receives eitc

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 receives the Earned Income Tax Credit (EITC), a refundable tax credit for low- to moderate-income working individuals and families.

Trained on survey data, it incorporates employment duration, education, and neighborhood income patterns to identify likely recipients.

Organizations can use these scores for outreach around tax filing assistance, benefit awareness, and economic mobility programs.

  • approach: deep learning
  • training rows: 4,992
  • last updated: 2026-08-13 19:23:29

Validation

metric value
area under ROC curve 0.88
area under precision-recall curve 0.47
accuracy 0.57
precision 0.16
recall 0.96
f1 score 0.27
mean log loss 0.80
gini coefficient 0.76
brier skill score 0.27

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
10 122 0.48 0.47 0.003 5.87
9 121 0.13 0.17 0.04 1.63
8 121 0.07 0.06 0.001 0.82
7 121 0.07 0.03 0.04 0.92
6 121 0.02 0.02 0.002 0.31
5 121 0.008 0.02 0.009 0.10
4 121 0 0.01 0.01 0.00
3 121 0 0.01 0.01 0.00
2 121 0.02 0.007 0.01 0.20
1 121 0.008 0.002 0.006 0.10

Architecture

This model is a deep neural network trained with TensorFlow.

Network structure

layer units activation
↓ dense 128
↓ dropout (0.40 rate)
↓ dense 91
↓ dropout (0.40 rate)
↓ dense 54

Hyperparameter configuration

  • loss: binary_focal_crossentropy
  • optimizer: adamax
  • batch size: 64
  • activation: hard_sigmoid
  • label smoothing: 0
  • learning rate: 5e-04

Features

Importance

Sourcing

Agency for Healthcare Research and Quality

  • received snap / food stamps (%)

CDC WONDER

  • utility shut-off threat, past 12 mo (adults)

Revelio Labs

  • avg salary differential
  • current position salary
  • industry sector

State voter files (via TargetSmart)

  • age
  • registered republican

TargetSmart

  • child tax credit support score
  • children present score
  • climate change concern score
  • college graduate score
  • double negative democrat score
  • double negative gop score
  • midterm general turnout score
  • mormon score
  • partisan score
  • presidential general turnout score
  • pro-choice score
  • race: black
  • race: white
  • recession sensitivity score

U.S. Census American Community Survey

  • block group % high school diploma

Verisk (via TargetSmart)

  • charity volunteer
  • home value
  • net worth

Implementation

Prediction distribution

statistic value
mean 0.10
std dev 0.15
min 0
25th 0.01
median 0.03
75th 0.12
max 0.70

Segments

segment lowest score highest score share of records
Most likely 0.33 0.70 10.2%
Somewhat likely 0.18 0.32 9.4%
Least likely 0 0.17 80.4%

External validity

Our model indicates that 10% of people with scored records claim the Earned Income Tax Credit. Per the IRS EITC Statistics for Tax Year 2023, ~24M working families and individuals received the EITC. Applying the IRS filing-status breakdown (~19% married filing jointly), this works out to roughly 28–29M adults benefiting — about 11% of all U.S. adults.