Introduction

1. • According to UNICEF, 12 million girls are married in childhood per year (UNICEF, 2020).

   • The consequences are that the children are ill prepared for forced adult responsibilities, such as early fatherhood or motherhood (UNICEF, 2020). They will feel economic and social pressure of household responsibilities (UNICEF, 2020). Child marriage limits access to education, restricts opportunities for career advancement, and perpetrates social isolation (UNICEF, 2020).

   • These are the impacts of child marriage, which can also be variables for countries’ happiness scores.

   • "Understanding Happiness and Psychological Wellbeing Among Young Married Women in Rural India" found that the younger the girls were married, the worse the psychological wellbeing (Ghosh, Lahiri, & Datta, 2017). Martial happiness was significantly higher for girls married at 18 than girls younger (Ghosh, Lahiri, & Datta, 2017). Child marriage, restricted education, and age was predictive of well-being.

   • "Investigation of the key factors that influence
the girls to enter into child marriage: A metasynthesis of qualitative evidence" identified six themes. This includes human insecurity and conflict; legal issues; family values and circumstances; religious beliefs; individual circumstances, beliefs, and knowledge; and social norms (Ayako et al., 2020). The studies that found these factors are intertwined, and interventions need to account for all the factors to work (Ayako et al., 2020). The researchers pointed out that there lack studies focusing on income level, cultural and religious background, and geographical region (Ayako et al., 2020).

   • Our research aims to find the relationship between child marriage and geographical region/countries, happiness score and GDP.

Data Curation and Ethics

1. • As for data collection and analysis ethics, international dataset are held to a high standard. We chose the datasets of child marriage and happiness scores in 2019. The biggest limitation is that the UNICEF dataset only had a limited amount of countries listed. The datasets vary in the year the data sourced from.

   • UNICEF has three general guidelines.

     1. First, integrity includes honesty, truthfulness and fairness. They must maintain an objective and independent attitudes, avoid conflicts of interest and engage only in fair trade.

     2. Second, they value high accountability, transparency, and fairness for their functions, decisions, and actions. All outside procurement activities must be in compliance with UNICEF ethical values and principles. They have a zero tolerance policy for violations.

     3. Finally, they uphold respect, care and trust. This includes respect for human rights, dignity, and worth of all persons. They shall act with understanding, tolerance, sensitivity and respect for diversity without discrimination of any kind.

   • For child marriage, UNICEF has a 93-page document, “Ethical Principles, Dilemmas and Risks in Collecting Data on Violence against Children.” The emphasis is on ensuring that findings are disseminated to those who are able to use them effectively to promote positive action. Ethical principles for child data concern privacy, confidentiality, child protection, dissemination of findings, the training of researchers, and the welfare of researchers.

   • Our data curation, preparation, and analysis have not been dramatically altered or misrepresented

   • The data file that we have read into our Colab can be found below

Data Preparation and Exploration: Hypothesis Statements

1. • Hypothesis:

     • HA: If a specified country has a low happiness level, the same country will have a greater higher average of child marriages/year

     • H0: There is no difference in average of child marriages/year based on a countries’ happiness level

Data Preparation and Exploration: Variables

1. • The variables in the happiness2019 data frame are of interest to us since they could be the underlying factors as to why some countries experience higher vs. lower rates of child marriages per year.

   • Our group wanted to look specifically at the ways a country's happiness score influences the average rate of child marriages per year. All the factors that make up the happiness score are also included in the data frame (GDP, generosity, perceptions of corruption, etc.), so we thought it made more sense to look at the value that combined all of these aspects.

   • As the project endured, we proceed to take a closer look at the Happiness Score's components to see if they have any correlation to the average rates of child marriages per year individually. GDP per capita was the component of the Happiness Score that seemed to have the greatest affect on child marriages out of the others so we spent more time investigating the relationship between Happiness Score & Child Marriages and GDP per capita & Child Marriages.

   • The size and shape of the two primary data frames are listed below:

Data Preparation and Exploration: Combining DataFrames

1. • We had to reshape the data frames in order to match one another size wise for some of our tests to run successfully. The Child Marriages data file that we used had considerably fewer countries listed than in the Happiness Scores file so we had to locate the countries that existed on both data frames in order to combine them into one.

   • After combining the two, there were still quite a few countries within the Child Marriages data frame that didn't exist in the Happiness Scores data frame. Those countries are listed below:

     • Cabo Verde

     • Eritrea

     • Equatorial Guinea

     • Kiribati

     • Marshall Islands

     • Republic of Moldova

     • Tonga

     • Tuvalu

   • The new combined data frame has the following data types, shape and size below:

Data Preparation and Exploration: Visualization of Data

• The graphs below represent the top ten countries with the highest Happiness Scores, ten lowest Happiness Score, top ten highest rates of Child Marriages, ten Countries with lowest Child Marriage rates, and the countries happiness score from the combined happy marriages data frame compared to the child marriages for each gender and age demographic. Female Child Marriages was the main demographic we initially intended to focus our research on in addition to the two variable (Score and GDP). But with the data frames being condensed down to 12 countries, we decided to investigate across all age and gender demographics applicable given the study's limitations.

• Since we're also interested in the role GDP plays, we made a series of bar graphs that show the relationship between the two variables of interest.

Data Preparation and Exploration: Visualization of Data

• We also did a series of graphical comparisons with GDP per capita and Child Marriages:

Data Preparation and Exploration: Descriptive Statistics

• Below is a small report summary/descriptive statistics of each variable - means, SDs, Variance

Normality Analysis: Happiness & Child Marriages

1. • Each Regression line formulated shares one thing in common with one another. A negative slope. This continues to affirm our idea that Happiness Scores impact Child Marriage rates. As Happiness score for country's decrease, the average rate of child marriages increases alongside. This negative slope and correlation depicts a negative relationship between the variables of interest here.

   • The normality tests could also arguably support our findings as the results do not stray that far from our initial findings. The values generated adhere closely to the our own data, confirming our data as being normal since we didn't have any statistical significance from the distributions to prove otherwise.

   • In order to run these regression tests, a series of code had to precede to make the results possible. We started by prepping with a code cell to initiate a "cdf" or a cumulative distribution function. This helps set the foundation for the real randomized variables we pull throughout the duration of the code series.

   • Before making the regression lines, three graphs are procured. In our case, we had a CDF graph of Happiness scores and Child Marriages. We then combined the randomized scores and child marriage value into a scatter plot that would be the basis of the regression lines we see below.

Normality Analysis: GDP & Child Marriages

1. • A similar analysis can be made for the GDP and child marriage Regression lines below. Each of the regression lines below also have a negative slope. Thus confirming again, our idea that other economic factors impact child marriage rates.

Bootstrap Test

1. • In order to conduct the bootstrap test, we had to create hypothetical data much like in our normality code section. To start, the columns that contained data for "Score" and "GDP per capita" had to be converted into a 1-dimensional array so they could be used for the tests. From there replicate data was made that consisted of means, probability density, and variance. The graphs comparing Means&PDFs can be found below, as well as graphs pertaining to the variance for each variable.

   • After running multiple bootstrap tests, we ran tests that displayed the confidence intervals for GDP per capita and Happiness scores, as well as tests to determine the Standard error mean for each variable.

   • The reason we run repeated measures is because a designated percent of the observed value would lie within the confidence interval. The confidence interval that this hypothetical data falls into is between [.46, .96]. Since the hypothetical data falls within this category we know we can be confident about the results as they fall within the interval. Running a test just once doesn't always ensure that the values are what we predicted them to be so running more measurements helps to validate the results and conclusions we come to.

Discussion

1. • The goal of this project was to determine which economic factors impacted countries' average rate of child marriages per year. Our hypothesis explicitly sought to find out what role the happiness score has on countries with high or low rates of child marriages. We were able to deduce that for the most part, a negative relationship existed between happiness score and child marriages. This meant that as countries experienced higher yearly averages of child marriages, their happiness scores were increasingly smaller than countries with lower yearly averages of child marriages. In figure(), the graph has a steady rise and the drop in child marriages through out the descending happiness score across all ages and genders investigated in this study. Upon reviewing these charts, our group started investigating the possibility of other factors that could be contributing to the cause of child marriages. With the use of matplotlib, we were able to analyze and investigate a wide array of data relationships. We ultimately determined that while happiness scores was a top contender for influencing a country's average rate of child marriages per year, GDP per capita had the strongest influence over all. When we developed regressions for the relationship between happiness score & child marriages, and GDP per Capita & Child Marriages, the values for GDP per capita & child marriages were much tighter around the regression line than the former grouping. Since regression lines are designed to create a line that fits best with the relationship between two variables, it's arguably safe to say that the two sets of variables (GDP & child marriages) have a stronger relationship than happiness score & child marriages. This conclusion doesn't necessarily refute our hypothesis entirely, but does show us that happiness scores are not the dominating factor behind child marriages like we anticipated.

   • Our research has an incomplete data for child marriage rates. After combining the child marriage and happiness score dataframe, there are only twelve less developed countries left. Our findings about the negative relationship between happiness score and child marriage rate might be biased due to the limited data. In addition, the child marriage data we use has child marriage rates of different reference years. Some statistics are collected in 2013 and are out of date.

   • Future studies need to find the most recent data of child marriage rates from more countries.

   • This research points that child marriage can be and should be addressed. Policies should educate and provide employment opportunities, and economic independence for girls. Grassroots organizations improve community engagement and awareness with social justice programs for human rights. Laws should enforce child marriage laws across the globe.

Works Cited

Ghosh, S., Lahiri, S., & Datta, N. (2017). Understanding Happiness and Psychological Wellbeing Among Young Married Women in Rural India 1. Journal of Comparative Family Studies, 48(1), 113–131,4–5,9,12–13. https://doi.org/10.3138/jcfs.48.1.113

Kohno, A., Techasrivichien, T., Suguimoto, S. P., Dahlui, M., Farid, N. D. N., & Nakayama, T. (2020). Investigation of the key factors that influence the girls to enter into child marriage: A meta-synthesis of qualitative evidence. PloS One, 15(7), E0235959.

UNICEF. (2020, November 11). Child marriage. UNICEF for every child. https://data.unicef.org/topic/child- protection/child-marriage/

United Nations Children's Fund (UNICEF). (2019, December 12). Ethical Principles, Dilemmas and Risks in Collecting Data on Violence Against Children. UNICEF DATA. https://data.unicef.org/resources/ethical- dilemmas-risks-collecting-data-violence-children-findings-work-cp-merg-technical-working-group-violence- children/