Module-2

Traffic Engineering Studies: Statistical Methods In Traffic Safety Analysis – Regression Methods, Poisson
Distribution, Chi- Squared Distribution, Statistical Comparisons- Traffic Management Measures And
Their Influence On Accident Prevention.

1. Explain how regression analysis is used in traffic safety studies. What are the benefits of using
regression models for predicting accident frequency at specific locations?
Regression analysis is a powerful statistical tool used extensively in traffic safety studies to
understand the relationships between crash frequency and various influencing factors, such as traffic
volume, road geometry, and environmental conditions. By modeling how different variables affect
accident occurrence, regression analysis helps researchers and traffic engineers predict crash patterns,
evaluate risk levels, and make informed decisions to improve road safety.
Use of Regression Analysis in Traffic Safety Studies: In traffic safety, regression models are commonly
applied to develop Safety Performance Functions (SPFs), which estimate the expected number of
crashes at a location based on its characteristics. These functions are used to identify high-risk areas,
evaluate countermeasures, and perform cost-benefit analyses. A typical regression model in this
context might relate accident frequency (dependent variable) to one or more independent variables
such as:
 Traffic volume (e.g., Average Daily Traffic - ADT)
 Roadway design features (e.g., number of lanes, shoulder width)
 Intersection type (e.g., signalized, unsignalized)
 Environmental factors (e.g., weather, lighting conditions)
Different types of regression models are used depending on the nature of the data. Since crash data
are usually count based and often over-dispersed (variance greater than the mean), Poisson regression
or Negative Binomial regression is preferred over standard linear regression. These models are well-
suited for predicting the frequency of rare events like road crashes.
Steps in Applying Regression Analysis in Traffic Safety:
 Data Collection: Collect historical crash data along with relevant variables for road segments or
intersections.
 Model Specification: Choose an appropriate regression model (Poisson, Negative Binomial, or Zero-
Inflated models) depending on the distribution of crash data.
 Parameter Estimation: Use statistical software to estimate the coefficients that show the strength
and direction of the relationship between each independent variable and crash frequency.
 Model Validation: Evaluate the model’s accuracy using statistical tests (e.g., goodness-of-fit,
likelihood ratio tests) and by comparing predicted crash values with actual crash data.
 Prediction and Analysis: Apply the model to new locations to estimate expected crash frequencies,
which can then inform safety prioritization and intervention planning.
Benefits of Using Regression Models for Predicting Accident Frequency:
 Data-Driven Decision-Making: Regression models provide objective, quantitative insights that help
identify high-risk locations based on expected crash frequencies rather than relying only on historical
crash counts.
 Control for Exposure: By including traffic volume as an independent variable, regression models
adjust for differences in exposure, allowing for fair comparisons between locations with varying traffic
levels.
 Identification of Contributing Factors: These models help isolate and quantify the effect of specific
roadway or environmental features on crash likelihood, supporting targeted improvements.
 Prediction for Untreated Sites: Regression models allow prediction of crash frequency at locations
without prior crash history, enabling proactive safety management.
 Support for Cost-Benefit Analysis: By estimating crash reduction potential, regression models help
assess the economic viability of proposed safety countermeasures. Regression analysis is a
cornerstone of modern traffic safety analysis, offering a statistically sound method to predict accident
frequency and identify influencing factors. By using regression models, transportation agencies can
prioritize interventions more effectively, implement evidence-based safety strategies, and ultimately
reduce traffic crashes and save lives.
2. Discuss the application of the Poisson distribution in modeling accident occurrences. Why is the
Poisson distribution commonly used in traffic accident studies?
The Poisson distribution is a discrete probability distribution that expresses the probability of
a given number of events (such as traffic accidents) occurring in a fixed interval of time or space,
assuming these events happen independently and with a constant average rate. In the context of traffic
safety, the Poisson distribution is widely used to model accident occurrences at specific locations, such
as intersections, road segments, or highways, over defined time periods.
Application of the Poisson Distribution in Traffic Accident Modeling: In traffic accident studies, the
primary objective is often to model the frequency of crashes and identify factors that influence crash
risk. Since traffic accidents are random, rare, and discrete events, the Poisson distribution provides a
suitable framework for statistical modeling.
Key Assumptions of the Poisson Model:
 Events are independent – One accident does not influence the probability of another.
 The average rate (λ) is constant – The mean number of accidents over a time or space interval
remains stable.
 Only one event can occur at a specific point in time or space – Two accidents occurring at exactly
the same instant or location are highly unlikely.
Using these assumptions, the Poisson distribution can be applied to estimate the probability of
observing a certain number of crashes (k) in a location or time frame using the formula:
> P(k; λ) = (λ^k \ e^-λ) / k
Where:
k is the number of crashes,
λ is the expected number of crashes (mean),
e is the base of the natural logarithm.
For example, if the expected number of crashes per year at a road intersection is 3 (λ = 3), the Poisson
model can calculate the probability of observing 0, 1, 2, or more crashes in any given year.
Why the Poisson Distribution is Commonly Used in Accident Studies:
 Modeling of Rare Events: Traffic accidents are generally infrequent at any specific location. The
Poisson distribution is particularly well-suited for modeling such rare and count-based events.
 Simplicity and Interpretability: The model is mathematically straightforward and easy to implement
using statistical software. The single parameter λ (mean crash rate) is intuitive and directly related to
crash frequency.
 Foundation for Advanced Models: While the basic Poisson model assumes equal mean and
variance, real world crash data often show overdispersion (variance > mean). In such cases, the Poisson
distribution forms the basis for more advanced models like the Negative Binomial distribution, which
introduces an extra parameter to account for variability.
 Useful for Safety Performance Functions (SPFs): Poisson regression models are used to develop
SPFs, which predict crash frequencies based on factors like traffic volume, road geometry, and
environmental conditions. These models help agencies estimate expected crashes and assess safety
performance.
 Support for Empirical Bayes Methods: The Poisson model is integral to the Empirical Bayes (EB)
approach, which combines predicted and observed crash data to better estimate expected crash
frequencies. This method reduces the impact of random fluctuations in crash data and supports
evidence-based safety decisions.
The Poisson distribution plays a foundational role in traffic accident modeling due to its
suitability for rare, discrete events. It allows researchers and traffic engineers to estimate crash
probabilities, develop predictive models, and support proactive safety planning. While simple, it serves
as a gateway to more complex and realistic models in modern traffic safety analysis.
3. What are the limitations of using a Poisson distribution in traffic crash data analysis, and how can these
be addressed?
The Poisson distribution is widely used in traffic crash data analysis to model the frequency of
crashes at specific locations over time. It is suitable for rare, discrete events and serves as a
foundational model in traffic safety studies. However, despite its usefulness, the Poisson model has
several limitations when applied to real-world crash data. Understanding these limitations is crucial
for selecting the appropriate analytical approach and improving the accuracy of traffic safety
evaluations.
Key Limitations of Using the Poisson Distribution:
1. Assumption of Equal Mean and Variance (Equidispersion): The Poisson distribution assumes
that the mean and variance of crash frequency are equal. In practice, traffic crash data often
exhibit overdispersion, where the variance exceeds the mean. This is due to unobserved
heterogeneity — factors such as weather, driver behavior, or time-of-day effects that are not
accounted for in the model but still influence crash frequency.
Impact: Underestimating the variance can lead to understated standard errors, inflated
significance levels, and misleading inferences, potentially resulting in the identification of non-
existent safety issues or overlooking real ones.
2. Inability to Handle Excess Zeros: Crash data at many road segments or intersections may show
zero crashes during the observation period. The standard Poisson model does not account for
these excess zeros, leading to poor model fit and inaccurate predictions.
Impact: The model may overpredict crashes at locations with zero counts or underpredict at
high-crash locations, weakening its reliability in safety performance evaluations.
3. Assumption of Independent Events: The Poisson distribution assumes that crash events are
independent of each other. However, in real-world settings, crashes may be correlated due to
spatial (nearby locations) or temporal (within short periods) effects.
Impact: Ignoring such correlations can compromise the model's accuracy and reduce its
effectiveness in identifying high-risk locations or assessing safety interventions.
4. No Room for Unobserved Heterogeneity: The Poisson model assumes a homogeneous
process, meaning all locations or time periods are assumed to be governed by the same
underlying crash mechanism. In reality, different sites have unique characteristics affecting
crash risk, such as road design, traffic volume, and driver demographics.
Impact: Failure to incorporate heterogeneity may result in biased or inconsistent parameter
estimates.
Limitations Can Be Addressed:
 Use of Negative Binomial (NB) Regression: To address overdispersion, analysts often
replace the Poisson model with a Negative Binomial (NB) model, which introduces a dispersion
parameter allowing the variance to exceed the mean. This model is more flexible and better
suited for real-world crash data.
 Zero-Inflated and Hurdle Models: To handle excess zeros, Zero-Inflated Poisson (ZIP)
and Zero-Inflated Negative Binomial (ZINB) models are used. These models combine two
processes: one predicting the probability of zero crashes and another modeling positive crash
counts. Hurdle models are also applied when different processes govern zero and non-zero
crash occurrences.
 Random Effects and Hierarchical Models: To account for spatial and temporal
correlations and unobserved heterogeneity, random-effects models or hierarchical Bayesian
models are used. These allow for varying effects across locations or time, improving accuracy
and generalizability.
While the Poisson distribution provides a simple starting point for crash data analysis,
it has significant limitations, especially when dealing with overdispersed data, excess zeros,
and unobserved variability. Advanced models like the Negative Binomial, zero-inflated models,
and random-effects frameworks offer more accurate and reliable tools for modern traffic
safety research and decision-making.
4. Describe the role of the Chi-squared test in traffic engineering. How can it be used to determine the
statistical significance of changes in crash rates before and after implementing safety measures?
The Chi-squared (χ²) test is a widely used statistical tool in traffic engineering for testing the
significance of differences between observed and expected frequencies. In the context of road safety,
it plays an important role in evaluating the effectiveness of traffic safety interventions, such as new
signage, signal timing changes, speed limit reductions, or road redesigns. Specifically, it helps
determine whether the change in crash rates observed after implementing a countermeasure is
statistically significant or could have occurred due to random variation.
Role of the Chi-squared Test in Traffic Engineering: In traffic engineering, the Chi-squared test is used
to compare categorical data—that is, data that fall into different categories such as the number of
crashes occurring before and after a safety intervention, or the number of crashes by type, location,
or severity level. The purpose is to assess whether observed differences in crash counts are significant
or merely due to chance.
The test is commonly used in:
 Before-and-after safety studies
 Crash type distribution comparisons
 Intersection or segment-based evaluations
 Severity analysis of crashes (e.g., fatal vs. injury crashes)
Using the Chi-squared Test to Evaluate Crash Rate Changes: To assess the effectiveness of a safety
measure, traffic engineers often conduct a before-and-after study. Suppose a safety intervention was
implemented at an intersection, and crash counts were collected for a similar time period before and
after the change.
Let’s say:
O₁ = Observed number of crashes before the intervention
O₂ = Observed number of crashes after the intervention
Using traffic volume data or exposure levels, expected crash counts can be estimated under the
assumption that the safety measure had no effect. These expected counts are compared with the
actual (observed) values using the Chi squared formula:
> χ² = Σ \[(Oᵢ - Eᵢ)² / Eᵢ]
Where:
Oᵢ = observed crash frequency in category i
Eᵢ = expected crash frequency in category i
The sum is taken over all categories (e.g., before and after periods)
This calculated Chi-squared value is then compared to a critical value from the Chi-squared
distribution table (based on degrees of freedom and a selected significance level, typically 0.05). If
the χ² value exceeds the critical value, the difference is considered statistically significant, indicating
the safety measure had a real effect.
Benefits of Using the Chi-squared Test
 Simplicity and Clarity: The Chi-squared test is straightforward and easy to compute,
making it suitable for practical engineering evaluations.
 No Assumption of Normality: It is a non-parametric test and doesn’t require data to follow
a normal distribution.
 Objective Evaluation: It provides an evidence-based way to assess whether safety
improvements are effective, supporting data-driven decision-making.
Limitations and Considerations:
 The test requires a sufficient sample size; small numbers can produce unreliable results.
It only tests association, not causation.
 It assumes that the expected frequencies are based on a valid and consistent baseline.

The Chi-squared test serves as a valuable analytical tool in traffic engineering for evaluating
the statistical significance of changes in crash rates following the implementation of safety
countermeasures. By comparing observed and expected crash frequencies, it helps determine
whether safety interventions have had a meaningful impact, thereby supporting the continuous
improvement of roadway safety strategies.
5. Outline the process of conducting a before-and-after study to evaluate the impact of a traffic
intervention. What statistical tools would you use to confirm the effectiveness of the intervention?
A before-and-after study is a widely used method in traffic engineering to evaluate the
effectiveness of safety interventions, such as installing traffic signals, redesigning intersections,
modifying speed limits, or adding pedestrian crossings. The goal is to compare the frequency and
severity of traffic crashes before and after the implementation of a measure, thereby assessing its real-
world impact on road safety.
Process of Conducting a Before-and-After Study:
1. Define Objectives and Select Study Sites: The first step involves clearly defining the purpose of the
study— typically, to assess whether a particular intervention led to a statistically significant reduction
in crashes. Suitable sites where the intervention has been implemented are identified, ensuring
consistency in traffic conditions and geometric design.
2. Determine Study Periods: Select appropriate before and after periods for data comparison. These
periods should be of similar duration (commonly 1 to 3 years each) and free from anomalies such as
construction work or external disruptions (e.g., natural disasters or pandemics).
3. Collect and Prepare Data: Gather crash data (frequency, types, severity, location) for both periods.
In addition, collect exposure data, such as traffic volumes (AADT – Average Annual Daily Traffic), to
control for changes in traffic flow, which might independently affect crash rates.
4. Account for External Influences: To isolate the effect of the intervention, control for confounding
factors such as seasonal variations, general crash trends in the region, or economic and environmental
changes. Using comparison sites —similar sites without interventions—can help control for broader
trends.
5. Calculate Crash Frequencies and Rates: Compute the number of crashes and the crash rate (e.g.,
crashes per million vehicle miles traveled) for both the before and after periods. These rates provide a
normalized measure of safety performance.
6. Analyze the Data: Apply appropriate statistical methods to determine whether observed differences
are significant and not due to random variation.

Statistical Tools to Confirm Effectiveness:

 Chi-Squared (χ²) Test: Used to compare observed crash frequencies before and after an intervention,
assessing whether changes are statistically significant. It is most effective when working with
categorical crash data.
 Paired t-Test or Z-Test: When crash rates are normally distributed, a paired t-test (or z-test for large
samples) can be used to evaluate the difference in means between the before and after periods.
 Empirical Bayes (EB) Method: A more advanced technique, the EB method adjusts for regression-
to-the-mean bias and incorporates data from similar sites (reference groups). It compares the expected
crash frequency (had no intervention occurred) with the observed post-intervention crashes to
estimate the actual safety effect.
 Percent Change Method: A simple approach that calculates the percentage change in crash
frequency or severity between the two periods. While easy to interpret, it does not account for
variability or statistical significance on its own.
A before-and-after study is a structured approach to evaluate the safety impacts of traffic
interventions. When combined with robust statistical tools like the Chi-squared test, Empirical Bayes
method, and paired t-tests, it enables traffic engineers and planners to make evidence-based
decisions. These evaluations are essential for prioritizing investments, refining road safety strategies,
and ultimately reducing crash risks on road networks.
6. List and briefly describe three traffic management measures aimed at accident prevention. How can
their effectiveness be quantified statistically?
Effective traffic management measures are critical components of road safety strategies,
aiming to reduce the frequency and severity of accidents. These measures address driver behaviour,
road conditions, and traffic flow through engineering, enforcement, and education. Below are three
commonly implemented traffic management measures focused on accident prevention, along with an
explanation of how their effectiveness can be quantified statistically?
1. Speed Management (Speed Limits and Calming Measures): Speed is one of the most
significant factors contributing to crash severity. Speed management includes the
enforcement of speed limits, installation of speed cameras, and implementation of traffic
calming devices such as speed humps, raised pedestrian crossings, and chicanes. These
interventions aim to reduce vehicle speeds, especially in high-risk areas like school zones and
residential neighbourhoods.
Statistical Evaluation: To evaluate effectiveness, a before-and-after study can be conducted by
comparing crash frequencies or severity before and after the measure’s implementation.
Statistical tools such as the paired t-test (for comparing mean crash rates) or Chi-squared test
(for categorical crash data) are used to determine if observed reductions are significant. Speed
data can also be collected to assess changes in average and 85th percentile speeds.
2. Traffic Signal Installation or Optimization: Installing traffic signals at previously unsignalized
intersections or optimizing existing signal timings can enhance vehicle coordination, reduce
conflict points, and improve safety for both drivers and pedestrians. Measures may include
adding dedicated turning phases, pedestrian signals, or adaptive signal control systems.
Statistical Evaluation: Effectiveness is quantified by comparing crash data from similar time
frames before and after the signal changes. For more robust results, the Empirical Bayes (EB)
method is commonly used. This method accounts for regression-to-the-mean bias and adjusts
expected crash frequencies using data from similar untreated sites. Statistical significance can
then be assessed using confidence intervals or hypothesis testing techniques.
3. Access Management: Access management involves regulating the location, design, and
number of driveways and intersections on roadways. Measures include consolidating
driveways, adding medians, and limiting left-turn movements. These strategies reduce the
number of conflict points and smooth traffic flow, especially on arterial roads.
Statistical Evaluation: Crash modification factors (CMFs) and Crash Reduction Factors (CRFs)
are often used to quantify the expected change in crash frequency due to access management
interventions. A Poisson or Negative Binomial regression model can be applied to develop
Safety Performance Functions (SPFs) that predict crash frequency based on traffic volume and
roadway characteristics. Changes in crash rates are then compared with predicted values to
evaluate effectiveness.
Traffic management measures such as speed control, traffic signal optimization, and
access management play a pivotal role in reducing road traffic crashes. Their effectiveness
must be evaluated through statistical methods to ensure data-driven decision-making. Tools
like before-and-after studies, Empirical Bayes analysis, regression modeling, and Chi-squared
tests provide the quantitative basis to assess safety outcomes. By applying these methods,
transportation agencies can prioritize the most effective interventions and improve overall
road safety performance.
7. Given a dataset with traffic volume and accident frequency, explain how you would develop a
regression model to identify the relationship between the two variables.
Developing a regression model to analyze the relationship between traffic volume and
accident frequency is a key technique in traffic safety studies. It helps transportation planners and
engineers understand how changes in traffic volume influence crash occurrences and is foundational
in the creation of Safety Performance Functions (SPFs). The following are the step-by-step outline of
how to develop such a model using a dataset containing traffic volume (typically measured as Average
Annual Daily Traffic, AADT) and accident frequency (number of crashes over a specific period).
1. Data Collection and Preparation: The first step involves collecting and organizing the dataset, which
includes:
 Traffic volume (independent variable, X)
 Accident frequency (dependent variable, Y)

Other variables such as roadway length, speed limits, number of lanes, and intersection type may be
included later to improve model accuracy, but the basic model focuses on the volume-accident
relationship.

Data cleaning is necessary to remove outliers, missing values, or inconsistent entries. Logarithmic
transformations may also be applied if the data is highly skewed.

2. Exploratory Data Analysis (EDA): Before modelling, it is essential to:

 Plot a scatter diagram of traffic volume vs. accident frequency to visually inspect the
relationship.
 Calculate summary statistics (mean, variance, etc.) for both variables.
 Check for linear or non-linear patterns, and assess whether variance increases with volume,
indicating overdispersion.
This step helps in deciding whether to use linear regression or a count-based model such as
Poisson or Negative Binomial regression.
3. Model Selection:
A. Linear Regression (Basic Case): If a rough linear trend is observed and assumptions of
normality and constant variance hold, a simple linear regression model can be applied:
> Y = β₀ + β₁X + ε
Where:
Y = accident frequency
X = traffic volume (AADT)
β₀, β₁ = regression coefficients
ε = error term B.
B. Poisson Regression: Since crash data are non-negative integers (count data), Poisson
regression is more appropriate in many cases:
> E(Y|X) = exp(β₀ + β₁X)
However, if the variance exceeds the mean (common in crash data), Negative Binomial
regression is preferred due to its ability to handle overdispersion.
4. Model Fitting and Interpretation: Using statistical software (e.g., R, Python, SPSS, or Excel), fit
the chosen model to the data. Key outputs include:
 Regression coefficients (β), showing the relationship strength
 p-values to test statistical significance
 Goodness-of-fit measures (e.g., R² for linear, Log-likelihood or AIC for
Poisson/Negative Binomial)
A positive and significant β₁ would suggest that crash frequency increases with traffic volume.
5. Validation and Residual Analysis: Check model assumptions and fit by:
 Analyzing residual plots
 Performing cross-validation or using a test dataset
 Comparing predicted vs. actual crash counts
6. Application: Once validated, the model can:
 Predict accident frequency at locations with known traffic volume
 Help in prioritizing sites for safety improvements
  Serve as the basis for Crash Modification Factors (CMFs) or SPFs in roadway safety analysis.

By systematically developing a regression model using traffic volume and crash data, transportation
professionals can quantify the relationship between exposure and crash risk. This analytical approach
supports evidence-based decision making in traffic safety planning and resource allocation.

8. How can statistical methods be used to evaluate the impact of road design changes on crash reduction?
Statistical methods play a critical role in evaluating the effectiveness of road design changes—
such as the installation of roundabouts, speed bumps, median barriers, or traffic-calming measures—
in reducing traffic crashes. By providing a rigorous framework for analysis, statistics help determine
whether observed changes in crash frequency or severity are the result of the design intervention or
random variation.
1. Before-and-After Studies: The most common approach for evaluating road design changes is
the before-and-after study. This involves comparing crash data for a site before a design
change (e.g., replacing an intersection with a roundabout) with data from the same site after
the change has been implemented. Key metrics include crash frequency, severity, and crash
rates adjusted for traffic volume.
Simple Before-and-After Study: A direct comparison of crash counts before and after the
intervention. While easy to implement, this method does not account for external influences,
such as regional crash trends or traffic volume changes.
Empirical Bayes (EB) Method: To improve accuracy, the Empirical Bayes method adjusts for
regression-to-the-mean and accounts for background trends using reference data from similar
untreated sites. This method compares the expected number of crashes (had no change
occurred) with the actual number of crashes observed after the intervention.
> Crash Reduction = Expected Crashes (EB estimate) – Observed Post-treatment Crashes
2. Statistical Tests for Significance: To determine whether the change in crash frequency is
statistically significant, several tests may be used:
Chi-squared Test: Used for categorical crash data (e.g., crash counts by type or severity),
testing if the distribution has significantly changed.
Paired t-test or z-test: Compares average crash rates or crash frequencies before and after the
intervention.
Confidence Intervals: Helps estimate the uncertainty around the mean reduction in crashes;
if the interval does not include zero, the reduction is statistically significant.
3. Regression Models: Poisson and Negative Binomial regression models are often used when
crash data are treated as count data. These models can incorporate variables like traffic
volume, road type, and environmental factors. For example, a Negative Binomial regression
might model crash frequency as a function of whether a roundabout is present, controlling for
AADT and other confounders. If the coefficient for the roundabout variable is negative and
significant, it suggests a crash-reducing effect.
4. Crash Modification Factors (CMFs): CMFs quantify the expected change in crash frequency
after a road design change. A CMF less than 1.0 indicates a reduction in crashes. For example,
a CMF of 0.60 for roundabouts suggests a 40% expected reduction in crashes. CMFs are often
derived using EB methods or regression analysis from observational data.
Statistical methods provide an objective, data-driven approach to evaluating road
design interventions. Whether through before-and-after studies, Empirical Bayes techniques,
or regression models, these methods enable transportation professionals to measure the
safety impact of changes like roundabouts or speed bumps. By doing so, they support
informed decision-making, resource allocation, and continuous improvement in road safety
strategies.
9. What is the importance of confidence intervals in interpreting statistical results in traffic safety studies?
Provide an example relevant to accident rate comparisons.
Confidence intervals (CIs) are a fundamental concept in statistics that play a vital role in
interpreting results from traffic safety studies. They provide a range of values within which the true
value of a population parameter (such as a mean accident rate or crash reduction) is expected to lie,
with a specified level of confidence—commonly 95%. In road safety research, confidence intervals help
assess the reliability and statistical significance of estimated effects, such as changes in crash rates
before and after interventions.
Why Confidence Intervals Matter in Traffic Safety:
1. Quantifying Uncertainty: Traffic data are often subject to natural variation due to differences in
traffic volume, weather, driver behavior, and random chance. A single number like an average crash
rate may not capture this variability. Confidence intervals account for this uncertainty and provide a
range of plausible values for decision making.
2. Assessing Statistical Significance: Confidence intervals help determine if an observed change is
statistically significant. If a 95% CI for the difference in accident rates does not include zero, it indicates
a significant change with 95% confidence. This is a more informative approach than relying solely on
p-values.
3. Supporting Comparative Analysis: When comparing accident rates between two sites or two time
periods (e.g., before and after a safety improvement), overlapping confidence intervals can suggest no
significant difference, while non-overlapping intervals suggest a likely effect.
4. Decision-Making and Policy Justification: Transportation agencies often need to justify investments
in road safety. Confidence intervals provide evidence-based support, showing not only the estimated
benefit but also the degree of confidence in that estimate.
Example:
Accident Rate Comparison: Imagine a study comparing accident rates at an intersection before and
after the installation of a roundabout. The average accident rate before the intervention was 3.5
crashes per million entering vehicles (MEV), and after the roundabout, it dropped to 2.1 crashes per
MEV. Using statistical analysis, we calculate:
Before CI: 3.5 ± 0.6 → (2.9, 4.1)
After CI: 2.1 ± 0.4 → (1.7, 2.5)
The confidence intervals do not overlap, indicating with 95% confidence that the roundabout
installation significantly reduced crash rates. Had the intervals overlapped (e.g., (2.9, 4.1) and (2.2,
3.0)), we would not be confident in claiming a significant reduction, as the difference could be due to
random chance.
This result helps practitioners conclude that the roundabout was effective in reducing accidents, and
similar treatments may be considered at other high-risk locations.
Confidence intervals are crucial in traffic safety studies because they provide a more complete
picture than point estimates alone. By quantifying uncertainty, guiding statistical significance, and
supporting transparent decision-making, CIs strengthen the reliability of conclusions drawn from
accident rate comparisons. Their application ensures that safety recommendations and investments
are based on sound, statistically validated evidence.
10. How can combining regression analysis with Poisson or negative binomial distributions improve the
accuracy of traffic crash prediction models? Illustrate your answer with a practical scenario.
Combining regression analysis with Poisson or Negative Binomial (NB) distributions
significantly enhances the accuracy and reliability of traffic crash prediction models, especially when
dealing with count data such as the number of accidents at road segments or intersections. These
models allow traffic safety analysts to identify patterns, estimate crash frequencies, and evaluate the
impact of roadway features or traffic volume on crash risk.
1. Why Poisson and Negative Binomial Regression Are Used: Traffic crash data are typically
discrete, non-negative, and often skewed, which makes standard linear regression
unsuitable due to its assumptions of normally distributed errors and constant variance
(homoscedasticity). Instead:
 Poisson regression is suited for modelling count data under the assumption that the mean
equals the variance.
  Negative Binomial regression is an extension of the Poisson model that accounts for
overdispersion (when the variance exceeds the mean), a common feature in crash datasets
due to unobserved heterogeneity such as weather, driver behaviour, or road conditions.
2. How These Models Improve Accuracy: By incorporating Poisson or NB distributions into
regression analysis, the model:
 More accurately represents the distributional nature of crash data.
 Allows inclusion of explanatory variables such as traffic volume, lane width, speed limit,
or number of intersections.
 Can handle a wide range of data conditions, especially overdispersed or zero-inflated
datasets.
 Produces more reliable predicted crash frequencies, essential for identifying high-risk
locations and prioritizing interventions.
3. Practical Scenario: Predicting Crashes at Urban Intersections: Imagine a city traffic
department wants to predict the number of crashes expected at 100 urban intersections to
prioritize safety upgrades. The dataset includes:
 Crash frequency over the past 3 years (dependent variable)
 Average daily traffic (ADT)
 Number of lanes
 Signalization (yes/no)
 Presence of pedestrian crossings
Using Poisson regression, the model may look like:
> E(Y) = exp(β₀ + β₁ADT + β₂Lanes + β₃Signal + β₄PedCross)
If residual analysis shows that the variance of the crash counts is much higher than the
mean (common in real-world traffic data), the model is likely overdispersed, making the
Poisson assumptions invalid. Switching to a Negative Binomial regression model allows the
analyst to incorporate a dispersion parameter (α), which better fits the observed variance,
leading to more accurate crash predictions.
4. Application of the Model: Once fitted, the model can:
 Predict crash frequencies at new or modified intersections.
 Identify locations where the observed crashes are significantly higher than predicted,
suggesting a need for safety interventions.
 Evaluate potential impact of design changes (e.g., reducing lanes or adding pedestrian
crossings).
For example, the model may predict that intersections with pedestrian crossings experience
20% fewer crashes, all else being equal. This insight can be used to justify pedestrian
infrastructure improvements citywide.
Combining regression analysis with Poisson or Negative Binomial distributions allows
for more accurate and interpretable traffic crash prediction models by aligning the statistical
method with the nature of crash data. This approach supports data-driven decision-making,
better resource allocation, and more effective road safety planning in both urban and rural
environments.