How maternal education shapes fertility patterns and infant health across the United States (2016 – 2021).
We stitched together 6 years of provisional NCHS records (≈ 20 M births), cleaned 1.2 GB of raw CSVs into a single tidy table, and fed it through a fully reproducible R × tidyverse workflow. The result is an interactive report that quantifies regional disparities, isolates key predictors, and pin-points policy levers for healthier outcomes.
♡ Research questions
Education → birth outcomes
How does maternal schooling predict birth weight and maternal age across states and years?Education → family size
Are mothers with lower formal education more likely to have multiple children than highly-educated mothers?
♡ Functions & methods
| Stage | Key tools & notes |
|---|---|
| Cleaning / prep | readr::read_csv() · dplyr::mutate() / as_tibble() · na.omit() |
| Descriptive stats | group_by() + summarise() for year-by-state aggregates |
| Inferential stats | One-way ANOVA (aov) + linear / interaction models (lm) |
| Classification | Logistic glm predicting High_EDU (Bachelor’s +) from age & birth-weight |
| Visualisation | ggplot2 bar / violin / scatter + facet_wrap() and geom_text_repel() |
♡ Code & data
♡ Headline findings
Education inversely tracks fertility.
High-school-only and some college mothers account for most births, while PhD / MD holders delay childbirth and have fewer children overall.Higher education ⇒ healthier babies.
Each jump in schooling adds ≈ 85 g to mean birth weight (p < 0.001).Age amplifies the effect.
In the interaction model, every extra year of maternal age adds +41 g for PhD mothers but only +1 g for high-school drop-outs.Regional split.
New England & the West host the most college-educated, older mothers; the Deep South clusters at younger ages & lower schooling.
♡ Policy take-aways
Education is a lever.
Keeping women in school correlates with later, healthier births.Target by geography.
Southern states stand to gain most from prenatal-care subsidies and health-literacy programs.Next steps.
Add paternal education and local healthcare spend to explain the remaining variance.