ABCDEFGH
1
PORTLAND STATE UNIVERSITY AT TRB 2024
2
*Due to the evolving status of speaker attendance and TRB programming, please check your TRB schedule for the most current information in the event of changes.*
3
DateTime (Eastern)#TypeSessionPresentation TitlePresentersAbstract
4
1/7/20249:00 AM-
12:00 PM		1006WBicycle and Pedestrian Data Fusion: Learning from Each OtherExploring Data Fusion Techniques to Estimate Network-Wide Bicycle VolumesSirisha Kothuri, Portland State UniversitySirisha Kothuri, Portland State University, will share their work on recently completed project “Exploring Data Fusion Techniques to Estimate Network-Wide Bicycle Volumes” which drew data from Location Based Services (LBS), app, ad bike share, as well as more traditional infrastructure and socio-demographic datasets.
5
1/8/202410:15 AM- 12:00 PM2109PTransit Management and Performance PapersShifts in Public Perception about Returning to Transit: A Qualitative Analysis of Longitudinal Survey Data Amidst the COVID-19 PandemicSameer Aryal, University of Tennessee, Knoxville; Christopher Cherry, University of Tennessee, Knoxville; Mojdeh Azad, The University of Tennessee Knoxville; Candace Brakewood, University of Tennessee; John MacArthur, Portland State UniversityThis paper presents a longitudinal survey analysis that investigates the dynamic perceptions of individuals regarding the resumption of transit services amidst the challenging COVID-19 pandemic. We analyzed responses to an open-ended question regarding users' thoughts on returning to transit during COVID-19. The data was collected through a smartphone app over six survey waves nationwide. Eight distinct themes emerged from the analysis, initially coded using text analysis software, and subsequently, they were manually verified and updated. These themes encompassed critical factors such as reopening physical locations, cleanliness and sanitization, mask requirements, vaccination requirements, social distancing, availability of more services, number of active COVID-19 cases, and feeling safe to ride. Using a panel regression model, we further investigated the relationship between these themes and the socio-demographic data of the users, providing an understanding of the factors influencing individuals' attitudes towards returning to transit. By exploring these multifaceted dimensions and incorporating statistical analysis, this study illuminates the evolving attitudes, concerns, and priorities of the public over time. It provides evidence of the socio-demographic factors that shape these perceptions. The insights from this analysis will benefit policymakers and transit authorities with the knowledge to develop targeted strategies and interventions that effectively address public sentiment and facilitate the safe and efficient return to transit services during infectious disease outbreaks.
6
1/8/20241:30 PM-
3:15 PM		2149LThe Unhoused in Transportation: A Conversation on the Issue in Transportation Rights-of-Way and AssetsPerspectives in the use ROW outlined in the the study "Homelessness: A Guide for Public Transportation (TCRP J-11/Task 40)John MacArthur, Portland State UniversityThe homelessness crisis has reached epidemic levels in countless regions across the United States. There are multiple strategies employed by policymakers to bring all the stakeholders together to understand, strategize, and act upon this crisis. But there is not one size fits all strategy and the outcomes vary. This session will bring one set of perspectives that will focus on the research work and engagement with the unhoused relative to transportation rights of way. This will also be an informal forum to bring to light related strategies that could be explored in the future.
7
1/8/20243:45 PM-
5:30 PM		2215PNovel Paradigms for the Use of Statistical and Econometric Methods in Transportation ResearchAddressing Uncertainties in Transportation Econometrics with Missing Data: A Comparative Study between Single and Multiple ImputationMd Istiak Jahan, University of Central Florida; Tanmoy Bhowmik, Portland State University; Lauren Hoover, University of Central Florida; Naveen Eluru, University of Central FloridaWhile several approaches for imputation exist for data imputation, these approaches are not commonly applied in transportation. The current paper is geared towards assisting transportation researchers and practitioners in developing models using datasets with missing data. The study begins with a data simulation exercise to evaluate different solutions implemented for missing data including :(a) single imputation (SI) approach and (b) multiple imputation (MI) approach (with varying number of realizations) and complete case data (CCD) approach (dropping records with missing values). The comparison is carried out by adopting the appropriate inference process for MI approach with multiple realizations. From the simulation exercise, we recognize the MI approach consistently performs better than SI approach. Among various realizations, MI approach with five realizations is selected based on our results. The MI approach with 5 realizations is compared with the complete case data approach under different conditions. In the presence of a small share of missing data, it might be beneficial to simply develop a CCD model by dropping observations with missing values as opposed to developing imputation models. However, when the share of missing data warrants variable exclusion, it is important and even necessary that multiple imputation approach be employed for model development. In the second part of the paper, based on our findings, we implemented the MI approach for real empirical datasets with missing values for four discrete outcome variables.
8
1/8/20243:45 PM-
5:30 PM		2223PAdvanced Analytical Techniques Used to Address the Unique Transport Issues of Developing CountriesExamining Driver Injury Severity in Motor Vehicle Crashes: A Copula-Based Approach Considering Temporal Heterogeneity in a Developing Country ContextShahrior Pervaz, University of Central Florida; Tanmoy Bhowmik, Portland State University; Naveen Eluru, University of Central FloridaUsing data from a developing country, the current study develops a joint modeling framework to study crash type and driver injury severity as two dimensions of the severity process. A copula-based multinomial logit model (for crash type) and generalized ordered logit model (for driver severity) is estimated. The data for our analysis is drawn from Bangladesh for the years of 2000 to 2015. Given the presence of multiple years of data, we develop a novel spline variable generation approach that facilitates easy testing of variation in parameters across time in crash type and severity components. A comprehensive set of independent variables including driver and vehicle characteristics, roadway attributes, environmental and weather information, and temporal factors are considered for the analysis. The model results identify several important variables affecting crash type and severity while also isolating temporal instability for a subset of parameters. The superior model performance was further highlighted by testing its performance using a holdout sample.
9
1/9/20241:30 PM-
3:15 PM		3161PPedestrian Accessibility, Walkability, and DesignUnder the Influence of Parents: A Longitudinal Study of Children’s WalkingKyuri Kim, Portland State University; Jennifer Dill, Portland State UniversityMany researchers have studied children's active travel; however, they have mostly been cross-sectional studies dealing with commuting to school and parental perceptions and attitudes. To find ways to promote children's active travel, this longitudinal study uses panel data (two time periods) to examine how parents' actual walking and safety perception correlated with children’s walking. Using data from 240 children aged 4-16 and their parents in Portland, Oregon, we estimated a cross-lagged panel model (CLPM) to analyze the continuous relationships between variables. We found that parents with more positive attitudes toward their walking walked more, and their children also walked more in both periods. In addition, parents' safety perceptions about their children's walking environment in terms of strangers and traffic significantly affected children's walking time. Children’s safety perception (only in the second period) and time-related perceived behavior control (in both periods) affected their attitude toward walking but not walking time. Moreover, previous children’s walking behavior can alleviate the subsequent negative perception of the time barrier of walking. This study is meaningful in finding the relationships between the attitudes and behaviors of parents’ and children's walking based on repeated observations. Given our results, neighborhoods that are good for parents to walk in affect children’s walking. Parental walking needs to precede for encouraging continuous children’s walking.
10
1/9/20241:30 PM-
3:15 PM		3171PDwight David Eisenhower Transportation Fellowship Program Poster (Session 3)Mitigating Liquefaction Risk for Transportation Infrastructure using Microbially Induced DesaturationKayla Sorenson, Portland State University
11
1/9/20241:30 PM-
3:15 PM		3176PFreight Demand and Choice Modeling Using Advanced Data SetsExploratory Analysis of Factors Affecting Home Delivery ReturnsMichael Bronson, Portland State University; Miguel Figliozzi, Portland State University; Ali Riahi Samani, University of Memphis; Sabyasachee Mishra, University of MemphisE-commerce and house deliveries have experienced a rapid growth in the last two decades. The return of online shopping products is an undesirable side effect of online shopping that has not been properly studied in the transportation literature. Utilizing binary logit models, this research answers two novel research questions focusing on the online shopping channel: (i) What household characteristics are associated with a higher or lower propensity to return online purchases? and (ii) What type of products contribute to positive return delivery rates? To answer these questions models are developed using data collected from a household online survey of e-commerce and shopping activities in the state of Tennessee. The results clearly indicate that the number of online purchases, the percentage of fashion and beauty products purchased online, the presence of a delivery subscription, higher household income, and lower age are the key factors that increase the probability of having returns for products purchased online. Implications of the findings for transportation planning and online shopping are analyzed and discussed.
12
1/9/20243:45 PM- 5:30 PM3214PBicycle Safety, Planning, and Design: Research to Support Better Bicycling ConditionsConsumer Purchase Response to E-bike Incentives: Results from a Nationwide Stated Preference StudyLuke Jones, Valdosta State University; Cameron Bennett, Portland State University; John MacArthur, Portland State University; Christopher Cherry, University of Tennessee, KnoxvilleElectric bike (e-bike) incentives are gaining popularity as a means to increase ownership of ebikes, often with the intention of increasing mobility while reducing greenhouse gas emissions from car use. To date, several incentive programs have been developed with little empirical analysis of effectiveness and optimization of incentives to maximize e-bike ownership. This study is among the first to experimentally analyze prospective purchase behavior of e-bike consumers to assess price sensitivity, behavioral response to incentives, and consumer value of e-bike attributes. To this end, this paper describes the result of a nationwide survey (2,241 respondents across 16 cities). We find that e-bike incentives are powerful levers to shift behavior and find heterogeneity across different income groups. Different incentive mechanisms have different behavioral utility. Purchase discounts are the most influential at shifting behavior, followed by rebates, then tax refunds. Purchase discounts are 30% more effective than tax refunds. For purchase discounts, under an optimum discount scheme, the inframarginal cost of an added e-bike purchase is $4252. Policy makers should evaluate this investment across benefits from added e-bike ownership that includes greenhouse gas reductions, health benefits, safety, and congestion reductions.
13
1/9/20246:00 PM-
7:30 PM		3231PAnalytical Methods of Safety PerformanceAn Integrated Multi-Resolution Framework for Jointly Estimating Crash Type and Crash SeverityShahrior Pervaz, University of Central Florida; Tanmoy Bhowmik, Portland State University; Naveen Eluru, University of Central FloridaThe current research effort contributes to safety literature by developing an integrated framework that allows for the influence of independent variables from crash type and severity components at the disaggregate level within the aggregate level propensity to estimate crash frequency by crash type and severity. The proposed framework can also incorporate unobserved heterogeneity in the model system. The empirical analysis is based on 2019 crash data drawn from the city of Orlando, Florida. The disaggregate level analysis uses 15,518 crash records of three crash types: rear end, angular and sideswipe. Each crash record contains crash specific factors, driver and vehicle factors, roadway, temporal, road environmental and weather information. For aggregate level model analysis, the study aggregates the crash records by crash type over 300 traffic analysis zones. The empirical analysis is further augmented by employing several goodness of fit and predictive measures. A validation exercise is also conducted using a holdout sample to highlight the superiority of the proposed integrated model relative to the non-integrated model system. The findings of the study indicate that the proposed framework is advantageous for capturing the variable effects simultaneously across the aggregate and disaggregate levels.
14
1/10/20248:00 AM-
9:45 AM		4003LSafety Performance and Analysis ResearchA Systematic Unified Approach for Addressing Temporal Instability in Road Safety AnalysisKazi Redwan Shabab, University of Central Florida; Tanmoy Bhowmik, Portland State University; Naveen Eluru, University of Central Florida; Mohamed Zaki, University of Western OntarioMultivariate models are widely employed for crash frequency analysis in traffic safety literature. In the context of analyzing data for multiple instances (such as years), it becomes essential to evaluate the stability of parameters over time. The current research proposes a novel approach, labelled the mixed spline indicator pooled model, that offers significant enhancement of current approaches to capture temporal instability. The proposed entails carefully creating additional independent variables that allow us to measure parameter slope changes over time and can be easily integrated into existing methodological frameworks. The current research effort compares four multivariate model systems: year specific negative binomial model, year indicator pooled model, spline indicator pooled model, and mixed spline indicator pooled model. The model performance is compared using log-likelihood and Bayesian Information Criterion. The empirical analysis is conducted using the Traffic Analysis Zone (TAZ) level crash severity records from Central Florida for the years from 2011 to 2019. The comparison results indicate that the proposed mixed spline indicator pooled model outperforms the other models providing superior data fit with significantly fewer parameters. The proposed mixed spline model can allow a piece-wise linear functional form for the parameter and is suitable to forecast crashes for future years as illustrated in our predictive performance analysis.
15
1/10/20248:00 AM-
9:45 AM		4027POmnibus Session on Bicycle Modeling and Shared Micromobility ResearchShared micromobility fees: current patterns, impacts, and best practicesCalvin Thigpen, Lime; Kevin Fang, Sonoma State University; John MacArthur, Portland State UniversityAs cities look to foster sustainable transportation systems, shared bicycle and scooter systems offer a promising option. However, shared micromobility programs have encountered financial turbulence, with companies bowing out of cities and many still seeking to reach financial sustainability. Absent public subsidy, one promising option for cities seeking long-term stability in their shared micromobility programs is to reconsider the program fees they charge private operators. This paper seeks to assess the current state of program fees, identify their impacts on business and relate them to the broader transportation context, and suggest best practices. From a sample of 74 cities from 3 continents, we find a broad array of fees, including per-vehicle, per-trip, annual, and one-time fees of varying amounts and in different combinations. Two thirds of jurisdiction employ two or more different fee types, while one in seven charge no fee. The resulting fee revenues consequently differ considerably, ranging from cities that charge no fees to cities receiving more than US$1M annually and fees constituting 0% to 34% of companies’ fare revenues. We find that on a per-mile basis, shared scooters and bicycles are charged at fourteen times the rate in fees than drivers pay in the gas tax. Considering goals related to program longevity, low costs, and minimization of negative externalities, we suggest fees should be set to no more than the city’s cost of program administration, fee revenues should be directed to program expenses, and cities should consider setting fee structures to align companies’ operations with city goals.
16
1/10/202410:15 AM-
12:00 PM		4053LNavigating the Future: Travel Behavior in the Age of New MobilitiesUnderstanding the Role of Transportation Network Companies in Addressing Transportation Demand: A Chicago Case StudyDewan Ashraful Parvez, University of Central Florida; Sudipta Dey Tirtha, University of Central Florida; Tanmoy Bhowmik, Portland State University; Srinivas Peeta, Georgia Institute of Technology; Naveen Eluru, University of Central FloridaThis study builds a systematic framework to analyze variations in Transportation Networking Companies (TNCs) usage patterns across the urban region. As opposed to comparing the TNC usage patterns based on TNC demand, we focus on TNC usage relative to overall transportation demand. Comparing TNC demand to overall transportation demand, an ordinal metric is generated at the census tract resolution using 2019 data from Chicago. Based on the comparison, census tracts were categorized into five categories: High TNC surplus, TNC surplus, balanced, TNC deficit and High TNC deficit. A univariate descriptive analysis is conducted to understand the characteristics of census tracts in the five categories. The results highlight the presence of significant differences among socio-demographics and transportation infrastructure across the categories. To further analyze the differences in TNC usage patterns, a multivariate analysis is conducted by developing a Generalized Ordered Logit model for the ordinal metric. The model results further highlight the interactions of different independent variables influencing TNC usage patterns. It is observed that the composition of communities heavily underserved are likely to overrepresent low income households, non-Caucasian households and older adults (age ≥ 65). The model applicability is illustrated through elasticity analysis that provides a clear mechanism to illustrate the contribution of the different variables affecting TNC usage. The proposed framework can be applied to any urban region to identify spatial pockets and demographic segments underserved by TNCs. Transportation agencies can devise policies to address these imbalances through additional incentives for TNCs to operate in currently underserved census tracts.
17
1/10/202410:15 AM-
12:00 PM		4070PSafety Impact on Road Users Including Pedestrians, Bicyclists, and OthersAnalysis of Lag Time of Pedestrian Fatalities: A Copula ApproachNafis Anwari, University of Central Florida; Tanmoy Bhowmik, Portland State University; Mohamed Abdel-Aty, University of Central Florida; Naveen Eluru, University of Central Florida; Juneyoung Park, Hanyang University, AnsanThis study uses a copula-based joint modelling framework to investigate the survival outcome and lag time of pedestrian fatalities. The upper level model investigates whether or not the pedestrian died instantly (lag time = 0 for instant death), while the lower level model investigates lag time for pedestrians who did not die instantly. The joint model was run on a dataset of 33615 observations obtained from the Fatality Accident Reporting System (FARS) for the period of 2015-2019. The effect of roadway and traffic characteristics were investigated on lag time using six copula structures, among which Gaussian parameterized copula was found to have the best fit. Weather, Driver age groups, Drunk/ distracted/ drowsy drivers, Hit and Run, Involvement of Large Truck, VRU age group, Sex, Presence of Sidewalk, Presence of Intersection, Light Condition, Speeding, and Race. The factor found to be significant exclusively for the Binary Logit model includes Area type. Factors found to be significant exclusively for the Ordered Logit model include Presence of Crosswalk and Fire station nearby. Most of the analysis results in this study are in line with those obtained in previous crash severity and crash count studies, validating the use of lag hours as an alternative to crash count and crash severity analysis. Based on model results, the study recommended some countermeasures to improve pedestrian safety.
18
1/10/20243:45 PM-
5:30 PM		4092PCurrent Issues in Alternative Fuels and TechnologiesUnderstanding the Dynamic Change of Electric Vehicle Adoption in California: An analysis of Stated Preference DataMd Istiak Jahan, University of Central Florida; Tanmoy Bhowmik, Portland State University; Naveen Eluru, University of Central FloridaEV adoption is still in its infancy with a market penetration rate of about 8%. Thus, a major challenge toward understanding the behavioral preferences of EV adoption is the paucity of revealed preference data on households making EV vehicle purchases vis-à-vis traditional vehicle purchases. Further, current EV owners might not be representative of the general population and are potentially skewed towards high income households. A good solution to address the data paucity/representativeness problem is the adoption of stated preference survey-based data for examining how households evaluate new emerging vehicle alternatives. We conduct our analysis using the stated preference (SP) part of the California Vehicle Survey for two waves (2017 and 2019) to understand the factors influencing EV adoption decisions. In our model development, we recognize that vehicle purchase decisions might follow different decision processes and compare the traditional random utility maximization framework with the random regret minimization framework. To accommodate for the influence of temporal unobserved factors specific to each wave we employ a heteroscedastic model. The vehicle purchase decision model is developed considering an exhaustive set of independent variable groups including (a) vehicle attributes, (b) household socioeconomic characteristics, (c) charging infrastructure characteristics, and (d) interactions of the various variable groups. A trade-off analysis is conducted to highlight the impact of different factors on vehicle purchase decisions.