15.4 ITE Trip Generation Handbook: Average Rates vs. Regression Equations
Key Takeaways
- The ITE Trip Generation Manual (11th Edition) provides empirical trip generation rates and fitted curve regression equations based on Land Use Codes (LUC) and independent variables X (e.g., 1,000 sq ft GLA, dwelling units, employees).
- Per the ITE selection criteria protocol, the fitted regression curve equation is preferred over the weighted average rate when: (1) sample size N >= 4; (2) coefficient of determination R^2 >= 0.75; and (3) the independent variable X falls within the observed data range.
- When R^2 < 0.75, N < 4, or X lies outside the data boundaries, the weighted average rate (T = R * X) is used, provided the standard deviation is reasonable and non-linear scale effects are absent.
- Pass-by trips are intermediate stops made by vehicles already traveling on the adjacent street; they add turning movement volumes to site access driveways but do NOT add new vehicle trips to the surrounding regional through network.
- Mixed-use development (MXD) internal capture matrices account for unconstrained and constrained trips interacting between on-site complementary land uses without using the public street network.
15.4 ITE Trip Generation Handbook: Average Rates vs. Regression Equations
PTOE Exam Focus: Mastery of the ITE Trip Generation Manual (11th Edition) and Handbook is essential for Domain 5. Candidates must know the strict 3-part statistical decision protocol for choosing between weighted average rates and fitted curve regression equations ($N \ge 4$, $R^2 \ge 0.75$, within data range), calculate trip generation using linear and logarithmic equations, differentiate Primary, Pass-By, and Diverted Linked trips, adjust driveway versus network link volumes, and calculate internal capture reductions in mixed-use developments.
1. Structure of the ITE Trip Generation Manual (11th Edition)
The ITE Trip Generation Manual provides empirical trip generation data collected across thousands of nationwide site studies. Data is categorized by standardized Land Use Codes (LUC) grouped into 10 major series:
- 100s: Industrial / Manufacturing / Warehousing
- 200s: Residential (Single-Family, Multi-Family, Senior Housing)
- 300s: Group Care (Assisted Living, Nursing Homes)
- 400s: Recreational (Parks, Health Clubs, Golf Courses)
- 500s: Institutional (Schools, Universities, Churches, Day Care)
- 600s: Medical (Hospitals, Medical Office Buildings, Clinics)
- 700s: Office (General Office, Corporate Headquarters, R&D)
- 800s: Retail (Shopping Centers, Supermarkets, Pharmacies)
- 900s: Services (Banks, Restaurants, Gas Stations, Auto Care)
Independent Variables ($X$)
Trip generation is estimated as a function of one or more physical scale characteristics ($X$), most commonly:
- Gross Floor Area (GFA) / Gross Leasable Area (GLA): Expressed in units of $1,000\text{ sq ft}$ ($KSF$).
- Dwelling Units (DU): For residential developments.
- Employees / Beds / Fueling Positions / Hotel Rooms: For specialized uses.
2. Average Rates vs. Fitted Curve Regression Equations
For each land use and analysis period, ITE provides two distinct mathematical models:
+-----------------------------------------------------------------------------------+
| ITE TRIP GENERATION MATHEMATICAL MODELS |
| |
| 1. Weighted Average Rate Model: |
| T = R × X |
| Where R = (Sum of Trips across all sites) / (Sum of X across all sites). |
| Assumes a strictly linear relationship passing through the origin (0,0). |
| |
| 2. Fitted Curve Regression Models: |
| • Linear Equation: T = a × X + b |
| • Logarithmic / Power Curve Equation: ln(T) = a × ln(X) + b |
| (Expressed algebraically as: T = e^b × X^a) |
+-----------------------------------------------------------------------------------+
3. The 3-Part ITE Selection Criteria Protocol
To ensure statistical rigor and prevent misapplication, ITE establishes a strict hierarchical protocol for choosing between the fitted curve equation and the weighted average rate:
+-----------------------------------------------------------------------------+
| ITE FITTED CURVE VS RATE SELECTION PROTOCOL |
| |
| USE THE FITTED CURVE REGRESSION EQUATION IF ALL 3 CONDITIONS ARE MET: |
| 1. Sample Size: N >= 4 study sites (preferably N >= 20). |
| 2. Coefficient of Determination: R^2 >= 0.75 (explains >= 75% variance). |
| 3. Data Range: The proposed development size X falls WITHIN the range |
| of independent variables observed in the ITE dataset. |
| |
| OTHERWISE, USE THE WEIGHTED AVERAGE RATE (T = R × X), PROVIDED: |
| • The standard deviation is low (standard deviation / mean rate <= 0.55). |
| • The data plot does not show a pronounced non-linear curvature. |
| • Never extrapolate a regression equation beyond the data boundaries! |
+-----------------------------------------------------------------------------+
Why Equations are Preferred When $R^2 \ge 0.75$
Large developments exhibit economies of scale (diminishing marginal trip generation per unit of floor area). A linear rate ($T = R \cdot X$) forces a straight line through the origin, overestimating trips for large developments and underestimating trips for small sites. A logarithmic fitted curve captures this non-linear saturation effect.
4. Trip Typology: Primary, Pass-By, & Diverted Linked Trips
Total trips arriving at a site are classified into three distinct categories based on trip origin, destination, and routing behavior:
+-----------------------------------------------------------------------------+
| ITE TRIP TYPE CLASSIFICATION |
| |
| 1. Primary Trips: Direct travel between origin and destination with the |
| sole primary purpose of visiting the site. Adds NEW trips to the |
| regional roadway network and site driveways. |
| |
| 2. Pass-By Trips: Trips made as intermediate stops by motorists ALREADY |
| traveling on the adjacent roadway (e.g., stopping at a gas station or |
| coffee shop during an existing commute). |
| • Adds turning volumes to site driveways. |
| • Does NOT add new trips to adjacent through road links! |
| |
| 3. Diverted Linked Trips: Trips diverted from another roadway in the |
| vicinity to visit the site before resuming their original journey. |
| Adds new trips to the local street segment but not the regional network.|
+-----------------------------------------------------------------------------+
Driveway vs. Adjacent Street Link Volume Adjustment for Pass-By Trips
For a retail site with $T_{\text{gross}} = 200\text{ veh/h}$ and an ITE pass-by rate of $40%$:
- Pass-By Trips: $200 \times 0.40 = 80\text{ veh/h}$.
- Primary Trips (New External Trips): $200 \times (1 - 0.40) = 120\text{ veh/h}$.
- Driveway Volume: Sized for full $200\text{ veh/h}$ (all 200 vehicles enter and exit the driveway).
- Adjacent Roadway Through Link Volume: Sized for only $120\text{ veh/h}$ new trips (the 80 pass-by trips were already in the through traffic stream and are simply converted to right/left turns at the driveway).
5. Mixed-Use Development (MXD) Internal Capture
In a Multi-Use Development (MXD) containing complementary land uses (e.g., residential apartments + retail shops + office space + restaurants), a significant portion of trips begin and end entirely within the development without accessing the public external street system.
The ITE Internal Capture Estimation Procedure
- Unconstrained Internal Trips ($T_{\text{unconstrained}}$): Calculated using ITE interaction percentage matrices between land use pairs (e.g., percentage of retail patrons originating from on-site residential):
- Constrained Internal Trips ($T_{\text{internal}}$): The actual internal exchange is constrained by the smaller of the supply demand pairs:
- Net External New Trips Formulation:
ITE Trip Generation Selection Criteria & Model Comparison
| Model Type | Mathematical Formulation | Mandatory Selection Prerequisites | Key Strengths | Common Engineering Pitfalls |
|---|---|---|---|---|
| Fitted Curve Equation (Logarithmic / Power) | ln(T) = a × ln(X) + b -> T = e^b × X^a | N >= 4, R^2 >= 0.75, and X within survey data range | Accurately captures economies of scale and non-linear trip saturation | Extrapolating outside observed data bounds produces extreme errors |
| Fitted Curve Equation (Linear) | T = a × X + b | N >= 4, R^2 >= 0.75, and X within survey data range | Simple linear relationship with non-zero intercept | May produce positive trip estimates at X = 0 or negative values for small X |
| Weighted Average Rate | T = R × X | Used when R^2 < 0.75, N < 4, or X is outside equation data bounds | Straightforward; passes through origin (0,0) | Overestimates trips for large sites; understates trips for small sites |
| Pass-By Trip Adjustment | T_primary = T_gross × (1 - Pass-By %) | Applied only to retail/service uses with documented ITE pass-by data | Prevents double-counting existing corridor traffic on network links | Deducting pass-by trips from site access driveways (driveways must serve 100%) |
| MXD Internal Capture | T_external = T_gross - T_internal - T_passby | Multi-use sites with internal pedestrian/vehicular connectivity | Accounts for internal interaction between live/work/shop land uses | Double-counting internal capture across overlapping trip pairs |
According to the ITE Trip Generation Handbook selection criteria protocol, under which set of conditions should an engineer select the fitted curve regression equation instead of the weighted average rate?
How do pass-by trips affect traffic volumes on the surrounding public street network and site access driveways?
A proposed mixed-use development generates 500 gross vehicle trips during the PM peak hour. The ITE multi-use matrix establishes an internal capture reduction of 100 trips between the residential and retail uses. The retail component has an approved pass-by reduction of 120 trips. How many net new primary vehicle trips does the development add to the surrounding external roadway network?