NGUYEN MINH DUC – S3977946

ASSIGNMENT 2

INDIVIDUAL CASE STUDY

Maternal Mortality – Data Set #5

Subject Code: ECON 1193

Subject Name: Business Statistics

Student Name: Nguyen Minh Duc

Student ID: S3977946

Tutorial Group: Group 02

Tutorial Lecturer: Duc LM

Word Count: 3000

NGUYEN MINH DUC – S3977946

Part A: Introduction
According to The World Health Organization (n.d.), the Maternal Mortality Ratio (MMR) is the
number of maternal deaths during a given time period per 100,000 live births during the same
time period. It reflects the risk of a woman demise due to complications from pregnancy or
childbirth (UNICEF 2023). Unfortunately, the maternal mortality level is still unacceptably high
despite experiencing a 34% decline globally between 2000 and 2020 period (WHO n.d.). In
2020, the global MMR was 223:100,000 live births which was more than triple the target set by
United Nation in 2030 (WHO n.d.).

Since some health indicators are used to assessed either the overall health care system or
different aspects of a country’s structure, maternal mortality is accepted as a key index of health
and socioeconomic development (Sima et al. 2015). Determining MMR is necessary for
indentifying issues or evaluating the effectiveness of existing systems and take actions if needed
(DHS n.d.). For example, a low MMR indicates adequate healthcare services and overall
maternal well-being and vice versa. Hence, reducing MMR is part of the 17 Goals of Sustainable
Development Goals, being included under SDG 3: Good Health and Well-being (UN n.d.). The
Goals is set in order to solve global poverty and other deprivations by improving health along
with education, mitigating inequality and stimulating economic growth (UN n.d.). By 2030, the
United Nation aims to reduce the worldwide maternal mortality ratio to smaller than 70:100,000
per live births (UN n.d.). As a consequence, countries will need to improve their healthcare
system and the well-being of women population which benefit the society’s overall health (SDG
3).

Because of the MMR significance, researchers have conducted various researches in order to
examine the correlations of MMR with gross national income (GNI) along with other indexes. It
is found that maternal mortality ratio and the GNI capita have a significantly inverse correlation
(Tadele and Wasie 2017). Correspondingly, the Pearson correlation also shows maternal
mortalities and GNI to be moving in the opposite direction (Jean and Adrien 2020). This means
as GNI increases, the number of maternal mortalities decreases since higher-income countries
can invest for a more adequate healthcare system. Data reveals that the lifetime risk of maternal
death for a 15-year-old girl is just 1:5,300 in high-income countries while it is 1:49 in low-
NGUYEN MINH DUC – S3977946

income countries (UNICEF 2023). However, MMR also depends on difference socio-economic
factors; thus, GNI per capita is not directly proportional to the MMR. For example, regardless
having a higher GNI per capita compared to France (The World Bank n.d.), the American’s
MMR is considerably higher than French’s (Roosa et al. 2020).

Part B: Descriptive Statistics and Probability

A high Maternal Mortality Ratio is higher than 50.

Maternal Mortality Ratio

High (A) Low (A’) Total
Low-Income
4 0 4
(C)
Middle-Income
Country’ 12 12 24
(C’)
s income
High-Income
0 12 12
(C’’)
Total 16 24 40
Figure 1: Contingency table of Maternal Mortality Ratio and Country’s income

Probability of a high Maternal Mortality Ratio for a random country:

P(A) = 16/40 = 40%

Probability of a high Maternal Mortality Ratio for:

- A Low-Income country: P(A|C) =4/4= 100%

- A Middle-Income country: P(A|C’) =12/24 =50%
- A High-Income country: P(A|C’’) =0/12 =0%

1. Statistically independent events

NGUYEN MINH DUC – S3977946

P(A) and P(A|C) are different, thus Maternal Mortality Ratio and GNI of High-Income countries
are dependent events and statistically dependent. The event of a country being Low-Income if
occurred will influence the probability of A.

Similarly, P(A) is different with P(A|C’) and P(A|C’’); therefore, Maternal Mortality Ratio and
GNI of Middle and Low-Income countries are dependent events and statistically dependent.

The probability of a high MMR in a random country is different when considering their income.
As a result, it can be concluded that income and Maternal Mortality Ratio are dependent events
and statistically dependent.

2. Higher MMR probability

P(A|C) > P(A|C’) > P(A|C’’)
The probabilities show that low-income countries are more likely to have a high Maternal
Mortality Ratio. As none of the high-income countries experienced a high MMR, while the
middle-income countries have a 50% probability and all of the low-income countries witnessed a
high MMR. In conclusion, the lower the income category, the higher the possibility of a country
to have high MMR.

3. Descriptive Statistics
a. Outliers

Min >,<,= Lower Bound Max >,<,= Upper Bound Outlier

LI 353 > 336.125 725 < 849.125 0
MI 7 > -48.25 602 > 155.75 4
HI 4 > -2.5 18 > 17.5 1
Figure 2: Table of outliers test

There are 4 outliers in middle-income category and 1 outlier in high-income countries while low-
income group witnessed 0 outliers.

b. Central Tendency and Dispersion

NGUYEN MINH DUC – S3977946

Measure of Central Tendency

LI MI HI
Mean 574.75 111.875 7.917
Median 610.5 51.5 7
Mode 30 5

Figure 3: Table of Measure of Central Tendency (per 100,000 live births)

The mean can be heavily influenced by extreme values (Pennsylvania n.d.) since it accounts for
all values in the data set to measure central tendency (PennState n.d.). Hence, it is not suitable in
this case despite being the most popular measure. Moreover, the mode of low-income category
of the sample is not detected due to no repetition of value; therefore, it cannot be used to
compared between different categories. As a result, the median is the most appropriate measure
of central tendency in this analysis due to the insensitiveness with outliers.

Variation
LI MI HI
Interquartile Range
128.25 51 5
(IQR)

Figure 4: Table of Variation (per 100,000 live births)

Similarly, outliers can affect measures such as variance and standard deviation by distort the
interpretation of data (ncl n.d.). Additionally, the range only displays the simplest span of data, it
cannot adequately express the variation (Mark et al. 2015). Thus, IQR is the best measure for this
analysis.

Cooperatively, both measures can give reader a comprehensive understanding of the center and
dispersion of a dataset.

c. Interpretation
NGUYEN MINH DUC – S3977946

The median of high-income group is the lowest at 7 deaths per 100,000 people while the
numbers are 51.5 and 610.5 for middle and low-income. Respectively, those are 7.35 - and 87-
times bigger compared to the high-income median, showing a significant difference. This pattern
correlates with the calculated probabilities in previous part.

The IQR is the range of middle 50% of the data set, illustrating the variation of the data around
its median (Mark et al. 2015). Initially, low-income group’s IQR can be seen as the biggest
spread of the middle 50% data and high-income IQR as the smallest. However, since the ranges
of 3 categories are different, it is also important to consider the scale of the data sets.
Correspondingly, the IQR of the low-income fluctuated around the median within a 10% radius
compared to its scale. Identically, the numbers for middle- and high-income are 49.51% and
35.71%. Hence, showing the middle 50% of low-income countries actually spread the least while
its peer in middle-income group spreads the most compared to the median if the difference in
data scale is considered.

Integrating both measures together, they show consistent results with earlier calculations in
probability part. Regarding the low-income countries, its median and IQR implies that 75% of
their data is higher than 50, being high Maternal Mortality Ratio. Similarly, the possibility are
50% higher than 50 for middle-income countries while 75% of high-income group is certain to
be lower than 50 deaths per 100,000 live births, hence reinforcing that lower-income countries
are more likely to have a higher Maternal Mortality ratio.

d. Box-and-whisker plot

Graph Skewness Calculation

LI Negatively-skewed 3(Mean-Median)/Standard Deviation=-0.67

MI Positively-skewed 3(Mean-Median)/Standard Deviation=1.07

HI Positively-skewed 3(Mean-Median)/Standard Deviation=0.68

NGUYEN MINH DUC – S3977946

Figure 5: Box-plots of 3 income groups

The box-plot provides a graphical distribution of the data, considering the five-number summary
and showing the range, interquartile range and quartiles (Mark et al. 2015).

Looking closer, it can be seen that the middle-income group has the widest range while the high-
income range is significantly smaller than its counterparts by observing the graph length. This is
due to the fact that there are 4 outliers being higher than the upper bound in middle-income
group, hence its right whisker length is drastically longer and being disproportionate. The same
applies to high-income graph although the effect is more subtle since the outlier is slightly over
the upper bound. The difference in central tendency can also be observed through the distant
median of each graph.

Moreover, the difference in IQR can be visualized by the distinctive sizes of the boxes. The
boxes in low-income are the longest, following by the middle-income then the high-income. This
is consistent with the values of the IQRs in previous calculations.

Additionally, the diagrams describe the data skewness. For low-income graph, the values are
skewed to the right, showing the half data is concentrated at concerning MMR which is even
higher than the biggest MMR of the middle-income group. In contrast, high-income diagram is
skewed to the left; therefore, most of the high-income countries has an under 10 ratio.

The 3 box-plots on the same axis also illustrate the probability of having high maternal mortality
ratio for each countries group. For example, the high-income diagram is completely under the 50
value, thus there is no country having high MMR or the probability is 0%. Similarly, it can be
NGUYEN MINH DUC – S3977946

seen the probability for low-income category is 100% since all of the values is higher than 50.
Lastly, the number is 50% for middle-income category due to half of the graph being higher than
50. One more time shows the persistence with earlier analysis.

Figure 6:

Figure 7:

Figure 8:
NGUYEN MINH DUC – S3977946

Part C: Confidence Intervals

1. Calculation

The confidence level is assumed to be 95% for the calculation of world Maternal Mortality ratio
confidence intervals.

Significance level α 0.05

Confidence level (1- α) × 100% 95 %
Population standard σ
deviation
Sample standard S 209.259 deaths per 100,000
deviation live births
Sample mean 𝑋 126.975 deaths per 100,000
live births
Sample size n 40 country
Degree of freedom d.f. 39
t-critical value t 2.0227

As the sample size is greater than 30, the Central Limit Theorem is applicable, hence the
sampling distribution is normally distributed. Moreover, the t student table will be used since
population standard deviation is unknown.

Confidence interval:

𝑆 209.259
𝑋±t∝/2,d.f. × 𝑛
= 126.975 ± 2.0227 × 40
= 126.975 ± 66.924

→ 60.050 ≤ μ ≤ 193.899

It is 95% confident that the true mean of world Maternal Mortality ratio is between 60.050 and
193.899 deaths per 100,000 live births.

2. Assumption
NGUYEN MINH DUC – S3977946

Despite the population standard deviation is unknown, the sample size is greater than 30;
therefore, the Central Limit Theorem is applicable. The CLT stated that regardless of the shape
of population, the sampling distribution becomes almost normal as the sample size gets large
enough (OpenStax n.d.). Thus, no assumption is required for the calculation.

3. Impacts on confidence interval outcomes when the population standard deviation is

known

Knowing the world standard deviation of Maternal Mortality ratio, the calculation will use the Z-
table instead of the t-table and does not need to substitute by using the sample standard
deviation.

Although both population and sample deviation measure variability, there are differences in
certainty due to their nature. The population standard deviation is a fixed value calculated from
individuals in the total population. However, a sample standard deviation is calculated from
individuals in a specific sample of the population which alter depending on that sample (Taylor
2019). Hence, it has greater variability and introduces extra uncertainty. For that reason, the t-
distribution has a ‘fatter’ tails and less in the center than the Z distribution despite both being
bell-shaped and symmetric (Mark et al. 2015). As sample size increases, S becomes a better
estimate of the σ and the t-distribution gradually approaches the standardised normal
distribution, thus decreasing the confidence intervals equaling to a more accurate result.

In conclusion, if the world standard deviation of Maternal Mortality ratio is known, the
confidence interval will decrease to a more accurate result.
NGUYEN MINH DUC – S3977946

Part D: Hypothesis Testing

1. Hypothesis Testing

In 2014, the world average Maternal Mortality ratio was 221 deaths per 100,000 live births.
Meanwhile, it is 95% confident that the global average MMR is in between 60.050 and 193.899
deaths based on previous calculation. Therefore, the MMR is forecasted to decrease in the future.

Significance level α 0.05

Confidence level (1- α) × 100% 95 %
Population standard σ
deviation
Sample standard S 209.259 deaths-per-100,000-
deviation live-births
Population mean μ 221 deaths-per-100,000-
live-births
Sample mean 𝑋 126.975 deaths-per-100,000-
live-births
Sample size n 40 country

Step 1: Check for CLT

The sample size is greater than 30, hence the Central Limit Theorem is applicable and the
sampling distribution is normally distributed.

Step 2: Determine the null and alternative hypothesis

The null hypothesis H0: μ ≥ 221

The alternative hypothesis H1: μ < 221 (the forecast)

Step 3: Identify the test

α = 0.05 and n = 40
NGUYEN MINH DUC – S3977946

The rejection region is only in one tail and the alternative hypothesis contains “<”, thus the
lower-tailed test is suitable for the test.

Step 4: Determine table

The population standard deviation is unknown, thus the t-table will be used.

Step 5: Determine Critical Value (CV)

α = 0.05, hence the critical t value for left-tailed is -1.685

Step 6: Calculate test statistic t

𝑋―μ 126.975 ― 221

tn-1 = 𝑆
= 209.259
=-2.842
𝑛 40

Step 7: Make statistical decision

tn-1 = -2.842 < -1.685 = t, thus the test falls in the rejection region. Therefore, H0 is rejected and
H1 is accepted.

Step 8: Make managerial decision

As H1 is accepted, it is 95% level of confidence that the world Maternal Mortality ratio will
decrease in the future.

Step 9: Possible errors

NGUYEN MINH DUC – S3977946

Since H0 is rejected, it is possible that the test has made Type I error by rejecting a true null
hypothesis (Mark et al. 2015). This means the world Maternal Mortality ratio may have a 5%
probability to remain unchanged or even increase instead of decreasing in the future.

2. Potential impacts of a reduction of the dataset by half

Reducing the dataset by half means the sample size will decrease by 50% down to 20.

Firstly, since the sample size is smaller than 30, the CLT will not be applicable and assumption
that the population is normally distributed will need to be made (Mark et al. 2015).

Secondly, as the degree of freedom will also decrease, sample standard deviation becomes a
worse estimate of population standard deviation, resulting in the t distribution becomes flatter
and getting further away than the Z distribution (Mark et al. 2015). The bigger sample standard
distribution creates greater error variance and less power (Sage n.d.), thus decreasing the
accuracy of the test. Moreover, less power also causes a higher probability of making type II
error when it is failed to reject a false null hypothesis (Mark et al. 2015). Therefore, the result
may change to the null hypothesis being accepted.

In conclusion, a half reduction in the dataset may impact the reliability of the test, leading to
higher chance of commit type II error and may possibly change the result of the test to accepting
the null hypothesis.
NGUYEN MINH DUC – S3977946

Part E: Policy Brief

Title: Maternal Mortality Ratio and Income Disparity

Executive Summary:

In 2020, 95% of all maternal deaths occurred in low and middle-income countries. Accordingly,
findings have discovered that Gross National Income and Maternal Mortality Ratio exist an
inverse relationship. The policy brief will elaborate on the issue along with possible
recommendations.

Introduction:

Everyday, almost 800 women or a woman per two minutes passed away from pregnancy related
causes which can be preventable (WHO n.d.). The risk of maternal death between a low-
compared to a high-income countries is striking with a 108 times gap (UNICEF n.d.), resulting in
the global ratio to be 3 times higher than UN goal (UN 2023). The issue can exert pressure on
economic development and create a wider gap between low- and high-income countries (Wilson
Center 2010). However, having established the association between GNI and Maternal Mortality
Ratio helps identify the possible effecting factor, thus allowing us to find solutions.

Findings:

Data showed that income has an influence on the possibility of a country having high Maternal
Mortality Ratio. While high-income countries experience a low MMR, all of the low-income
countries suffered from uncontrollable high MMR and the figure for middle-income group is
50%. Although testing result found the global ratio is having a tendency to decrease in the future,
the outcome is affected by limited sample size with just 40 countries. More accurate calculations
will require bigger sample size for the best conclusions (Andrade 2020).

Recommendations:

First, one of the biggest reasons for high Maternal Mortality ratio in low-income countries is
poor healthcare system. (Abraham 2018). Therefore, governments of developing countries
should put more priority in medical care. As researchers found that skilled delivery service
coverage has a strong negative correlation with MMR (Tadele and Wasie 2017), all woman
NGUYEN MINH DUC – S3977946

should have access to high quality healthcare with skilled health professionals (WHO n.d.).
According to WHO (n.d.), this is extremely essential since timely treatment can make a life and
death for either the mother or newborn in these situations. As an example, high-income countries
have invested in maternal health care and almost successfully eliminated maternal deaths.

As low-income usually comes with limited education (Scott and Jessica 2017), adult literacy rate
is proved to be negatively correlated with MMR (Tadele and Wasie 2017). Local authority in
low-income area should focus on educate woman who is lack of understanding about the risks
associated with pregnancy (Myanmar UNFPA n.d.). Reducing total fertility rate through family
planning can also have positive effect due to a positive relationship with MMR in low-income
countries (Tadele and Wasie 2017).

Conclusion:

The issue of Maternal Mortality in low-income countries is urgent and need to be tackled as soon
as possible since it can negatively affect economic development (Wilson Center 2010). To solve
this problem, developing countries may consider putting more emphasis to their healthcare
system while providing adequate maternal education and effective family planning.

References:

Abraham h (2018) ‘Maternal and Newborn Mortality- Still the Greatest Disparity between Low-
Income and High-Income Countries’, Ethiopian Journal of Health Science, 28(4):368.

Andrade C (2020) ‘Sample Size and its Importance in Research’, Indian Journal of
Psychological Medicine, 42(1):102-103.

Myanmar UNFPA (n.d.) Policy Brief on Maternal Mortality, Myanmar UNFPA website,
accessed 29 August 2023. https://myanmar.unfpa.org/sites/default/files/pub-
pdf/policy%20brief%20and%20infographics_Maternal%20Mortality.pdf

Scott AW and Jessica S (2017) Education, Income, and Wealth, Federal reserve bank of St.
Louis website, accessed 29 August 2023. https://research.stlouisfed.org/publications/page1-
econ/2017/01/03/education-income-and-
wealth#:~:text=The%20relationship%20between%20education%20and,incomes%20(see%20the
%20table)
NGUYEN MINH DUC – S3977946

Wilson Center (2010) The Impact of Maternal Mortality and Morbidity on Economic
Development, Wilson Center website, accessed 29 August 2023.
https://www.wilsoncenter.org/event/the-impact-maternal-mortality-and-morbidity-economic-
development

Contact Information:

Nguyen Minh Duc – [email protected]

References
Abraham h (2018) ‘Maternal and Newborn Mortality- Still the Greatest Disparity between Low-
Income and High-Income Countries’, Ethiopian Journal of Health Science, 28(4):368.

Andrade C (2020) ‘Sample Size and its Importance in Research’, Indian Journal of
Psychological Medicine, 42(1):102-103.

DHS (n.d.) Maternal Mortality, DHS website, accessed 29 August 2023.

https://dhsprogram.com/pubs/pdf/FR183/12Chapter12.pdf

Jean PN and Adrien M (2020) ‘Forecasting Maternal Complications Based on the Impact of
Gross National Income Using Various Models for Rwanda’, Journal of Environmental and
Public Health, 2020: 7692428.

Mark B, David L, Kathryn S, Martin O, Judith W and Nicola J (2015) Basic Business Statistics
EBook, Pearson Education Australia, ProQuest Ebook Central.

Myanmar UNFPA (n.d.) Policy Brief on Maternal Mortality, Myanmar UNFPA website,
accessed 29 August 2023.

https://myanmar.unfpa.org/sites/default/files/pub-
pdf/policy%20brief%20and%20infographics_Maternal%20Mortality.pdf
NGUYEN MINH DUC – S3977946

ncl (n.d.) Other Measures of Dispersion, ncl website, accessed 29 August 2023.
https://www.ncl.ac.uk/webtemplate/ask-assets/external/maths-resources/statistics/descriptive-
statistics/other-measures-of-dispersion.html

OpenStax (n.d.) Introductory Business Statistics, OpenStax, Houston.

PennState (n.d.) 2.2.4.1 - Skewness & Central Tendency, PennState website, accessed 29 August
2023.

https://online.stat.psu.edu/stat200/lesson/2/2.2/2.2.4/2.2.4.1

Pennsylvania (n.d.) T-chart, Pennsylvania website, accessed 29 August 2023.

https://www.education.pa.gov/K-
12/Career%20and%20Technical%20Education/Documents/Electrical%20T-Chart%20-
%20Describe%20how%20outliers%20affect%20measures%20of%20central%20tendency.pdf

Roosa T, Munira ZG, Molly FG and Laurie CZ (2020) Maternal Mortality and Maternity Care in
the United States Compared to 10 Other Developed Countries, The Commonwealth Fund
website, accessed 29 August 2023.

https://www.commonwealthfund.org/publications/issue-briefs/2020/nov/maternal-mortality-
maternity-care-us-compared-10-
countries#:~:text=Key%20Findings%3A%20The%20U.S.%20has,mortality%20rate%20among
%20developed%20countries

Sage (n.d.) 33 Power and the Factors Affecting It, Sage website, accessed 29 August 2023.
https://www.sagepub.com/sites/default/files/upm-binaries/35401_Module33.pdf

Scott AW and Jessica S (2017) Education, Income, and Wealth, Federal reserve bank of St.
Louis website, accessed 29 August 2023.

https://research.stlouisfed.org/publications/page1-econ/2017/01/03/education-income-and-
wealth#:~:text=The%20relationship%20between%20education%20and,incomes%20(see%20the
%20table)

Sima S, Reza M, Abou AV, Mahmoud GT and Kazem M (2015) ‘Maternal mortality: a cross-
sectional study in global health’, Globalization and Health, 11(4).
NGUYEN MINH DUC – S3977946

Tadele G and Wasie A (2017) ‘Correlates of maternal mortality in developing countries: an

ecological study in 82 countries’, Maternal Health, Neonatology and Perinatology, 3(19).

Taylor C (2019) Differences Between Population and Sample Standard Deviations, ThoughCo.
website, accessed 29 August 2023.

https://www.thoughtco.com/population-vs-sample-standard-deviations-
3126372#:~:text=The%20population%20standard%20deviation%20is,the%20individuals%20in
%20a%20population

The World Bank (n.d.) GNI per capita, Atlas method (current US$) [data set], The World Bank
website, accessed 29 August 2023.
https://data.worldbank.org/indicator/NY.GNP.PCAP.CD?most_recent_value_desc=true

UN (United Nation) (n.d.) The 17 Goals, UN website, accessed 29 August 2023.

https://sdgs.un.org/goals

UNICEF (2023) Maternal mortality, UNICEF website, accessed 29 August 2023.

https://data.unicef.org/topic/maternal-health/maternal-mortality/

WHO (World Health Organization) (n.d.) Maternal mortality ratio (per 100 000 live births),
WHO website, accessed 29 August 2023.

https://www.who.int/data/gho/indicator-metadata-registry/imr-
details/26#:~:text=Maternal%20mortality%20ratio%20%3D%20(Number%20of,household%20s
urveys%20or%20other%20sources.)

Wilson Center (2010) The Impact of Maternal Mortality and Morbidity on Economic
Development, Wilson Center website, accessed 29 August 2023.
https://www.wilsoncenter.org/event/the-impact-maternal-mortality-and-morbidity-economic-
development
NGUYEN MINH DUC – S3977946