has children at home

WarningStaging

Built from the latest training run on 2026-09-30 17:45 UTC, not from a released model set. The released documentation is at https://matter.indigo.engineering/docs/.

Overview

This model predicts whether a voter has children living at home.

Trained on survey data, it uses age, household composition scores, and social attitude signals to identify parents and guardians.

These scores help organizations target outreach around education policy, child care, family tax credits, and children’s health coverage.

  • approach: dnn classifier
  • training rows: 5,375
  • last updated: 2026-09-06 18:55:12

Validation

metric value
area under ROC curve 0.85
area under precision-recall curve 0.60
accuracy 0.83
precision 0.64
recall 0.47
f1 score 0.54
mean log loss 0.37
gini coefficient 0.70
brier skill score 0.30

Lift Chart

Each bar shows how much more common the outcome is within a score decile than in the population overall. Bars taller than the dashed line (lift of 1) mean the model’s top-ranked groups concentrate the outcome; a steady decline from decile 1 to 10 indicates the model is ranking effectively.

Calibration Plot

Bars of similar height within each decile indicate well-calibrated predictions. Systematic differences reveal where the model over- or under-predicts.

Cumulative Gains

The further the model curve bows above the diagonal, the better the model concentrates the outcome in its top-ranked predictions.

Decile Analysis Table

decile count mean actual mean predicted mae lift
10 138 0.70 0.75 0.05 3.33
9 137 0.53 0.51 0.02 2.52
8 137 0.34 0.30 0.05 1.63
7 138 0.22 0.16 0.06 1.06
6 137 0.10 0.10 0.003 0.48
5 137 0.04 0.07 0.03 0.21
4 138 0.07 0.06 0.001 0.31
3 137 0.02 0.06 0.04 0.10
2 137 0.04 0.05 0.01 0.21
1 138 0.03 0.05 0.02 0.14

Architecture

This model is a deep neural network classifier trained with BigQuery ML.

Hyperparameter configuration

  • max iterations: 20
  • learn rate: 0.0015
  • l1 regularization: 0
  • l2 regularization: 0
  • min relative progress: 0.01
  • warm start: FALSE
  • early stop: TRUE
  • input label columns: family_has_children_at_home
  • data split method: AUTO_SPLIT
  • hidden units: 64, 32, 16
  • batch size: 128
  • dropout: 0.2206
  • num trials: 18
  • max parallel trials: 5
  • hparam tuning objectives: ROC_AUC
  • enable global explain: FALSE
  • tf version: 1.15
  • auto class weights: FALSE
  • activation fn: elu
  • optimizer: adagrad

Features

Importance

Sourcing

Agency for Healthcare Research and Quality

  • tract % medicare only

Attom Property Data API

  • bedrooms
  • rooms

CDC PLACES

  • any disability (adults)

Revelio Labs

  • avg tenure (days)
  • current position tenure (days)

State voter files (via TargetSmart)

  • age

TargetSmart

  • abortion issue motivation score
  • child tax credit support score
  • children present score
  • early vote / absentee vote score
  • marriage score
  • paid leave support score

U.S. Census American Community Survey

  • block group % age 0–17
  • block group % some college

U.S. Census CPS Annual Social and Economic Supplement

  • current medicare coverage

Verisk (via TargetSmart)

  • children age 0–5 in household
  • senior present in household

Implementation

Prediction distribution

statistic value
mean 0.30
std dev 0.27
min 0.04
25th 0.06
median 0.17
75th 0.52
max 1

Segments

segment lowest score highest score share of records
Most likely 0.45 1 29.4%
Somewhat likely 0.11 0.44 30.5%
Least likely 0.04 0.10 40.1%

External validity

Our model indicates that 30% of people with scored records have their own children under 18 at home. Census ACS 2023 (B11005 / B23008) and Pew Research place the share of U.S. adults living with their own children under 18 at ~27%.