Geospatial Project: African Americans & Instructional Expenditure By Mikayla Edwards

Hypothesis

My hypothesis is that states with higher Black or African American populations get less instructional expenditure. I based my hypothesis solely on diversity per state and upon creating it realized it would be better to focus on one specific race and thus decided to focus on individuals identifying as Black or African American. I believe that due to the historic relationships between Black/African Americans and the federal government when it comes to education (ex: slavery, segregation, Brown v Board of Education, HBCU's, etc.) that there will be a negative correlation between states with higher (in regard to the rest of the continental United States) populations of Black or African American individuals and the amount of instructional expenditure they receive from the government.

Data Sources: https://www.kaggle.com/noriuk/us-education-datasets-unification-project & https://www.kaggle.com/noriuk/us-education-datasets-unification-project

Percent Black

Here is a map of the percent of the population in each state that self-identified as Black or African American.

Correlation?

I decided to merge the 2 datasets in order to run a correlation analysis on two variables: Percent Black or African American & Instructional_Expenditure To do this I used a left join to incorporate all the data from states_all_2 (the dataset with the instructional expenditure amounts) with the diversityindex dataset on the basis of State which is found in both datasets under different column names.

library(data.table)
library(dplyr)
Black_edu <- states_all_2 %>%
  left_join(diversityindex, by = c("STATE" = "Location"))
cor(Black_edu$INSTRUCTION_EXPENDITURE,Black_edu$`Black or African American alone, percent, 2013`, use= "complete.obs")

[1] 0.1156634

I decided to complete the same analysis on indivudlas who identified as Two or More Races as many Americans, 2.9% (9 million individuals) according to the 2010 US Census, self-identified as multiracial. Studies have shown that not only has this number increased since then but it will only continue to grow even more in the future.

cor(Black_edu$INSTRUCTION_EXPENDITURE,Black_edu$`Two or More Races, percent, 2013`, use = "complete.obs")

[1] -0.1086686

Conculsion

In conclusion my hypothesis was incorrect as there appears to be a weak linear relationship between the percent of the population that self-identifies as Black or African American and the amount of educational expenditure received per state.

This research could be furthered by looking deeper into the "Two or more races" category and what that entails as there is a negative correlation between those indivudals and the amount recedived in instructional expenditure.

It would also be interesting to creat a decision tree based on the Black_edu data to see what variable most affects instructional expenditure as well as doing further resreach to see how the federal government decides how much money to give states.