At the 5% significance level, the regressor cigs is statistically insignificant. With the use of a
p-value test, where the decision rule is that it is statistically significant if the p-value < 𝛼, the
value of cigs is 0.2617, which is greater than the 5% significance level. Thus, we are 95%
confident that the cigs regressor is statistically insignificant, holding all else equal.
g)i)
Population A: 𝑙𝑏𝑤𝑔ℎ𝑡 = 4.684 + 0.061(0) + 0.026(1) − 0.002(10) + 0.008(𝑙𝑛50) −
0.003(0)
𝑙𝑏𝑤𝑔ℎ𝑡 = 4.684 + 0.026 − 0.02 + 0.031
𝑙𝑏𝑤𝑔ℎ𝑡 = 4.721
𝑏𝑤𝑔ℎ𝑡 = 112.314
The average male birth weight is 112.314 ounces when the mother is a person of colour, has a
family income of $50,000 and smokes 10 cigarettes a day.
g)ii)
Population B: 𝑙𝑏𝑤𝑔ℎ𝑡 = 4.684 + 0.061(1) + 0.026(0) − 0.002(5) + 0.008(𝑙𝑛50) −
0.003(1 ∗ 5)
𝑙𝑏𝑤𝑔ℎ𝑡 = 4.684 + 0.061 − 0.01 + 0.031 − 0.015
𝑏𝑤𝑔ℎ𝑡 = 4.751
𝑏𝑤𝑔ℎ𝑡 = 115.700
The average female birth weight is 115.700 ounces when the mother is white, has a family
income of $50,000 and smokes 5 cigarettes a day.
g)iii)
Under the null:
𝛽 + 𝛽 (0) + 𝛽 (1) + 𝛽 (10) + 𝛽 (𝑙𝑜𝑔50) + 𝛽 (0)
= 𝛽 + 𝛽 (1) + 𝛽 (0) + 𝛽 (5) + 𝛽 𝑙(𝑙𝑜𝑔50) + 𝛽 (5)
𝛽 (1) + 𝛽 (10) = 𝛽 (1) + 𝛽 (5) + 𝛽 (5)
𝛽 = 𝛽 − 𝛽 (5) + 𝛽 (5)
We denote:
𝛿 = 𝛽 − 𝛽 + 𝛽 (5) − 𝛽 (5)
Which implies:
𝛽 = 𝛽 − 𝛽 (5) + 𝛽 (5) + 𝛿
We use this to reparametrize the model:
𝑙𝑏𝑤𝑔ℎ𝑡 = 𝛽 + 𝛽 𝑤ℎ𝑖𝑡𝑒 + (𝛽 − 𝛽 (5) + 𝛽 (5) + 𝛿)𝑚𝑎𝑙𝑒 + 𝛽 𝑐𝑖𝑔𝑠 + 𝛽 𝑙𝑓𝑎𝑚𝑛𝑖𝑐
+ 𝛽 𝑐𝑖𝑔𝑠 ∗ 𝑤ℎ𝑖𝑡𝑒 + 𝑢
𝑙𝑏𝑤𝑔ℎ𝑡 = 𝛽 + 𝛽 (𝑤ℎ𝑖𝑡𝑒 + 𝑚𝑎𝑙𝑒) + 𝛿𝑚𝑎𝑙𝑒 + 𝛽 (𝑐𝑖𝑔𝑠 − 5𝑚𝑎𝑙𝑒) + 𝛽 𝑙𝑓𝑎𝑚𝑛𝑖𝑐
+ 𝛽 (5𝑚𝑎𝑙𝑒 + 𝑐𝑖𝑔𝑠 ∗ 𝑤ℎ𝑖𝑡𝑒) + 𝑢

𝑙𝑏𝑤𝑔ℎ𝑡 = 𝛽 + 𝛽 (𝑤ℎ𝑖𝑡𝑒 + 𝑚𝑎𝑙𝑒) + 𝛿 𝑚𝑎𝑙𝑒 + 𝛽 (𝑐𝑖𝑔𝑠 − 5𝑚𝑎𝑙𝑒) + 𝛽 𝑙𝑓𝑎𝑚𝑛𝑖𝑐

+ 𝛽 (5𝑚𝑎𝑙𝑒 + 𝑐𝑖𝑔𝑠 ∗ 𝑤ℎ𝑖𝑡𝑒)
g)iv)
Unrestricted Model:
𝑙𝑏𝑤𝑔ℎ𝑡 = 𝛽 + 𝛽 𝑤ℎ𝑖𝑡𝑒 + 𝛽 𝑚𝑎𝑙𝑒 + 𝛽 𝑐𝑖𝑔𝑠 + 𝛽 𝑙𝑓𝑎𝑚𝑛𝑖𝑐 + 𝛽 𝑐𝑖𝑔𝑠 ∗ 𝑤ℎ𝑖𝑡𝑒 + 𝑢
Restricted Model:

𝑙𝑏𝑤𝑔ℎ𝑡 = 𝛽 + 𝛽 (𝑤ℎ𝑖𝑡𝑒 + 𝑚𝑎𝑙𝑒) + 𝛿 𝑚𝑎𝑙𝑒 + 𝛽 (𝑐𝑖𝑔𝑠 − 5𝑚𝑎𝑙𝑒) + 𝛽 𝑙𝑓𝑎𝑚𝑛𝑖𝑐

+ 𝛽 (5𝑚𝑎𝑙𝑒 + 𝑐𝑖𝑔𝑠 ∗ 𝑤ℎ𝑖𝑡𝑒)
𝐻0: 𝛿 = 0
𝐻0: 𝛿 > 0
𝑆𝑖𝑔𝑛𝑖𝑓𝑖𝑐𝑎𝑛𝑐𝑒 𝐿𝑒𝑣𝑒𝑙: 𝛼 = 0.05

𝛿
𝑡 = ~𝑡 = 1382 𝑢𝑛𝑑𝑒𝑟 𝐻0
𝑠𝑒 𝛿
−0.026474
𝑡 = = −1.288048
0.020554
𝑡 = 1.658
Decision: Since -1.288<1.658 there is not enough evidence to reject the null in favour of the
alternative hypothesis.
Conclusion: Due to the 𝑡 of -1.288 being less than the 𝑡 of 1.658, there is not enough
evidence to reject the hypothesis that the average birthweight of male infants in population A
is the same as the average birthweight of female infants in population B, in favour of the
hypothesis that the average birthweight of male infants in population A is greater than the
average birthweight of female infants in population B.
Appendix
Appendix(a):

Appendix(b):
Appendix(c):

Appendix(e):
Appendix(f):

Appendix(g):