Trends in New Mexico School Districts Serving Low-Income Communities

Basic Information

• Paper ID: 2510.09993
• Title: Trends in New Mexico School Districts Serving Low-Income Communities
• Authors: U.E. Nelson, O.J. Okeke
• Classification: physics.soc-ph (Sociophysics)
• Publication Date: October 14, 2025
• Paper Link: https://arxiv.org/abs/2510.09993

Abstract

This study employs a mixed-methods geospatial framework integrating longitudinal enrollment data from 2019-2023, school district poverty indicators from the Small Area Income and Poverty Estimates (SAIPE), and community demographic data from the American Community Survey (ACS) to comprehensively analyze enrollment trends in New Mexico public school districts and their socioeconomic drivers. The research reveals that while statewide K-12 enrollment has declined overall, this decline is not uniformly distributed. School districts with high concentrations of low-income students experienced more pronounced losses; however, the relationship between poverty and enrollment change at the state level is complex and weak (r=0.011). Spatial analysis reveals a checkerboard pattern of growth and decline, suggesting that local factors such as housing instability, labor market conditions, and programmatic changes are important drivers. The study identifies 18 high-risk districts requiring targeted stabilization efforts.

Research Background and Motivation

Core Research Questions

This study addresses the enrollment decline problem facing New Mexico's public education system, with particular attention to the disproportionate impacts on school districts serving low-income communities.

Significance of the Problem

1. Severity of Enrollment Decline: Statewide K-12 enrollment declined by over 5% from 2019-2023, with projections indicating an additional 1.2% decline by 2025
2. Unequal Distribution: School districts where over 60% of students qualify for free or reduced-price lunch experienced average losses of 7.8%, compared to only 2.1% in affluent districts
3. Resource Disparities: In 2023, per-pupil funding in low-income districts was approximately $1,200 less than in affluent districts
4. Multiple Challenges: Chronic absenteeism rates reach 32%, with persistent 15-percentage-point gaps in reading and mathematics proficiency

Limitations of Existing Research

While national studies have documented post-pandemic changes in school engagement, the specific spatial and socioeconomic characteristics of New Mexico remain insufficiently researched.

Research Motivation

This study fills this research gap through a mixed-methods geospatial framework, providing evidence-based recommendations for policymaking.

Core Contributions

1. Established a Comprehensive Analytical Framework: Created a reproducible Python workflow integrating multi-source data (CCD, SAIPE, ACS)
2. Quantified Spatial Disparities: Systematically analyzed spatial differences in enrollment trajectories between low-income and affluent districts from 2019-2025
3. Identified High-Risk Districts: Identified 18 high-risk districts requiring immediate attention through composite indicators
4. Proposed a Dual-Track Policy Approach: Immediate stabilization support for high-risk districts combined with long-term structural investments for all high-need districts
5. Revealed Complex Relationships: Demonstrated that poverty concentration and enrollment loss are not simply deterministic

Methodology

Task Definition

Input: School district enrollment data, poverty indicators, community demographic variables Output: Enrollment change trend analysis, risk stratification, policy recommendations Constraints: Geographic boundary consistency, temporal series completeness

Data Architecture

Core Data Sources

1. NCES Common Core of Data (CCD): Enrollment time series accessed through the Urban Institute Education Data Explorer
2. Small Area Income and Poverty Estimates (SAIPE): School-age poverty statistics by district
3. American Community Survey (ACS): Community-level socioeconomic indicators
4. TIGER/Line Boundary Data: 2023 school district geographic boundaries

Data Processing Pipeline

# Key processing steps
1. Geographic identifier standardization (GEOID/LEAID 7-digit zero-padding)
2. Population-weighted areal interpolation (tract-to-district aggregation)
3. Equal-area coordinate system projection (EPSG:5070)
4. Panel data construction (district-year observations)

Analytical Methods

1. Descriptive Analysis

• Choropleth Maps: Visualization of enrollment numbers and annual changes
• Small Multiples Trend Lines: Trajectory analysis of largest movers
• Bivariate Maps: Comparative mapping of poverty and change

2. Statistical Modeling

Fixed Effects Model:

YoY_change = α + β₁×SAIPE_poverty + β₂×ACS_controls + δᵢ + γₜ + εᵢₜ

Where:

• δᵢ: District fixed effects
• γₜ: Year fixed effects
• εᵢₜ: District-clustered standard errors

3. Risk Stratification

Composite Risk Score: Early warning system combining poverty intensity and enrollment dynamics

Technical Innovations

1. Population-Weighted Spatial Aggregation: 100% use of population weights rather than area weights for improved accuracy
2. Multi-Scale Poverty Definition: Integration of tract-level (≥25% threshold) and district-level (state 80th percentile) poverty measures
3. Robustness Checks: Tail trimming at ±30% change, multi-threshold sensitivity analysis
4. Reproducible Workflow: Open-source Python toolchain ensuring transparency

Experimental Setup

Dataset Characteristics

• Coverage: 89 school districts in New Mexico
• Time Span: 2018-2023 (focus on 2019-2023)
• Sample Size: Total enrollment of 240,510 in 2023
• Poverty Distribution: Average school-age poverty rate of 30.26%, with 38 districts exceeding 30%

Evaluation Metrics

1. Enrollment Change: Annual absolute and percentage changes, compound annual growth rate (CAGR)
2. Correlation Analysis: Pearson correlation coefficient between poverty rate and enrollment change
3. Spatial Patterns: Comparative analysis of adjacent districts
4. Risk Stratification: Accuracy of high-risk district identification

Quality Assurance

1. Data Validation: Cross-verification of CCD with PED STARS reports
2. Outlier Detection: IQR method for identifying statistical anomalies
3. Diagnostic Tests: Residual analysis, Q-Q plot normality testing
4. Spatial Accuracy: 100% coverage of population-weighted interpolation

Experimental Results

Main Findings

1. Enrollment Trajectories (2018-2023)

• Overall Trend: Relative stability from 2018-2022, moderate contraction in 2023
• Recent Change: Average annual change of -1.41 percentage points in 2023
• Scale Effects: Smaller systems show greater percentage fluctuations

2. Poverty Landscape

• Statewide Average: School-age poverty rate of 30.26%, with heavy-tailed distribution
• High Poverty Concentration: 38 districts exceed 30%, 21 exceed 40%
• Four-Quadrant Classification:
  • Low poverty & stable growth: 27 districts (30.3%)
  • Low poverty & decline: 24 districts (27.0%)
  • High poverty & stable growth: 22 districts (24.7%)
  • High poverty & decline: 16 districts (18.0%)

3. Relationship Between Poverty and Enrollment Change

Key Finding: 2023 correlation coefficient r=0.011, near-zero correlation

Extreme Case Comparisons:

• Largest Decline: Ruidoso (-9.35%, poverty rate 9.4%), Zuni (-8.57%, 56.1%)
• Largest Growth: Belen (+10.99%, 18.0%), Santa Fe (+10.36%, 54.5%)

Spatial Pattern Analysis

The study reveals a "checkerboard" pattern of growth and decline, where adjacent high-poverty districts may exhibit opposite trends, indicating the importance of local factors (housing mobility, labor cycles, programmatic composition).

Risk Stratification Results

High-Risk Districts Identified (18 total):

• Clovis (Score 5, poverty rate 36.8%, change -7.9%)
• Hobbs (Score 5, poverty rate 57.1%, change -5.4%)
• Zuni (Score 5, poverty rate 56.1%, change -8.6%)

Related Work

This study builds upon the following research foundations:

1. National Education Statistics: Post-pandemic enrollment trend studies
2. Educational Equity Research: Resource allocation and student outcomes relationships
3. Rural Education: Geographic isolation impacts on educational opportunity
4. Geospatial Analysis: GIS applications in education policy research

Compared to existing work, this study's advantages include:

• Comprehensive framework integrating multi-source data
• Spatially explicit analytical methods
• Actionable policy recommendations

Conclusions and Discussion

Main Conclusions

1. Complexity Beyond Simple Relationships: Poverty intensity alone does not determine enrollment trajectories (r=0.011)
2. Importance of Local Factors: Adjacent high-poverty districts may exhibit opposite trends
3. Necessity of Stratified Interventions: 18 high-risk districts require immediate attention
4. Dual-Track Policy Needs: Immediate stabilization plus long-term structural investment

Policy Recommendations

Immediate Stabilization Track

• Flexible funding support for 18 high-risk districts
• Attendance support and rapid monitoring indicators
• Improved transportation and teacher recruitment

Long-Term Structural Investment Track

• Address root causes of concentrated disadvantage
• Enhance educational program accessibility in rural and border regions
• Establish sustainable resource pipelines

Limitations

1. Time Range Constraints: Primarily focuses on short-term changes from 2019-2023
2. Causal Inference: Observational study design limits causal interpretation
3. Local Factors: Cannot fully capture all district-specific driving factors
4. Policy Lag: Policy intervention effects may require longer timeframes to manifest

Future Directions

1. Longitudinal Tracking: Extend observation period to capture long-term trends
2. Mechanism Research: Deeper exploration of local driving factors
3. Intervention Evaluation: Assess effectiveness of policy interventions
4. Method Generalization: Apply framework to other similar states

In-Depth Evaluation

Strengths

1. Methodological Rigor:
   • Multi-source data integration ensures analytical comprehensiveness
   • Fixed effects model controls for unobserved heterogeneity
   • Robustness checks enhance result credibility
2. High Practical Value:
   • Identifies specific high-risk districts
   • Provides actionable policy recommendations
   • Establishes early warning system
3. Technical Innovation:
   • Population-weighted spatial aggregation improves accuracy
   • Open-source reproducible workflow
   • Multi-scale poverty measurement
4. Strong Policy Relevance:
   • Directly addresses real policy needs
   • Balances short-term and long-term strategies
   • Considers resource allocation efficiency

Limitations

1. Causal Inference Constraints:
   • Observational study design limits causal interpretation
   • Omitted variables may affect results
2. Relatively Short Time Span:
   • Main analysis period 2019-2023 may be influenced by pandemic-specific effects
   • Long-term trend assessment requires additional data
3. Local Heterogeneity:
   • While identifying importance of local factors, specific mechanisms are insufficiently analyzed
   • Different regions may require differentiated strategies
4. Absent Intervention Evaluation:
   • Does not evaluate effectiveness of existing policy interventions
   • Lacks cost-benefit analysis

Impact

1. Academic Contributions:
   • Provides new analytical framework for educational geography
   • Enriches methodology of educational equity research
   • Offers replicable template for other states
2. Policy Impact:
   • Provides direct guidance for New Mexico education policy
   • Supports evidence-based resource allocation decisions
   • May influence federal education policy development
3. Method Generalization:
   • Open-source tools increase research accessibility
   • Applicable to other geographic and social contexts
   • Establishes foundation for real-time monitoring systems

Applicable Scenarios

1. Direct Application:
   • New Mexico Department of Education policy development
   • School district resource allocation optimization
   • Early warning system development
2. Generalized Application:
   • Other rural states (Arizona, Mississippi, etc.)
   • International educational equity research
   • Urban education system analysis
3. Methodological Application:
   • Spatial analysis of other public services
   • Social policy impact assessment
   • Geographically weighted regression modeling

References

The paper cites 22 important references covering multiple domains including education statistics, policy analysis, spatial analysis methods, and open-source tools, providing a solid theoretical and methodological foundation for the research.

Overall Assessment: This is a high-quality applied research paper that is methodologically rigorous and highly policy-relevant, providing valuable insights and actionable recommendations for addressing educational equity issues in New Mexico. The study's reproducibility and open-source nature also establish a strong foundation for subsequent research.