Digital Life and Fertility Future

Explore how remote-work share, commute time saved, screen leisure, online social life, and in-person interaction shape projected fertility trends across U.S. states.

About

About this dashboard

This dashboard explores how projected fertility trends across U.S. states may shift under different digital-life scenarios, including remote-work time saved, screen leisure, online social life, and in-person interaction.

Users can compare a model-based reference path with alternative scenario paths and see whether each scenario places fertility above or below the reference path.

The dashboard covers all 50 states plus DC and provides state-year outputs for multiple predictive models.

Reference path

What the selected model projects for fertility if no extra digital-life scenario is added.

Scenario path

What fertility looks like after adding the selected scenario.

Scenario difference

Scenario path minus reference path. Positive values mean the scenario is above the reference path; negative values mean it is below.

Data & Methods

Data

The outcome is the General Fertility Rate: live births per 1,000 women aged 15–44. The dashboard uses state-year information to project fertility paths and compare how those paths change under different digital-life assumptions.

Digital-life proxies

The scenarios prioritize direct state-year measures where possible: ACS remote-work share, commute minutes, and ATUS-based digital media, screen leisure, in-person social, and care-burden minutes. When a scenario still uses a proxy, the dashboard labels it explicitly.

Projection models

Statistical baseline: Primary State benchmark. It uses observed fertility trends and state-level predictors to create the reference path. It is transparent, stable, and performs well out of sample for the main interpretation.

Tree ML benchmark: State robustness benchmark with strong held-out accuracy. In the Metro patterns view, it is the preferred MSA display model because it preserves metropolitan variation better.

Neural network benchmark: Exploratory prediction benchmark. It can capture more complex patterns, but it is less transparent and should be interpreted cautiously.

Scenario assumptions

Each scenario changes the projected fertility path by adding a transparent scenario adjustment to the model-based reference path. For example, the remote-work scenario uses the change in remote-work time saved relative to the reference path, while the screen-leisure scenario uses ATUS-based digital leisure minutes.

Validation

Model performance is checked using held-out prediction errors. Lower validation error means the model predicts historical fertility patterns more accurately, but good predictive performance does not make the scenario results causal.

Limitations

These are projections under assumptions, not causal estimates. The results should be read as scenario comparisons, not as predictions of what will definitely happen. Scenario outputs depend on model choice, proxy measures, and the assumptions selected by the user.

Download

Downloads

How fertility changes across states under this scenario

Click a state to see its trend below.

Scenario Horizon year

Scenario interpretation text updates here based on the selected scenario.

Selected state

Predictive benchmarks

Selected state trends

Loading state explorer

State

Compare digital-life scenarios

See how different digital-life changes could shape fertility across U.S. states.

How fertility changes across MSAs under this scenario

Select an MSA to see its trend below.

Counties are used only to display MSA-level estimates more clearly.

State filter MSA

Scenario Horizon year

Estimated MSA counties Metropolitan CBSA context only Micropolitan CBSA context only Outside CBSA / not in analytical universe

Displayed by county; estimated at MSA level.

This map shows MSA-level scenario estimates. For visibility, each MSA's estimate is displayed across the counties that belong to that MSA.

This map shows MSA-level scenario estimates. For visibility, each MSA’s estimate is displayed across the counties that belong to that MSA. Counties in micropolitan areas or outside CBSAs are shown only as geographic background unless an estimate is available.

Selected MSA

Predictive benchmarks

MSA estimates are exploratory and more uncertain than State estimates. Use this view to compare broad metropolitan patterns, not exact local forecasts for individual MSAs.

Selected MSA trends

Loading MSA explorer

Compare Scenarios

Compare two digital-life scenarios

Compare two scenarios for the active geography and selected display metric at the chosen horizon year.

Geography Model Horizon year Scenario A Scenario B

Scenario comparison = Scenario A difference − Scenario B difference

Comparisons follow the selected display metric for the active geography and horizon year.

Assumptions

Adjust assumptions

Adjust assumptions User-controlled

Why did the scenario change?

Direction of the scenario difference

Method

Data and method

Use the top Data & Methods panel for the full overview. This section keeps only the key interpretation notes and validation snapshot.