Research Presentation - Final Project
Title
The Association between income per person and life expectancy rate among countries worldwide with varying urbanization rates.
Vijaya Phanindra Sarma, Passion Driven Statistics Course, Wesleyan University
Introduction
It is a known fact that average income of a person of a country is a good indicator to study the standard of living of citizens of that country. A country’s Income per Person(1) or average wealth of a citizen tends to be closely related and linked to other socio economic indicators in general like better healthcare, longer life expectancy rates, high standard of living, economic freedom and ability to spend beyond needs and necessities. In general citizens of developed countries with high income have longer life expectancy rates with lower infant mortality rates. Furthermore the rapid urbanization as a result of rural migration caused by the lure of opportunities that cities can offer might be having both positive impact in the form of increased economic growth, better healthcare, educational and other facilities. It might be the case that high income per person of a country contributed to high life expectancy rates and also high urbanization rates in a country to higher income per person levels.
Research Questions
Is longer life expectancy rate associated with high income per person of a country
Is the association between high income per person and life expectancy rate is same across countries with high urbanization rates
Methodology
Sample
GNI (Gross National Income) Per Capita, Life Expectancy is taken from the World Bank development indicators for all the available countries for years 2001-2011.
Urban Rate is taken from the International Human Development Indicators for years 2000-2010
As there is missing data for some countries, average value of 10 years of data is considered.
Measures
The original data is was from the Gapminder(2) code book.
These data is of different years between different selected variables.
Data is missing for many countries, below tables gives snap shot of missing data.
So Based on the code book definition, generated an alternative dataset for same variables and from the same source
Main Source of search is http://www.google.co.in/publicdata/directory
Income Per Person Quantitative variable is binned into four categories as per world Development indicators classification(3)
Income below 1025 "LOW INCOME" Income between 1025 and 4035 "LOWER MIDDLE INCOME" Income between 4035 & 12475 "UPPER MIDDLE INCOME" Income more than 12475 "HIGH INCOME";
Mean and Standard Deviation of Income Per Person income_2001-11 variable. PROC Univariate Procedure
Income per Person Groups and Urbanization Rate (Categorized quantitative Variables). PROC FREQ procedure.
/* ANOVA Analysis Incomeperperson vs Life Expectacy Rate */
/* DUNCAN POST HOC TEST AS the CATEGORICAL VARAIBLES Have more than 2 levels */
PROC ANOVA; CLASS incomeperperson_grp;
MODEL life_expectancy_2001_11=incomeperperson_grp;
MEANS incomeperperson_grp /DUNCAN;
/* ANOVA Analysis Incomeperperson vs Urban Rate */
PROC ANOVA; CLASS incomeperperson_grp;
MODEL urbanrate_2000_10=incomeperperson_grp;
MEANS incomeperperson_grp /DUNCAN;
Results
Univariate
Among all the countries the average income per person is 7320 (Std Deviation 10425)
Of the four income level groups, 60% (31.72, 29.66) form the lower income and lower middle income range
Of the four Urban Rate groups, 35% the highest are from the upper middle urbanization group.
Bivariate
ANOVA analysis between Income Per Person Groups (Category Explanatory Variable) and Life Expectancy Rate (Quantitative Response Variable) shows that the association is statistically significant (F Value= 92.50 and P <0.0001)
Further Duncan Test on multiple categories of explanatory variable income per person shows that these 4 income level groups are statistically different.
ANOVA analysis between Income Per Person Groups (Category Explanatory Variable) and Urban Rate (Quantitative Variable) shows that the association is statistically significant (F Value= 62.33 and P <0.0001)
Further Duncan Test on multiple categories of explanatory variable income per person shows that these 4 income level groups are statistically different.
Moderation
Urban Rate confirmed positive and significant association between Income Per Person and Life Expectancy Rates.
Among countries with High Income Per Person and are in Highly Urbanized Cities there is high life expectancy rate.
A bar chart below clearly illustrates association between Income Per Person, Urban Rate and Life Expectancy rate.
Furthermore the highest life expectancy rate of age 80 is found in both High and Upper Middle Urban Cities of a country and of people with high income per person.
This group is visibly absent in the other two groups Middle Urban and Lower Urban Cities
Discussion
What might the results mean?
• Countries with higher income per person and with large population living in urbanized cities seems be having longer life expectancy rates.
Strengths
• Results are based on the data provided by the World Bank
• Average value of a decade (10 years) considered for analysis which fairly represents data over a considerable period
Limitations
• Further Research is required to determine if there are any other factors that influence Life Expectancy Rate
• A good example of existence of other factors is for instance a county might focus to provide better social security even though their income per person is very low, for instance Ecuador with income per person at around USD 1515 (which is Lower Middle Income group) has life expectancy rate of 74.79
(1) A country’s GNI (Gross National Income) Per Capita is dollar value of country’s yearly income divided by its total population
(2)www.gapminder.org
(3)http://data.worldbank.org/about/country-classifications