Skip to content

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

4 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

🎓 Identifying Key Policy Levers to Improve Immigrant Educational Attainment

🌟 About This Project

The problem addressed is the persistent educational attainment gap observed between foreign-born (immigrant) and native-born U.S. residents.

This analysis provides quantifiable, evidence-based insights to inform policy by using a stable Weighted Least Squares (WLS) model to isolate the effects of socioeconomic, structural, and linguistic factors.

⚙️Data Source and Methodology

Data Source

  • Source: ACS PUMS 1-Year Estimates, 2023.

  • Geography: State of Florida.

  • Key Variables Used: SCHL (Education), NATIVITY (Birth Status), POVPIP (Income Proxy), and PWGTP (Weight).

Final Model Specification (WLS)The model was significantly enhanced from the initial concept to include two critical language barrier variables and a technique to ensure stability

🧹 Data Preparation and Feature Engineering

This project transforms the raw PUMS codes into the precise variables required for the regression model.

  • Sample Definition: The dataset was filtered to include only individuals aged 25 and older (AGEP >= 25) to accurately capture completed education.

  • Dependent Variable (Y) (Years of Schooling): The categorical SCHL codes (e.g., Bachelor's Degree) were mapped to a continuous numerical variable representing the total years of schooling (e.g., 16 years).

  • Independent Variable $\mathbf{X}_1$ (Is_Immigrant):

    1. This variable is a binary dummy created from the NATIVITY variable.

    2. Classification: $1 = \text{Immigrant (Foreign-Born)}$ and $0 = \text{U.S. Born}$ (Reference Group).

Note: Due to the age filter (25+), the NOP variable for second-generation status was unusable, necessitating this robust binary proxy.

  • Weighting: The PWGTP (Person Weight) variable is applied in the MLR to ensure the sample estimates accurately reflect the true population.

A. Key Variables Added

The model's explanatory power was improved by adding two key linguistic variables:

  • Limited_English_household: Binary variable identifying low English proficiency in the household.

  • Speaks_Other_Language: Binary control for language spoken at home.

B. Statistical Stabilization

The model required stabilization to ensure reliable coefficients:

Problem: The initial model suffered from severe multicollinearity.

Solution: The socioeconomic variable $\text{POVPIP}$ was mean-centered to create $\text{POVPIP}_c$. This stabilized the model, reducing the Condition Number from $2.03e+03$ (Problematic) to $\mathbf{578.}$ (Stable).

📊 Results

  • The Limited English Household Barrier is more than double the magnitude of the residual Immigrant Gap, proving its importance as a policy focus.

  • The potential gain is (3.2 years) which more than four times the size of the Dominant Language Barrier (0.74 years).

📝Policy

  • Prioritize and significantly fund adult english as a Second Language (ESL) and family literacy programs.

  • Provide dedicated, language-appropriate financial aid and college Counseling for immigrant families.

  • Develop specialized high School integration pathways immigrant students.

About

A comparative analysis of education across first, second and later immigrant generations

Resources

Stars

0 stars

Watchers

0 watching

Forks

Releases

Packages

Contributors

Languages