Is the SAT unfair?

Exploring and visualizing the data to find out

==============================

Goal of this project

The goal of this project is to examine if the SAT (Scholastic Aptitude Test) is unfair to any particular groups of students. The annual SAT is used by U.S. colleges and universities to determine which students to admit into their programs.

Approach

We will explore SAT and NYC (New York City) high school data to answer this question. We will examine the Pearson correlation coefficients between SAT scores and the other columns in our dataset to identify the variables with the strongest connections to SAT performance. We will then explore the strongest correlated columns further to uncover any bias in the SAT. We will create various data visualizations, such as scatter, bar and map plots, to aid our investigation.

Summary of results

At first glance, there appears to be some racial bias in the SAT, in favor of White and Asian students and against Hispanic students. It also appears that schools with larger student populations and class sizes outperform smaller schools on the SAT. However, further analysis suggests that the socioeconomic status of the students' families and whether they are English language learners are the most relevant factors tied to SAT performance.

Viewing the project

For proper rendering, I recommend viewing the project in nbviewer here.

Alternatively, you can view the project directly in GitHub by going to the "notebooks" folder of this repository. However, there are some rendering issues when viewing Jupyter Notebooks in GitHub.

Installing and running the project

Download the project

• Download and extract, or clone this repository to your computer
• Do not alter the folder/file structure or the file references in the Jupyter Notebook will break

Install the conda package manager

• If you do not have conda on your computer, you can follow the installation instructions here

Reproduce the project environment and launch the Jupyter Notebook

• From your computer's command line interface, run the following commands:
  • Make sure you are in the parent directory of this repository (i.e. ".../analyzing-the-sat", ".../analyzing-the-sat-master" etc..., whatever you named it when you downloaded or cloned it): cd <path-to-repo-parent-directory>
  • Reproduce the project environment: conda env create -f environment.yml
  • Activate the environment:
    • On Windows: activate project_env
    • On macOS and Linux: source activate project_env
  • Launch the Jupyter Notebook: jupyter notebook notebooks/1.0-jt-analyze-the-sat.ipynb

Run the Jupyter Notebook

• From the Jupyter Notebook "1.0-jt-analyze-the-sat.ipynb":
  • To run each cell one at a time: click Run in the top control panel or press Shift + Enter
  • To run all of the cells: click Kernel in the menu bar and then Restart & Run All

Project organization

---

├── data               <- The final cleaned dataset produced from running the
│                         "cleaning-nyc-hs-data" project      
│
├── notebooks          <- Jupyter notebooks. Naming convention is a number (for ordering),
│                         the creator's initials, and a short "-" delimited description, e.g.
│                         "1.0-jt-analyze-the-sat"
│
├── references         <- Data dictionaries
│
├── .gitignore         <- Files and folders for Git to ignore tracking
├── LICENSE            <- The license governing the use and redistribution of this project
├── README.md          <- The top-level README for users of this project  
└── environment.yml    <- The file for reproducing this project's environment

---