Transportation Analysis Methodology

This document describes the methodology, assumptions, and limitations of each analysis in the independent transportation study. Every assumption is documented so that corrections can be applied directly when better data becomes available.

General Approach

This is order-of-magnitude analysis, not route-level precision. We use publicly available data, published research, and documented assumptions to estimate the transportation implications of the approved reconfiguration (Option A) and the alternatives that were considered. All outputs are presented as ranges, not point estimates. The analysis provides the transportation data needed to plan implementation before fall 2026.

split-family analysis: Split-Family Count Model

Method: Grade boundary probability model with sibling co-enrollment rates.

Calculation: For each configuration, identify the grade boundary, compute the probability that a multi-child family spans it, multiply by the estimated number of multi-child families.

Script: pipeline/transport/split_family.py

McKinney-Vento analysis: McKinney-Vento Exposure

Method: Proportional eligibility rate applied to displaced student population, with incremental transport cost multiplier.

Legal basis: 42 U.S.C. Section 11432 (McKinney-Vento Homeless Assistance Act)

Script: pipeline/transport/mckinney_vento.py

staffing analysis: SEA Staffing Adequacy

Method: Compare confirmed driver count against route requirements per configuration.

The 3/30/2026 BoE meeting provided the actual driver count (20 drivers, confirmed by Director of Operations Mike Natalie). The available driver pool is 17-20 (20 confirmed, minus 0-3 potentially lost to the SEA cut from 100 to 86 FTE — cut allocation is unknown).

Driver gap by configuration:

All configurations show a shortfall. This is a structural finding: even the lowest-impact option requires more drivers than the district has.

Script: pipeline/transport/sea_staffing.py

bell schedule analysis: Bell Schedule Tier Analysis

Method: Qualitative feasibility assessment of bus tier requirements under each configuration.

The 3/30/2026 meeting confirmed the driver schedule as 7 AM-4 PM with an idle window from 9:30-1:30. This is consistent with the existing 3-tier structure (HS, MS, elementary) and confirms that a 3-tier system is feasible with the current schedule. A 4th tier would require extending the driver workday or adding drivers.

Script: pipeline/transport/bell_schedule.py

care-gap analysis: Before/After Care Gap

Method: Combine split-family counts with bell schedule scenarios to estimate new care demand, price at market rates.

Script: pipeline/transport/care_gap.py

configuration comparison: Configuration Comparison

Method: Assembly of all metric outputs into structured comparison. Route expansion cost estimated from DOE per-pupil data scaled by route increase percentage.

The sensitivity analysis compares FY25 baseline ($2.99M) against FY23 baseline ($2.18M) to account for the anomalous 50.2% transport cost increase. All configurations are reported under both baselines. District and family costs are separated explicitly.

Sensitivity output (Option A — approved configuration):

Script: pipeline/transport/configuration_comparison.py

Multi-Year DOE Transport Cost Data

Maine DOE Resident Expenditure Per Pupil reports for FY23-FY25 reveal that South Portland’s transport costs are rising anomalously fast:

South Portland’s per-pupil transport costs rose 50.2% in two years — nearly 5x the state average — while enrollment fell 8.6%. The cause is unknown (contract renegotiation, SPED transport growth, or reporting reclassification are possibilities). This affects the route expansion estimate in two ways:

1. If the FY25 baseline is correct: The route expansion estimates stand, but the district is already spending significantly more on transport than peers — further increases compound on an already-stressed budget line.
2. If FY23 is more representative: The baseline should be ~$2.18M, not $2.99M, and route expansion estimates shrink accordingly. The sensitivity analysis presents both.

FY27 budget context: The district’s FY27 budget book (Cost Center 8) sets transportation at $3,326,606 — up 17.5% from FY26’s $2,830,679. This reflects contractual wage/benefit increases and includes $165K in reductions ($130K overtime, $35K vehicle supplies). It is a baseline budget with no reconfiguration adjustment. The route expansion estimates in this analysis use the DOE-derived baseline ($2.99M at FY25 rates) for consistency with the per-pupil methodology. If calculated against the district’s own $3.33M figure, the route expansion cost for Option A would be ~$832K instead of $748K.

Sources of Error

Ranked by estimated impact on the final comparison (highest first):

1. Route expansion cost — the largest and softest number. Derived from aggregate per-pupil data with adjustment heuristics (+10% per lost building, +15% per additional grade band), not route-level analysis. This is the single assumption most likely to move the total by a large amount in either direction.
2. Care gap rates — the 30%/85% care-need rates for 3-tier/4-tier schedules are logistics estimates, not survey data. If fewer split families actually need paid care, the family-borne cost drops. If more do (or if informal care networks are thinner than assumed), it rises.
3. McKinney-Vento displacement counts — the number of students who change schools under each option is estimated from the configuration structure, not from actual redistricting plans. The displacement estimate drives the MV cost.
4. Walkability assumptions — the 35-40% walker rate affects how many students need bus service. Actual walk zones are not publicly available; GIS-derived walkability from the redistricting tool is an approximation.
5. Sibling co-enrollment rates — the 30-40% multi-child family rate is a national average. South Portland’s actual rate could be higher or lower.
6. FY25 transport baseline anomaly — South Portland’s 50.2% per-pupil increase FY23-FY25 is unexplained. If the FY25 figure includes one-time costs or reclassified expenditures, the baseline is inflated and route expansion costs would be lower. The sensitivity analysis brackets this.
7. Driver cut allocation — the SEA cut (100 to 86 FTE) will remove 14 positions, but which positions is unknown. If drivers are spared entirely, the available pool stays at 20. If 3 drivers are cut, it drops to 17. The 17-20 range reflects this uncertainty.

All figures are order-of-magnitude estimates. The actual values could fall outside the stated ranges.

What District Data Would Change

The following data, all held by the district, would substantially improve this analysis:

1. Family-level enrollment — Exact split-family counts (replaces sibling rate estimates)
2. McKinney-Vento roster by school — Exact MV student counts (replaces proportional estimates)
3. SEA position-level cut list — Which of the 14 cut FTEs are drivers (replaces the 17-20 range)
4. Fleet size and route data — Actual route requirements and tier feasibility
5. Care waitlist numbers — Quantified unmet demand (replaces binary FULL/Available)
6. Walk zone policy — Not publicly available; needed for walker/rider classification
7. Transport consultant analysis — Confirmed “underway with a partner” at 3/30/2026 meeting; scope and timeline unknown

Reproducibility

All calculation scripts are in pipeline/transport/. All input data is in data/ and docs/troves/. The analysis can be fully reproduced from these inputs. Running any script produces both human-readable output and machine-readable JSON in data/.