Skip to content

Sensitivity Testing

Overview

Sensitivity testing evaluates whether ABM3 responds to changes in inputs and policies in a logical and defensible manner. Rather than testing model accuracy against observed data (which is the role of validation), sensitivity testing assesses the internal consistency and reasonableness of model behavior. A well-performing model should produce directionally correct and proportionally reasonable responses when key inputs are changed.

A broad range of sensitivity tests were conducted prior to using ABM3 for the 2025 Regional Plan.

Policy Sensitivity

Policy sensitivity tests examine how the model responds to changes in travel demand management (TDM) strategies and pricing mechanisms. Tests include:

  • Auto Operating Cost (AOC) increases — testing whether vehicle miles traveled (VMT) and mode shares respond appropriately to higher auto operating costs
  • Parking pricing — evaluating shifts in mode choice and destination choice when parking costs increase in employment-dense areas
  • Transit fare changes — assessing ridership sensitivity to fare increases and decreases
  • Managed lane pricing — testing diversion between general purpose and tolled facilities

Land Use and Demographic Sensitivity

These tests assess the model’s response to changes in the spatial distribution of population and employment, as well as shifts in household characteristics. Tests include:

  • Household income shifts — examining changes in auto ownership, mode choice, and trip generation under higher and lower income distributions
  • Employment density changes — testing whether trip attraction patterns respond appropriately to changes in job locations
  • Transit-oriented development — evaluating mode share and VMT responses to increased density near transit stations

Other

Several other tests were conducting to assess model response to teleworking levels, Electric Vehicle (EV) rebates, EV charging infrastructure, and microtransit service levels.

Further Reading

ABM3 Sensitivity Testing Report (2024)