Loading Last-Leg Toolkit…

Please wait while resources load

Texas State University

The "Last Leg" Impaired Driving
Challenge

Driving after using alcohol, cannabis, or other drugs remains a persistent highway safety problem. One part of it is the "last leg" of a multimodal trip: a traveler drives to a station or park-and-ride lot, takes transit to an evening out, returns late at night, and still has a few miles to drive home. This toolkit turns research evidence into tools for screening, strategy selection, and action planning.

Potential last-leg crashes
identified in ten states
Of travelers would drive
in the baseline scenario
1,395 Public survey respondents,
weighted to national benchmarks
28% Lower odds of driving
with safe overnight parking

What This Toolkit Provides

State-Level Crash Analysis

Ten-state map of potential last-leg crashes, screening steps, timing, and community characteristics.

Evidence-Based Strategies

Strategies rated by practitioners and travelers, with the evidence from the behavioral models.

Survey Insights & Station Simulator

What 1,395 travelers say about barriers, waiting, price, and messages, with a station simulator built on the final choice model.

Action Plan Builder

Generate a plan with the layers of protection, a strategy package, partners, a pathway for your agency, and performance measures.

Texas State University • Portland State University • Exponent

Potential Last-Leg Crashes in Ten States

Per 1,000 substance-impaired crashes · select a state

Four screens: weekend night · near a transfer point · residential area · not a freeway
  1. 5 PM to 3 AM, Friday to Sunday nights and holiday weekends
  2. Within 2 miles of a park-and-ride lot, station, or airport
  3. 80% or more residential land in the block group or a neighbor
  4. Not on an interstate or freeway

Places for closer review, not confirmed last-leg trips. 2020 excluded.

Rate per 1,000 impaired crashes Holdout state (used to test the models) Not studied

State Summary

Select a state on the map.

—

All crashes

—

Study period

Substance-impaired crashes

—

 

Potential last-leg crashes

—

Passed all four screens

Per 1,000 impaired crashes

—

 

Share by hour

—

 

Per 1,000 impaired crashes

Ten-state rate

Ten-State Comparison

Screening results, not a ranking of risk

StateRegionImpaired crashes Potential last-legPer 1,000 impaired crashes

—

—

Screening steps

Crashes remaining after each step · % kept from the previous step

What the screening shows

—

Rates use impaired crashes as the base because impairment coding varies by state (2.2% of crashes in Florida, 9.5% in Pennsylvania).

Share of potential last-leg crashes by hour

State All ten states

Bar chart of the hourly share of potential last-leg crashes for the selected state and for all ten states.

—

Share by night of the week

Monday: holiday weekends only

Block groups near transfer points

Community characteristics associated with potential last-leg crashes

Regional models, 40,000 block groups · occurrence: 1 or more crashes · concentration: 3 or more

CharacteristicOccurrenceConcentration

 

Average block group characteristics in the study states

Check that a state's block groups fall in this range before applying the models

How well the community models classify block groups

Correct classification vs. a share-only benchmark · best used to rank areas for review

MeasureInternal validation: benchmarkModelsHoldout states: benchmarkModels

Seven Strategies

Practitioner ratings (n = 78) and traveler support (n = 1,388)

Barriers Practitioners Report

Mean severity, 1 to 5 · n = 78

Combined Evidence and Recommended Use

Stakeholders, travelers, and behavioral models

StrategyStakeholdersTravelersBehavioral modelsRecommended use

Traveler Segments and Emphasis

SegmentEvidenceMain barrierEmphasis

Key gap: travelers rank overnight parking second (54.8%), but only 2.0% of agencies use it.

Strategies Side by Side

Select a column to sort · practitioners n = 78 · travelers n = 1,388

Strategy

Impact and Feasibility

Practitioner mean ratings, 1 to 5 · dot size = agency use

National traveler survey • 1,395 respondents • weighted

What Travelers Say and What Changes Their Choice

What Would Travelers Do?

Would drive, by scenario

How Risky Is Driving After Drinking?

say very or extremely risky, yet would drive.

What Is Associated with Choosing to Drive

Odds ratios with 95% confidence intervals · below 1 = less likely to drive

Beliefs and Concerns

Share who agree or are concerned · n ≈ 1,395 · ±3.1 pts

Station Simulator

Probability that a traveler chooses to drive, from the final choice model

Station conditions

10 min45 min

Probability of choosing to drive

0%40%55%100%

points vs. an unimproved station ()

What changes the probability

Probability of driving by waiting time

Solid line: current station

What to consider at this station

Barriers to Using Alternatives

Conditions for a Comfortable Wait

Strategies Travelers Chose

Top threeRating

Messages Rated Most Likely to Influence

Share rating 4 or 5 of 5 · mean rating

Rated 4 or 5Mean

Cannabis Compared with Alcohol

Awareness of Existing Programs

Layers of Protection

A traveler drives only when every layer has a gap

Plan Inputs

Barriers you see locally
Resources already available

Action Plan Report

Updates when you change the settings or the state

Your report will appear here

Choose the settings on the left, then select Generate report. Select a state on the map to add its crash results.

Implementation Pathways by Capacity

CapacityStart withData neededKey partnersNext step

Performance Measures

TypeMeasureSourceFrequency

Evaluation Designs

DesignWhen to useStrengths and limits
Real-world evidence

Lessons from Campaigns and Safe-Ride Programs

Approach, reported result, and lesson for each program in the guidance. Results are as reported by each program.

Results as reported by each program · illustrative photos from Wikimedia Commons (credit on each photo)

About This Toolkit

Evidence and tools to reduce substance-impaired driving on the last leg of a trip: the drive home from a park-and-ride lot or station after an evening out.

Evidence

How to Use It

Research Team

Texas State University, Portland State University, and Exponent.