PART 1 MULTIPLE REGRESSION applied to all data sets dependent variable (DV) adolescent fertility rate

Adolescent Fertility RatePART 1: MULTIPLE REGRESSION Applied to all data sets: Dependent variable DV: Adolescent fertility rate Independent variables IV: - Gross National Income per capi

Trang 1

Adolescent Fertility Rate

RMIT University Vietnam ECON 1193 – Business Statistics 1

0

Trang 2

Multiple Regression 2

All Countries (ALL) 2

Low-Income Countries (LI) 4

Middle-Income Countries (MI) 6

High-Income Countries (HI) 9

Team Regression Conclusion 11

Time Series 12

Brazil 12

Kenya 14

Vietnam 17

Ireland 19

Team Time Series Conclusion 22

Overall Conclusion 23

Reference 24

1

Trang 3

PART 1: MULTIPLE REGRESSION

Applied to all data sets:

Dependent variable (DV): Adolescent fertility rate

Independent variables (IV):

- Gross National Income per capita (GNI)

- Compulsory education (years)

- Domestic general government health expenditure per capita PPP (current international $)

- Life expectancy at birth (Years)Significance level α = 0.05

All countries (ALL)

Regression output 1 (All variables)

Regression Statistics

Multiple R

R SquareAdjusted R SquareStandard ErrorObservations

Significant variables: Life Expectancy at birth

Insignificant variables: Gross National Income per capita (Highest P-value), Compulsory education, Domestic general government health expenditure per capita

2

Trang 4

Regression output 2 (Exclude GNI per capita)

Multiple R

Intercept

Compulsory education

PPP

Life expectancy

Significant variables: Life expectancy at birth

Insignificant variables: Compulsory education and Domestic general government health expenditure per capita (Highest P-value)

Regression Output 3 (Exclude PPP)

Trang 5

Regression Equation:

AFR = 358.382 + 4.166 (Compulsory Education) – 4.845 (Life Expectancy at birth)

Interpretation

- For every year a citizen receives as compulsory education year, the number

of infants given by teenage moms will increase by 4.166 births

- For every year a newborn is expected to live, the number of babies of young women will decrease by 4.845

- r2=0.723 is coefficient of determination 72.3% of the variation in the number of births given by teenagers at the age of 15-19 could be clarified by the variation in the Life Expectancy at birth and Compulsory education years

Low-income countries (LI)

YEA

Country NameR

Multiple R

Intercept

4

Trang 6

Significant Variables: None

Insignificant Variables: Gross National Income per capita, Compulsory education (Highest P-value), Domestic general government health expenditure per capita and Life expectancy at birth

Regression Output 2 (Exclude Compulsory education)

Multiple R

Intercept

GNI

Life expectancy

PPP

Insignificant Variables: Gross National Income per capita, Domestic general government health expenditure per capita (Highest P-value) and Life expectancy at birth.

Regression Output 3 (Exclude PPP)

Multiple R

R SquareAdjusted R Square

5

Trang 7

Intercept

GNI

Life expectancy

Insignificant Variables: Gross National Income per capita (Highest P-value) and

Life expectancy at birth

Regression 4 (Exclude GNI)

Intercept

Life expectancy

Insignificant Variables: Life expectancy at birth

There is no variables that has a significant linear relationship with Adolescent Fertility Rate.Therefore, it is impossible to come up with a final regression model for Low-income countries

Middle-income Countries (MI)

YEA Country Name Country Adolescent GNI Compulsory PPP Life

Trang 8

2015201520152015

Regression Output 1 (All variables)

Trang 9

Significant Variables: Compulsory education

Insignificant Variables: Gross National Income per capita (Highest P-value), Domestic general government health expenditure per capita and Life expectancy at birth.

Regression Output 2 (Exclude GNI)

Intercept

Life expectancy

PPP

Significant Variables: Compulsory education

Insignificant Variables: Domestic general government health expenditure per capita and Life expectancy at birth (Highest P-value)

Regression output 3 (Exclude Life expectancy)

Intercept

8

Trang 10

- For every dollar the government spends for healthcare treatment for citizens, numbers of newborns given by teenagers will decrease by 0.044 births This means for every $1000 that each individual receive from the government as budget for healthcare, Adolescent Fertility rate decrease by 4.4 births.

- r2=0.547 is coefficient of determination 54.7% of the variation in the number of births given by young moms aged 15-19 can be explained by the variation in the Compulsory education years and Domestic general government health expenditure per capita

High-income Countries (HI)

YEAR201520152015

201520152015

Trang 11

9

Trang 12

Regression Output 1 (All variables)

Multiple R

Insignificant Variables: Gross National Income per capita (Highest P-value), Compulsory education, Domestic general government health expenditure per capita and Life expectancy at birth

Regression Output 2 (Exclude GNI)

0.71Multiple R

R Square

Trang 13

10

Trang 14

Adjusted R SquareStandard ErrorObservations

Intercept

Life expectancy

PPP

Insignificant Variables: Compulsory education (Highest P-value), Domestic general government health expenditure per capita and Life expectancy at birth

Regression Output 3 (Exclude Compulsory education)

Multiple R

Intercept

Life expectancy

PPP

11

Trang 15

Significant Variables: Domestic general government health expenditure per

capita Insignificant Variables: Life expectancy at birth

Regression Output 4 (Exclude Life Expectancy)

Multiple R

Intercept

PPP

Significant Variables: Domestic general government health expenditure per

capita Insignificant Variables: None

Interpretation

- In high-income countries, for every dollar that citizens receive from the

government as a healthcare treatment welfare, the number of babies given by

adolescent girls will decrease by 0.002 births This indicates that per every $1000 the government assists people with healthcare, number of infants of teenagers will decrease by 2 births

- r2=0.414 is coefficient of determination 41.4% of the variation in the number of

births given by teenagers aged 15-19 can be explained by the variation in the

Domestic general government health expenditure per capita

PART 2: TEAM REGRESSION CONCLUSION

- Each model has different independent significant variables Nevertheless, Domestic general government health expenditure per capita is the significant variables in both

Middle-income and High-income countries In Low-income countries, no variables is found to have a significant linear relationship with Adolescent Fertility Rate.

12

Trang 16

- Model of All Countries provide the best Adolescent Fertility rate estimation The reason for

this is since the model has a high number of observations, it will be more accurate Also, its

variables explain the highest percentages of Adolescent Fertility Rate (72.3%)

- In conclusion, it is seen that Government Health Expenditure is an important factor

that should be focused on in order to improve the AFR in middle-income and

high-income countries For low-high-income countries, further investigation should be conducted

to identify other determinants of AFR

PART 3: TIME SERIES

MADSSE

- Quadratic Trend Model (QUA)

Multiple R

R SquareAdjusted R SquareStandard Error

Trang 17

Intercept

Time period (x)

Square of time period (x^2)

Coefficient s

77.9561.028-0.047

Quadratic Equation:

Prediction & Errors

QUA predictionReality

ErrorAbsolute Error

- Exponential trend model (EXP)

Intercept

Time period (x)

14

Trang 18

ErrorAbsolute Error

Conclusion: Exponential model is suggested to predict the Adolescent Fertility Rate of

Brazil since its errors (MAD and SSE) are smallest (5.194 and 110.155, correspondingly)

Trang 19

Prediction & Errors

LIN predictionReality

ErrorAbsolute Error

Multiple R

R SquareAdjusted RSquareStandard ErrorObservations

ErrorAbsolute Error

16

Trang 20

-Exponential trend model (EXP)

ErrorAbsolute Error

MADSSE

Conclusion: Exponential model is suggested to predict the Adolescent Fertility Rate in

Kenya as its errors (MAD and SSE) are smallest (2.847 and 33.130, correspondingly)

17

Trang 21

Multiple R

18

Trang 22

Prediction & error:

-Exponential trend model (EXP)

Intercept

Time period (x)

or

19

Trang 23

Prediction & error:

EXP predictionReality

ErrorAbsolute error

Conclusion: The Linear Model is suggested to predict the Adolescent fertility rate in Vietnam since its error is the lowest (MAD 2.059 and SSE 24.073 correspondingly)

Trang 24

ErrorAbsolute error

MAD 4.695SSE 89.012

Multiple R

Intercept

Time period (x)

Square of time period (x^2)

Quadratic equation: ^ =¿ 17.214 + 0.104(year) - 0.005(year)2

AFR

Prediction & error:

QUA predictionReality

ErrorAbsolute error

21

Trang 25

- Exponential trend model (EXP)

ErrorAbsolute error

MAD

SSE

2013 2014 201515.920 15.860 15.801

11.555 11.082 10.608

4.365 4.778 5.192

4.57284.457

Conclusion: The Quadratic Model is suggested to predict the Adolescent fertility rate in Ireland since its error is the lowest (MAD 3.427 and SSE 47.216 correspondingly)

22

Trang 26

PART 4: TEAM TIME SERIES CONCLUSION

Adolescent fertility rate (births per 1,000 women ages 15-19) from 1980 to 2015

Overall, except for Vietnam, there is a downward trend among all the remaining countries; however, each line of each country has different figure Over the period of 35 years, while Brazil and Ireland both underwent an insignificant decrease (from 78.00 to 63.75 births, from 20.19 to 10.60 births respectively), Kenya drop dramatically from 166.91 to 83.09 births In contrary, Vietnam slightly increased from 20.28 to 29.03 births.

CountryBrazilKenya

23

Trang 27

IrelandVietnam

- Exponential trend model is followed by both Brazil and Kenya While Irelandfollows Quadratic trend model, Vietnam follows Linear trend

- According to the calculation, it is suggested that Linear Trend is the most suitable trend model to predict Adolescent fertility rate The SSE and MAD of

Vietnam is the lowest compare to other countries with different trend model which indicates the Linear trend’s accuracy

PART 5: OVERALL TEAM CONCLUSION

- According to 2 recent reports, it is suggested that there is a relationship between Income

level and Adolescent fertility rate Nevertheless, this relevance becomes

less significant through two researches While from assignment 1 (Figure 1), it can be seen that the relationship between income and AFR is certain, the assignment 2 suggested that GNI only explain approximately 39% of AFR

P(Low AFR | HI) P(Low AFR | MI) P(Low AFR | LI)

Figure 1: Result from Assignment 1

- From our regression test result, Domestic general government health

expenditure per capita appears in 2 out of 3 final models of all countries Moreover, the final model of Low-income countries only have Life expectancy at birth and of Middle-income one is include Compulsory education beside PPP Nonetheless, only Life expectancy and PPP are in agreement with the Hypothesis Testing from

assignment 2 as Compulsory Education only explains around 1% of Adolescent Fertility Rate

On the other hand, since there is no variable which has a significant relationship with AFR, more investigation should be done to find other factors

Country Kenya Brazil Ireland Vietnam

Figure 2:

Prediction of AFR in 2020 and 2030

24

Trang 28

- From figure 2, it suggested that Vietnam, Brazil and Ireland will experience the

upward trend between 2016 and 2020: Vietnam (29.038 to 36.249 births), Brazil (62.677 to 67.965 births), Ireland (10.135 to 12.9 births), meanwhile AFR of Kenya decrease

(81.791 to 71.947 births) From 2020 to 2030, AFR of Kenya, Brazil and Ireland decline, while Vietnam rise from 36.249 to 39.287 births

In our perspective, the ability of United Nation accomplishing their goal (SGD 5) is

insignificantFirstly, although the world’s Adolescent fertility rate decrease annually, the overall trend does not apply to several countries (World Bank 2017) (For example: Vietnam, Iraq) It is reported that 1 out of 5 women and girls have suffered physical or sexual

violence However, 49 countries have not yet come up with the laws to protect

women from such violence (United Nation) Lack of protection from the government would lead to the unwanted births at the early ageSecondly, although child marriage has been decreasing, the progress is not enough to reach 2030’s target It is predicted if the trend continues, between 2017 and 2030, 150 million will get married before 18, which is still a huge number

Overall, further method should be adopted to improve the situation such as raising government expenditure on healthcare in middle and high income countries and invest deeper in factors that have been affecting AFR in low-income countries

REFERENCE

1.United Nation n.d., Goal 5: Achieve gender equality and empower all women and girls,

United Nation, viewed 6 Jan 2018, < equality/?fbclid=IwAR2ACvboq71ZMX4GcVrpBmG8LV-

https://www.un.org/sustainabledevelopment/gender-pYXszjkgv_BgyssNgZJ3XVrxsytAnWSA >

1 Unicef 2018, Progress for every child in the SGD area 2018, Unicef, New York

25

Tiêu đề	PART 1 MULTIPLE REGRESSION applied to all data sets dependent variable (DV) adolescent fertility rate
Trường học	RMIT University Vietnam
Chuyên ngành	Business Statistics
Thể loại	Thesis
Thành phố	Hanoi

Định dạng
Số trang	28
Dung lượng	464,67 KB