🧠 NESARC Research: Social Anxiety Disorder (SAD) & Family Behavioral History

A data-driven exploration of the relationship between SAD severity and family history of behavioral problems.

📌 Overview

This project was developed as part of the Data Analysis and Interpretation Specialization from Wesleyan University on Coursera. It explores the relationship between Social Anxiety Disorder (SAD) and a family history of behavioral problems using data from the National Epidemiologic Survey on Alcohol and Related Conditions (NESARC). The analysis applies statistical hypothesis testing, including ANOVA, Chi-Square test, Pearson Correlation, and moderation analysis (Gender as a moderator) to assess these relationships.

📂 Dataset

• Source: National Epidemiologic Survey on Alcohol and Related Conditions (NESARC)
• File: source/NESARC Dataset.csv
• Size: 252.61 MB (tracked with Git LFS)
• Key Variables:

Variable	Type	Description
SAD_score	Numerical	Composite score derived from SAD-related survey responses.
SAD_spectrum	Categorical	Low (≤2), Medium (2-5), High (>5) severity categories.
behavior_problems_count	Numerical	Number of relatives with behavioral problems.
relatives_with_problems	Binary (Y/N)	Presence of ≥1 relative with behavioral problems.

🎯 Objectives

1. Primary: Determine if family history of behavioral problems correlates with higher SAD severity.
2. Secondary: Assess if gender moderates this relationship.

🔍 Methodology

1. Data Preprocessing

• Cleaning: Removed missing values and standardized variables.
• Feature Engineering:
  • Created SAD_score from symptom-related survey responses.
  • Binned SAD_score into Low/Medium/High categories.
  • Derived relatives_with_problems from behavior_problems_count.

2. Statistical Analysis

ANOVA: SAD Spectrum vs. Behavior Problems Count

A one-way ANOVA revealed significant differences in behavior problems across SAD severity groups:

Source	Sum Sq	df	F	p-value
SAD_spectrum	125.78	2	22.24	2.48e-10
Residual	10919.14	3862	–	–

Post-hoc Tukey's HSD: All group pairs showed significant differences (p < 0.05).

Chi-Square Test: Relatives With Problems vs. SAD Spectrum

• χ² = 34.56 (p = 3.13e-08)
• Cramér's V = 0.095 (small effect size)

Conclusion: Significant association between family history and SAD severity.

Pearson Correlation: SAD Score vs. Behavior Problems Count

• r = 0.08 (p = 3.82e-07)
• r² = 0.0067 (0.67% variance explained)

Conclusion: Weak but statistically significant correlation.

Moderation Analysis: Gender as a Moderator

An ANOVA with interaction terms tested if gender moderates the SAD-behavior relationship

Key Findings:

• Interaction term p = 0.187 → Gender does not significantly moderate the relationship.
• Main effects of SAD_spectrum remain significant.

📁 Repository Structure

NESARC_research/
├── source/                  # Raw data (tracked via Git LFS)
│   └── NESARC Dataset.csv
├── .gitattributes           # Git LFS configuration
├── DMV.ipynb                # Jupyter Notebook (full analysis)
├── LICENCE                  # MIT Licence
├── README.md                # Project documentation
└── requirements.txt         # Python dependencies

🛠️ Installation & Usage

1. Clone the Repository

git clone https://github.com/PogloLopez/nesarc_research.git
cd nesarc_research

2. Install Git LFS & Download Data

git lfs install   # Set up Git LFS
git lfs pull      # Download dataset

3. Set Up a Virtual Environment

python -m venv .venv
source .venv/bin/activate  # Mac/Linux
.\.venv\Scripts\activate   # Windows

4. Install Dependencies

pip install -r requirements.txt

5. Run the Analysis

Launch Jupyter Notebook:

jupyter notebook DMV.ipynb

📜 License

This project is licensed under the MIT License. See LICENSE for details.

---

💡 For questions or collaborations, contact Pablo López or connect on LinkedIn.