1
Kwame Darko-Mensah
INCOME AND COLLEGE TIERS
CAPSTONE PROJECT CONCEPT
Model: How are factors such as parent income, household income distributions, children married in 2014(a few years after graduation) , tiers of school attended by children have an effect on the children’s income earnings.
INTRODUCTION
“The concept of college tiers did not exist before the mid-1980s. This was just a term used by statisticians and became a successful marketing ploy that big-name colleges use to build their brands.” (Broderick 2018)
MOTIVATION OF MY STUDY
College tiers can matter if other careers interest you, college tiers take a backseat to other factors such as your ability to pay for the program. There are various factors that influence which college tier a student attends which includes parent’s household income. Little has been done to distinguish how these factors affect the school tier a child attends. Therefore, this study seeks to evaluate more on these factors and their association/relationship with school tier. College tiers are defined under the data set section.
DATASET
The data set was acquired from the data repository https://opportunityinsights.org/data/. The study will use a cross-sectional data that This table reports parents’ household income distributions and key statistics on children’s income distributions conditional on their parents’ incomes by college tier. The sample pools all children in the 1980, 1981 and 1982 birth cohorts. There is one row for each parent percentile and college tier.
The values in this table are exact statistics (not estimates) because they aggregate across multiple colleges. Following established disclosure standards, all dollar amounts are rounded to the nearest 100 and cells with fewer than 50 observations are omitted. Monetary values are measured in 2015 dollars.
The college tiers are defined based on a cross-tabulation the following set of variables:
Type: public / private not-for-profit / private for-profit Level: 4-year or above, 2-year, less than 2 years.
Selectivity:
1. Ivy Plus Schools
2. Elite schools other than Ivy Plus (Barron’s 2009 selectivity index of 1)
3. Highly selective schools (Barron’s selectivity index of 2)
4. Selective schools (Barron’s selectivity index of 3-5)
5. Nonselective schools (Barron’s selectivity index > 5 or unlisted in Barron’s)
Definition of Main variables provided below
TECHNIQUE
In this project, I will use a classification method such as Poisson regression, some cluster analysis and bootstrapping to check the association between dependent variable(tier) which is categorical in nature and independent variables (parent income, child earnings, parent household income, children married in 2014 and children not working). Through the model, I will determine which variables are significant in the study in determining the school tier attended by students.
Also, in my exploratory analysis, I will plot boxplots, histograms, scatterplots, and bar plots to have a graphical visualization of the variables of interest also how the predictor and response variable associate by observing the patterns within the dataset.
I will also perform correlation analysis to view how the variables are correlated within the dataset and descriptive statistics to obtain the central tendency of the dataset.
Research Question: What is the association between dependent variable (tier) and independent variables (parent income, child earnings, parent household income, children married in 2014 and children not working)?
HYPOTHESIS
Null hypothesis: Children’s earnings are influenced by tier of college attended, children’s percentile rank in schools and household income
Alternative hypothesis: Children’s mean earnings are not influenced by external factors such as parents household income etc.
STATISTICAL TOOL FOR THE ANALYSIS
I intend to use R programming language as the statistical tool to perform my analysis for this project.
REFERENCE
Broderick, Thomas (2018). What are college tiers? https://myklovr.com/blog/what-are-college-tiers