Alcohol Consumption and Infant Mortality Rates

Abstract

The purpose of this observational study was to discover if there is a relationship between alcohol consumption and infant mortality rates. Furthermore, if there is a relationship, how does wealth effect it? The more I looked the less likely a relationship seems to exist. In fact, in some instances, a negative correlation might exist.

Introduction

Research Question: Does a countries wealth class change the impact alcohol consumption has on infant mortality rates?

Alcohol is probably most abused drug in the world. Every years we get contradicting reports on whether alcohol is beneficial or harmful to your health. The point of this study is to see how alcohol consumption effects child birth. Obviously, drinking while pregnant is dangerous, but what about before a person is pregnant? Does the father’s alcohol consumption play a role? Does the society’s alcohol consumption play a role?

Gathering the Data

The data came from the World Health Organization and the World Bank. Both websites offer a tool that allows you to filter a giant dataset into the data you want and have it exported in the form of a csv or Excel sheet.

The observations in this case are countries. The explanatory variable is the amount of liters of pure alcohol consumed by a country. The response variable is the amount of infant deaths for every 1000 births.

The links are provided in the comments of each link in my GitHub.

#Alcohol consumption
alc_link <- "https://raw.githubusercontent.com/tylerbaker01/Data-606-Final-Project/main/Alcohol%20Consumption.csv"
#Infant mortality rate
infant_link <- "https://raw.githubusercontent.com/tylerbaker01/Data-606-Final-Project/main/infant%20mortality%20rates%20per%20country.csv"
#GDP
## Data from The World Bank
gdp_link <- "https://raw.githubusercontent.com/tylerbaker01/Data-606-Final-Project/main/GDP%20per%20capita%20by%20country.csv"
alc_df <- read.csv(alc_link)
infant_df <- read.csv(infant_link)
gdp_df <- read.csv(gdp_link)

Tidy Data

Infant Mortality Rate

# I only want data from the last 25 years.
infant_df <- infant_df[-c(2:38)]
# Blank variable
infant_df <- infant_df[-c(28)]
# Name Columns
colnames(infant_df) <- c("country", "1994", "1995", "1996", "1997", "1998", "1999", "2000", "2001", "2002", "2003", "2004", "2005", "2006", "2007", "2008", "2009", "2010", "2011", "2012", "2013", "2014", "2015", "2016", "2017", "2018", "2019")
# Remove unwanted rows
infant_df <- infant_df[-c(1:5), ]
# Remove NAs
infant_df <- na.omit(infant_df)
# Pivot longer
infant_df <- infant_df %>%
  pivot_longer(!country, names_to = "year", values_to = "deaths per 1000 births")

GDP per Capita

# Keep last 25 years
gdp_df <- gdp_df[-c(2:39)]
# Name Columns
colnames(gdp_df) <- c("country", "1995", "1996", "1997", "1998", "1999", "2000", "2001", "2002", "2003", "2004", "2005", "2006", "2007", "2008", "2009", "2010", "2011", "2012", "2013", "2014", "2015", "2016", "2017", "2018", "2019")
# Remove unwanted rows
gdp_df <- gdp_df[-c(1:5), ]
# Remove unwanted column
gdp_df <- gdp_df[-c(27)]
# Remove NAs
gdp_df <- na.omit(gdp_df)
# Pivot longer
gdp_df <- gdp_df %>%
  pivot_longer(!country, names_to = "year", values_to = "GDP per Capita")

Alcohol Cosumption

# Keep only needed variables
alc_df <- alc_df[c("Location", "Period", "Dim1", "Value")]
# Rename variables
colnames(alc_df) <- c("country", "year", "sex", "liters of pure alcohol")
# Sort by countries
alc_df <- alc_df %>%
  arrange(country)
# Only keep both sex's data
alc_df <- subset(alc_df,  sex == "Both sexes")
# Remove column
alc_df <- alc_df[c("country", "year", "liters of pure alcohol")]
# Convert data type
alc_df$year <- as.character(alc_df$year)
# Convert data type
alc_df$`liters of pure alcohol` <- gsub("\\[.*","", as.character(alc_df$`liters of pure alcohol`))
alc_df$`liters of pure alcohol` <- as.numeric(alc_df$`liters of pure alcohol`)

Combining the Data

I want only the countries in which all three data frames have in common.

combination <- inner_join(gdp_df, infant_df)

## Joining, by = c("country", "year")

country_stats <- inner_join(combination, alc_df)

## Joining, by = c("country", "year")

Analysis

Splitting the countries into groups based on GDP per capita

country_stats_2019 <- country_stats %>%
  filter(year == "2019")
gdp_quantiles <- quantile(country_stats_2019$`GDP per Capita`)
country_stats_2019 <- country_stats_2019 %>%
  mutate(`wealth class` = 
           case_when(
             `GDP per Capita` < 2401.7629 ~ "lower", `GDP per Capita` < 6724.3071 ~ "low-mid", `GDP per Capita` < 17544.3487 ~ "upper-mid", `GDP per Capita` >= 17544.3487 ~ "upper"
           )
  )

Lower class countries

How does alcohol consumption effect infant mortality rate in lower class countries?

lower_class_countries <- country_stats_2019 %>%
  filter(`wealth class` == "lower")

Lower Middle Class Countries

low_mid_countries <- country_stats_2019 %>%
  filter(`wealth class` == "low-mid")

Upper Middle Class Countries

upper_mid_countries <- country_stats_2019 %>%
  filter(`wealth class` == "upper-mid")

Upper Class Countries

upper_class_countries <- country_stats_2019 %>%
  filter(`wealth class` == "upper")

Summary Statistics (for 2019)

means

countries_alc_mean <- mean(country_stats_2019$`liters of pure alcohol`)
gdp_mean <- mean(country_stats_2019$`GDP per Capita`)
infant_mortality_mean <- mean(country_stats_2019$`deaths per 1000 births`)

medians

countries_alc_median <- median(country_stats_2019$`liters of pure alcohol`)
gdp_median <- median(country_stats_2019$`GDP per Capita`)
infant_mortality_median <- median(country_stats_2019$`deaths per 1000 births`)

range

countries_alc_range <- range(country_stats_2019$`liters of pure alcohol`)
gdp_range <- range(country_stats_2019$`GDP per Capita`)
infant_mortality_range <- range(country_stats_2019$`deaths per 1000 births`)

IQR

countries_alc_iqr <- IQR(country_stats_2019$`liters of pure alcohol`)
gdp_iqr <- IQR(country_stats_2019$`GDP per Capita`)
infant_mortality_iqr <- IQR(country_stats_2019$`deaths per 1000 births`)

Summary Statistics (Lower Class)

means

lower_class_countries_alc_mean <- mean(lower_class_countries$`liters of pure alcohol`)
lower_class_infant_mortality_mean <- mean(lower_class_countries$`deaths per 1000 births`)

medians

lower_class_countries_alc_median <- median(lower_class_countries$`liters of pure alcohol`)
lower_class_infant_mortality_median <- median(lower_class_countries$`deaths per 1000 births`)

range

lower_class_countries_alc_range <- range(lower_class_countries$`liters of pure alcohol`)

lower_class_infant_mortality_range <- range(lower_class_countries$`deaths per 1000 births`)

IQR

lower_class_countries_alc_iqr <- IQR(lower_class_countries$`liters of pure alcohol`)

lower_class_infant_mortality_iqr <- IQR(lower_class_countries$`deaths per 1000 births`)

Summary Statistics (Lower Middle)

means

lower_mid_countries_alc_mean <- mean(low_mid_countries$`liters of pure alcohol`)
low_mid_class_infant_mortality_mean <- mean(low_mid_countries$`deaths per 1000 births`)

medians

lower_mid_class_countries_alc_median <- median(low_mid_countries$`liters of pure alcohol`)
lower_mid_class_infant_mortality_median <- median(low_mid_countries$`deaths per 1000 births`)

range

lower_mid_class_countries_alc_range <- range(low_mid_countries$`liters of pure alcohol`)

lower_mid_class_infant_mortality_range <- range(low_mid_countries$`deaths per 1000 births`)

IQR

lower_mid_class_countries_alc_iqr <- IQR(low_mid_countries$`liters of pure alcohol`)

lower_mid_class_infant_mortality_iqr <- IQR(low_mid_countries$`deaths per 1000 births`)

Summary Statistics (Upper Mid)

means

upper_mid_class_countries_alc_mean <- mean(upper_mid_countries$`liters of pure alcohol`)
upper_mid_infant_mortality_mean <- mean(upper_mid_countries$`deaths per 1000 births`)

medians

upper_mid_countries_alc_median <- median(upper_mid_countries$`liters of pure alcohol`)
upper_mid_infant_mortality_median <- median(upper_mid_countries$`deaths per 1000 births`)

range

upper_mid_countries_alc_range <- range(upper_mid_countries$`liters of pure alcohol`)

upper_mid_infant_mortality_range <- range(upper_mid_countries$`deaths per 1000 births`)

IQR

upper_mid_countries_alc_iqr <- IQR(upper_mid_countries$`liters of pure alcohol`)

upper_mid_infant_mortality_iqr <- IQR(upper_mid_countries$`deaths per 1000 births`)

Summary Statistics (Upper Class)

means

upper_class_countries_alc_mean <- mean(upper_class_countries$`liters of pure alcohol`)
upper_class_infant_mortality_mean <- mean(upper_class_countries$`deaths per 1000 births`)

medians

upper_class_countries_alc_median <- median(upper_class_countries$`liters of pure alcohol`)
upper_class_infant_mortality_median <- median(upper_class_countries$`deaths per 1000 births`)

range

upper_class_countries_alc_range <- range(upper_class_countries$`liters of pure alcohol`)

upper_class_infant_mortality_range <- range(upper_class_countries$`deaths per 1000 births`)

IQR

upper_class_countries_alc_iqr <- IQR(upper_class_countries$`liters of pure alcohol`)

upper_class_infant_mortality_iqr <- IQR(upper_class_countries$`deaths per 1000 births`)

Visualization

All Countries

ggplot(data = country_stats_2019, aes(x = `deaths per 1000 births`, y = `liters of pure alcohol`))+
  geom_point()

ggplot(data = country_stats_2019, aes(x = `deaths per 1000 births`, y = `liters of pure alcohol`))+
  geom_point()+
  geom_smooth()

## `geom_smooth()` using method = 'loess' and formula 'y ~ x'

## Lower class countries

ggplot(data = lower_class_countries, aes(x = `deaths per 1000 births`, y = `liters of pure alcohol`)) +
  geom_point()

## Lower Middle Class Countries

ggplot(data = low_mid_countries, aes(x = `deaths per 1000 births`, y = `liters of pure alcohol`)) +
  geom_point()

## Upper Middle Class Countries

ggplot(data = upper_mid_countries, aes(x = `deaths per 1000 births`, y = `liters of pure alcohol`)) +
  geom_point()

## Upper Class Countries

ggplot(data = upper_class_countries, aes(x = `deaths per 1000 births`, y = `liters of pure alcohol`)) +
  geom_point()

What about the United States?

alc_df %>%
  filter(country == "United States of America") %>%
  ggplot( aes(x = year, y = `liters of pure alcohol`)) +
  geom_point()

Here we can clearly see that our alcohol consumption has been steadily increasing.

infant_df %>%
  filter(country == "United States") %>%
  ggplot( aes(x = year, y = `deaths per 1000 births`))+
  geom_point()

Here we can see that the US has a steady decline for infant deaths over the years.

Conclusion

The Result

To my surprise, alcohol consumption does not seem to play a significant role on infant mortality rates, regardless of how wealth the country is.

The Possible Explanations

Maybe there was an invention that significantly raised successful birth rates, regardless of development issues.
People that like to drink are more fertile than those that don’t.

Data 606 Final Project

Tyler Baker

12/8/2021

Alcohol Consumption and Infant Mortality Rates

Abstract

Introduction

Gathering the Data

Tidy Data

Infant Mortality Rate

GDP per Capita

Alcohol Cosumption

Combining the Data

Analysis

Splitting the countries into groups based on GDP per capita

Lower class countries

Lower Middle Class Countries

Upper Middle Class Countries

Upper Class Countries

Summary Statistics (for 2019)

means

medians

range

IQR

Summary Statistics (Lower Class)

means

medians

range

IQR

Summary Statistics (Lower Middle)

means

medians

range

IQR

Summary Statistics (Upper Mid)

means

medians

range

IQR

Summary Statistics (Upper Class)

means

medians

range

IQR

Visualization

All Countries

What about the United States?

Conclusion

The Result

The Possible Explanations