/education-data-source-saipe

name: education-data-source-saipe
description: >-
  SAIPE — annual Census poverty estimates for school districts (Portal; county/state not in Portal). Use for district poverty, Title I context, or trends. ~18-month lag. No race/ethnicity disaggregation at district level — use ACS 5-year for that.
metadata:
  audience: any-agent
  domain: data-source
  skill-authored: "2026-02-09"
  skill-last-updated: "2026-02-09"

SAIPE Data Source Reference

Census Bureau Small Area Income and Poverty Estimates (SAIPE) — annual model-based poverty estimates for school districts (Portal mirror; county and state data not in Portal). Use when district-level poverty is needed for Title I allocation interpretation, annual poverty trend analysis, or school-age children in poverty estimates. Estimates have ~18-month lag and no race/ethnicity disaggregation at district level — use ACS 5-year for race-disaggregated poverty.

Reference for understanding Census Bureau poverty estimates for school districts, counties, and states. SAIPE is the only annual, district-level poverty source and the legally mandated basis for Title I education funding allocations.

> **CRITICAL: Value Encoding** > > This document describes **Education Data Portal** integer encodings, which differ from Census Bureau raw file formats. The Portal uses integers for FIPS codes and standard missing data conventions. > > | Context | FIPS Alabama | FIPS California | Missing | Suppressed | > |---------|--------------|-----------------|---------|------------| > | **Portal (integers)** | `1` | `6` | `-1` | `-3` | > | Census raw files | `01` (string) | `06` (string) | varies | varies | > > **Key difference:** Portal FIPS codes are integers (no leading zeros), while Census files use 2-character strings. > > See `./references/variable-definitions.md` for complete encoding tables.

What is SAIPE?

SAIPE is the Census Bureau's program for producing **model-based** estimates of income and poverty:

• **Primary purpose**: Provide annual poverty estimates for Title I education funding allocations
• **Collector**: U.S. Census Bureau
• **Coverage**: All 50 states, 3,100+ counties, 13,000+ school districts
• **Key measure**: Related children ages 5-17 in families in poverty
• **Update frequency**: Annual (released each December, ~18-month lag)
• **Available years**: 1995-2023 (gaps at 1996, 1998; annual from 1999)
• **Primary identifier**: FIPS code + LEAID (district ID)
• **Methodology**: Model-based — combines ACS survey data with administrative records (IRS tax returns, SNAP data) using regression models with "shrinkage" techniques; school district estimates are allocated from county totals using within-county shares; all estimates contain uncertainty and confidence intervals are essential
• **Available through**: Education Data Portal mirrors (district-level only; state and county SAIPE are not in Portal mirrors — see Data Access section)

Reference File Structure

| File | Purpose | When to Read | |------|---------|--------------| | `estimation-methodology.md` | How state/county models work | Understanding model inputs and outputs | | `school-district-estimates.md` | How district estimates are derived | Working with school district data | | `variable-definitions.md` | Variables, codes, population universes | Interpreting specific data fields | | `data-quality.md` | Uncertainty, CV, limitations | Assessing estimate reliability | | `historical-changes.md` | Methodology changes over time | Comparing across years | | `comparison-other-sources.md` | SAIPE vs ACS, FRPL, CPS | Choosing between data sources |

Decision Trees

What do I need to understand?

Understanding SAIPE?
├─ How are estimates created?
│   ├─ State/county models → ./references/estimation-methodology.md
│   └─ School district shares → ./references/school-district-estimates.md
├─ What variables are available?
│   └─ Variable definitions → ./references/variable-definitions.md
├─ How reliable are estimates?
│   ├─ Confidence intervals → ./references/data-quality.md
│   └─ Small district uncertainty → ./references/data-quality.md
├─ Comparing data sources?
│   ├─ SAIPE vs FRPL → ./references/comparison-other-sources.md
│   ├─ SAIPE vs ACS → ./references/comparison-other-sources.md
│   └─ Why estimates differ → ./references/comparison-other-sources.md
└─ Year-to-year changes?
    ├─ Methodology breaks → ./references/historical-changes.md
    └─ Safe comparisons → ./references/historical-changes.md

Common research questions

Research question?
├─ District poverty rate for Title I
│   ├─ Use SAIPE (official source for Title I)
│   └─ Note: rates use different numerator/denominator universes
├─ Compare district poverty over time
│   ├─ Check methodology breaks → ./references/historical-changes.md
│   └─ Cannot compare school districts pre/post 2010
├─ Why doesn't SAIPE match FRPL?
│   └─ Different income thresholds → ./references/comparison-other-sources.md
├─ Poverty by race/ethnicity in districts
│   └─ SAIPE does NOT provide race breakdowns for districts
│       Use ACS 5-year estimates instead
└─ Very small district reliability
    └─ Check CV by population size → ./references/data-quality.md

Quick Reference: SAIPE Variables

CRITICAL: Field Name Prefix

All SAIPE estimate columns in the Education Data Portal use the `est_` prefix:

| Short Name | Portal Column Name | |------------|-------------------| | `population_total` | `est_population_total` | | `population_5_17` | `est_population_5_17` | | `population_5_17_poverty` | `est_population_5_17_poverty` | | `population_5_17_poverty_pct` | `est_population_5_17_poverty_pct` |

Key Identifiers

| ID | Format | Level | Example | Notes | |----|--------|-------|---------|-------| | `fips` | Integer | State | `6` | State FIPS code (no leading zeros in Portal) | | `leaid` | String | District | `0100005` | NCES district ID; join key to CCD | | `year` | Integer | Time | `2022` | Estimate reference yea