Hierarchical Ordered Probit Model Research Articles

Trust towards autonomous driving, worthwhile travel time, and new mobility business opportunities

ABSTRACT When driving autonomous cars, individuals have the opportunity to engage in non-driving activities, resulting in improved well-being. By incorporating trust into the technology acceptance modelling framework, this study extends the understanding of behavioural intention from the adoption of autonomous cars to the allocation of worthwhile travel time. Within this extended framework, the structural equation model is employed to identify significant paths to time allocation behaviour. Specifically, trust is identified as a crucial determinant of such behavioural intention. Then, we use the hierarchical ordered probit model to analyse the extent of such time allocation, accounting for scaling heterogeneity in self-reported responses. Our empirical evidence demonstrates the vital role of trust in encouraging the allocation of worthwhile and productive travel time. Based on these results, we present various new mobility business opportunities as well as some policy implications from the perspective of trust.

Effects of design consistency measures and roadside hazard types on run-off-road crash severity: Application of random parameters hierarchical ordered probit model

Run-off-road crashes are one of the most significant causes of road deaths worldwide. Given such significant safety concerns, a number of earlier studies examined the critical factors contributing towards run-off-road crash severity outcomes, mostly by using the information compiled in the official crash database. However, the official crash databases are less likely to have detailed information on driver behavior (errors/expectations) and roadway environment (roadway geometry and roadside attributes). This study aims to investigate the effects of design consistency measures on run-off-road crash severity mechanisms by applying a random parameters hierarchical ordered Probit model. This study contributes towards existing safety literature by demonstrating a complementary approach to capturing the effects of driver behavior and heterogeneity in roadway environment on run-off-road crash severity outcome through the composite measures of design consistency indices and cosmopolite measures of roadside hazard type variables. Specifically, 17 different functional forms of design consistency indices are developed to capture the behavioral factors from the road-geometric changes in developing run-off-road crash severity models. Further, in examining the effect of different types of the roadside environment on run-off-road crash severity outcomes, seven roadside hazard type variables are generated as a composite function of roadside object type and clear zone (lateral distance to roadside object). The empirical analysis of this study involves a two-step modelling approach - in the first step, the decision tree algorithm is applied to identify the higher-order interaction among independent variables, and in the second step, crash severity models are developed by employing several econometric approaches. The hybrid models are estimated by employing four econometric frameworks, which include Ordered Probit, Hierarchical Ordered Probit, Random Parameters Ordered Probit, and Random parameters Hierarchical Ordered Probit models. The run-off-road crash severity models are estimated by using crash data collected from the State of Queensland, Australia, for the years 2015 through 2019. Overall, this study reveals the importance of considering the interaction of drivers' behavior, road geometry, and roadside attributes along with other independent variables in developing run-off-road crash severity models.

A data fusion approach of police-hospital linked data to examine injury severity of motor vehicle crashes

Injury severity studies typically rely on police-reported crash data to examine risk factors associated with traffic injuries. The police crash database includes essential information on roadways, crashes and driver-vehicle characteristics but may not contain accurate and sufficient information on traffic injuries. Despite sizable efforts on injury severity modelling, very few studies have employed hospital records to classify injury severities accurately. As such, the inferences drawn from the police-recorded injury severity classifications may be questionable. This study investigates factors affecting road traffic injuries of motor vehicle crashes in two approaches (1) police-reported injury severity data and (2) a data fusion approach linking police and hospital records. Data from 2015 to 2019 were collected from the Abu Dhabi Traffic Police Department and linked with hospital records by the Department of Health, Abu Dhabi. A total of 6,333 casualty crashes were categorised into non-severe, severe, and fatal crashes following police-reported data and non-hospitalised, hospitalised and fatal crashes based on the police-hospital linked data. The state-of-the-art random thresholds random parameters hierarchical ordered Probit models were then employed to examine the differences in factors affecting crash-injury severities between police-reported and police-hospital linked data. While there are similarities between these two approaches, there are numerous notable differences in injury severity factors. For instance, head-on collisions are associated with high crash-injury severities in the model with police-hospital linked data, but they tend to show low injury severities in the model with police-reported data. In addition, the police-reported approach identifies that crashes occurred in remote areas and angle collisions are associated with low injury severities, which is not intuitive. These findings highlight that modelling the misclassified injury severity in police crash data may lead to wrong estimations and misleading inferences. Instead, the data fusion approach of police-hospital linked data provides critical and accurate insights into road traffic injuries and is a valuable approach for understanding traffic injuries.

Who is More Likely (Not) to Make Home-Based Work Trips During the COVID-19 Pandemic? The Case of Scotland.

In this study, we used survey data (n = 6,000) to investigate the work trip patterns of Scottish residents at various points of the COVID-19 pandemic. We focused specifically on the reported patterns of weekly work trips made during the government-enforced lockdown and subsequent phases of restriction easing. This was of particular importance given the widespread changes in work trips prompted by COVID-19, including a significant rise in telecommuting and a reduction in public transport commuting trips. The survey data showed that the vast majority of respondents (∼85%) made no work trips during lockdown, dropping to ∼77% following the easing of some work-related restrictions. Zero-inflated hierarchical ordered probit models were estimated to determine the sociodemographic and behavioral factors affecting the frequency of work trips made during three distinct periods. The model estimation results showed that the socioeconomic characteristics of respondents influenced work trips made throughout the pandemic. In particular, respondents in households whose main income earner was employed in a managerial/professional occupation were significantly more likely to make no work trips at all stages of the pandemic. Those with a health problem or disability were also significantly more likely to make no work trips throughout the pandemic. Other interesting findings concern respondents’ gender, as males were more likely to complete frequent work trips than females throughout the pandemic, and differences between densely populated areas and the rest of Scotland, as respondents from a large city (Edinburgh or Glasgow) were significantly more likely to make frequent work trips as restrictions were eased.

Analysis of factors affecting occupant injury severity in rear-end crashes by different struck vehicle groups: A random thresholds random parameters hierarchical ordered probit model

This study aims to explore the variability of risk factors affecting injury severity in rear-end crashes when different struck vehicle groups are involved. Two types of rear-end crash data, vehicle-strike-car data, vehicle-strike-truck data, are extracted from the Fatality Analysis Reporting System (FARS). Two likelihood ratio (LR) tests are firstly performed to validate the struck vehicle group variations, and then two separate random thresholds random parameters hierarchical ordered probit (RRHOP) models (Model 1 and Model 2) are established to capture unobserved heterogeneity. The results of LR test show significant differences in the effects of factors included in each model. Moreover, the model results suggest that SUVs, vans, and large trucks as striking vehicles are significant related to injury severity in both models with different effects. Factors such as speeding related, pickup, model year (struck vehicle), disabled damage, adverse weather, speed limit (≥60 mile/h), and young driver (struck vehicle) are found to be statistically significant in only model 1. These results provide a better understanding of differences in contributing factors of rear-end crashes, which help to propose effective countermeasures to mitigate its injury severity.

Детерминанты здоровья и гетерогенность ответов при самооценке здоровья в России: подход якорных виньеток

We show the reporting heterogeneity bias into all 7 domains of health using the anchoring vignettes approach with the hierarchical ordered probit model. The paper uses the data of the Russian sample of the first wave of the global survey SAGE provided by the World Health Organization Study. The heterogeneity bias correction has changed the significance level of education level, permanent income, rural/ urban living, and marital status.

Weekly variations and temporal instability of determinants influencing alcohol-impaired driving crashes: A random thresholds random parameters hierarchical ordered probit model

Alcohol consumption has been acknowledged as a critical determinant concerning the occurrence of vehicle crashes and their resulting injury severities. To investigate the weekly transferability and temporal stability of the contributors determining different injury severity levels in alcohol-impaired driving crashes, this paper employs two groups of random thresholds random parameters hierarchical ordered probit models. Three injury-severity categories are determined as outcome variables: no injury, minor injury and severe injury, while multiple factors are investigated as explanatory variables including driver characteristics, vehicle characteristics, roadway characteristics, environmental characteristics, crash characteristics and temporal characteristics. The weekly transferability and temporal stability of the models are examined through three groups of likelihood ratio tests. Marginal effects are also adopted to analyze the weekly transferability and temporal stability of the explanatory variables. Overall, the findings exhibit weekly variations and temporal instability while some indicators are also observed to be of relative weekly transferability including speeding, aggressive driving, proceeding, motorcycle, speed limit between 45 and 55 mph, curve, driveway, overturned, hit-fixed-object, vehicle age (5–9 years old). Besides, curve and passenger car indicators in weekday models present relative temporal stability. This paper can provide insights into preventing alcohol-impaired driving crashes and can potentially facilitate the development of the corresponding crash injury mitigation policies. More studies could be conducted integrating the advanced data-driven methods into the statistical models to simultaneously achieve inference and prediction.

A comparative analysis of factors influencing the injury severity of daytime and nighttime crashes on a mountainous expressway in China

This study focuses on analyzing the injury severity of crashes occurring at different times of the day and providing insights into crash prevention countermeasures. The data contained 1521 police-reported crashes that occurred on a mountainous expressway between 2012 and 2017, which were divided into two subsamples: the daytime crash sample and the nighttime crash sample. The statistical plausibility of developing separate models for the two subsamples was verified by the likelihood ratio test. Two random thresholds random parameters hierarchical ordered probit models were independently conducted to analyze the relationship between injury severity and risk factors. The results showed that the contributing factors to injury severity were different between the two subsamples. The three factors most likely to cause serious crashes during the daytime were ‘not keep a safe distance’, ‘drunk driving’ and ‘adverse weather’, while during the nighttime, these three factors were ‘fatigued driving’, ‘adverse weather’ and ‘drunk driving’. The same factors can cause more serious consequences during the nighttime than that during the daytime. After fully identifying the contributing factors related to injury severity, and in particular distinguishing the features of daytime and nighttime crashes, some useful suggestions were proposed to mitigate crash injury severity.

Investigation of factors contributing to bus-crash severity based on extended hierarchical ordered probit model with heteroscedasticity

A bus-involved crash unavoidably affects many passengers, causing severe consequences. Nonetheless, few studies have been comprehensively conducted to analyze the bus-crash severity. Accordingly, using the data of the bus crashes that have occurred in the UK in 2017–2019, this study constructs an extended hierarchical ordered probit model with heteroscedasticity (HHOPIT) to investigate the factors influencing bus-crash severity levels. Using AIC and BIC indicators, we confirm that considering the impact of variable heterogeneity on the threshold and error term is essential for interpreting bus-crash severity. Particularly, the frontal-impact-crash, wet-road, and moderate-speed-limit variables produce heteroscedasticity and significantly affect the crash severity level. Additionally, overturned-crash, multi-vehicle-crash, and morning-crash variables induce threshold variability. Based on the marginal effect of variables, it is confirmed that road-section, frontal-impact-crash, wet-road, junction-exit, and main-way variables increase the likelihood of severe injuries. Furthermore, the probability of slight injuries as consequences, increases with the influence of lower-speed-limit, roundabout, urban-crash, parking-crash, and moderate-speed-limit variables. This study additionally determines the factors that have inconsistent effects on the consequence of bus and passenger-car crashes. To improve the safety of bus services, the authorities must carefully consider the heterogeneity and specificities of the factors affecting bus-crash severity and appropriately formulate targeted policies.

Examining driver injury severity in left-turn crashes using hierarchical ordered probit models

Objective Few existing studies in the literature devoted efforts to examine the driver injury severity in left-turn crashes. To fill this research gap, this paper aims to provide a comprehensive study of the contributing factors of left-turn crashes and the corresponding injury severities. Methods The hierarchical ordered probit (HOPIT) model is first applied to study driver injury severity in left-turn crashes. The HOPIT model can overcome the limitations of traditional ordered probit models since its thresholds are always positive and ordered. It is a function of unique explanatory parameters that do not necessarily affect the ordered probability outcomes directly. Considering the driving condition during the wintertime could be significantly different from other seasons, this study divided the overall crash dataset into “winter” and “other-season” subsets based on the temperature, snowing condition, and other factors. Results With the “other-season” dataset, results demonstrated that contributing factors, such as young drivers, male drivers, clear, light, and ramp intersection with crossroad, are associated with a decrease in injury severity. On the contrary, factors like drug, alcohol, disregard traffic control device, high-speed limit, the protected left-turn signal, etc., are related to an increase in injury severity. In winter, results revealed that only nine contributing factors are significant to the left-turn crash. Alcohol, disregard traffic control device, nighttime, high-speed limit, head-on collision, and state road are associated with an increase in injury severity. Also, two-vehicle involved, snow, ramp intersection with crossroad are related to a decrease in injury severity. Conclusions The HOPIT model is applied to examine contributing factors of left-turn crashes and the corresponding injury severity, based on left-turn crash records from 2010 to 2017 in Utah. Eighteen significant factors of left-turn crash injury severity are identified in the overall dataset. In seasons rather than winter, the significant factors are almost the same as that of the entire year. In the winter, less significant factors and different patterns are found compared with the overall crashes.

Temporal stability of driver injury severity in single-vehicle roadway departure crashes: A random thresholds random parameters hierarchical ordered probit approach

This study examines contributing variables affecting the driver injury-severity in single-vehicle roadway departure crashes. To capture the threshold heterogeneity and unobserved heterogeneity, the random thresholds random parameters hierarchical ordered probit (HOPIT) approach is employed using the crash data collected from 2014 to 2017 in the State of North Carolina. Three injury severity levels are considered: severe injury, minor injury, and no injury. Attributes that potentially affect crash severity are examined, including driver, crash, roadway, and environmental characteristics. A series of likelihood ratio tests are conducted to examine the temporal stability of factors across different studied years. The marginal effects of factors in different injury-severity models are compared. Significant temporal instability is found among the studied year periods. The threshold value estimated using the random threshold random parameters HOPIT model is found to be random parameters and determined by specific explanatory variables. Additionally, the effects of some factors (e.g., alcohol, curved roadway, passenger car, SUV, and wet/water surface) on injury severity are relatively stable, while others (e.g., female driver, collector, and clear weather) present temporal unstable effects. Regarding the temporal instability, decision-makers need to treat the factors carefully to avoid developing incorrect countermeasures.

Specification and Testing of Hierarchical Ordered Response Models with Anchoring Vignettes

AbstractCollection and analysis of self-reported information on an ordered Likert scale is ubiquitous across the social sciences. Inference from such analyses is valid where the response scale employed means the same thing to all individuals. That is, if there is no differential item functioning (DIF) present in the data. A priori this is unlikely to hold across all individuals and cohorts in any sample of data. For this reason, anchoring vignettes have been proposed as a way to correct for DIF when individuals self-assess their health (or well-being, or satisfaction levels, or disability levels, etc.) on an ordered categorical scale. Using an example of self-assessed pain, we illustrate the use of vignettes to adjust for DIF using the compound hierarchical ordered probit model (CHOPIT). The validity of this approach relies on the two underlying assumptions of response consistency (RC) and vignette equivalence (VE). Using a minor amendment to the specification of the standard CHOPIT model, we develop easy-to-implement score tests of the null hypothesis of RC and VE both separately and jointly. Monte Carlo simulations show that the tests have good size and power properties in finite samples. We illustrate the use of the tests by applying them to our empirical example. The tests should aid more robust analyses of self-reported survey outcomes collected alongside anchoring vignettes.

Is the front passenger seat always the “death seat”? An application of a hierarchical ordered probit model for occupant injury severity

Although many studies have investigated the correlations between injury severities and seat positions, few researchers explored the correlates of injury severities (e.g., seat positions) within a crash that results in multiple occupant injuries. Therefore, we examine the injury correlates within and between crashes, and study the correlations between seat positions and occupant injury severity by constructing a hierarchical ordered probit model. A total of 20,327 occupant injuries in 16,405 motor vehicle crashes in South Australia (2012 − 2016) are used. The results of this study indicate that the rear left passenger seat is associated with a 7.66% higher chance of getting injured (including moderate and severe injury), and the front left passenger seat is associated with a 2.94% higher chance of getting injured compared with the driver seat. Besides, the higher injury chances for other passenger seats including the rear right and rear middle seats are 4.97% and 4.74%, respectively, compared with the driver seat. Thus, this study offers passengers insightful suggestions about how to protect themselves by choosing the right passenger seat in a vehicle.

Investigating the injury severity of single-vehicle truck crashes in a developing country

Trucking plays a vital role in economic development in every country, especially countries where it serves as the backbone of the economy. The fast growth of economy in Iran as a developing country has also been accompanied by an alarming situation in terms of fatalities in truck-involved crashes, among the drivers and passengers of the trucks as well as the other vehicles involved. Despite the sizable efforts to investigate the truck-involved crashes, very little is known about the safety of truck movements in developing countries, and about the single-truck crashes worldwide. Thus, this study aims to uncover significant factors associated with injury severities sustained by truck drivers in single-vehicle truck crashes in Iran. The explanatory factors tested in the models include the characteristics of drivers, vehicles, and roadways. A random threshold random parameters hierarchical ordered probit model is utilized to consider heterogeneity across observations. Several variables turned out to be significant in the model, including driver’s education, advanced braking system deployment, presence of curves on roadways, and high speed-limit. Using those results, we propose safety countermeasures in three categories of 1) educational, 2) technological, and 3) road engineering to mitigate the severity of single-vehicle truck crashes.

Inflation expectations in India: Learning from household tendency surveys

We use a large household survey that is being conducted by the Reserve Bank of India since 2005 to estimate the dynamics of aggregate inflation expectations over a volatile inflation regime. A simple average of the quantitative responses produces biased estimates of the official inflation data. We therefore estimate expectations by quantifying the reported directional responses. We perform quantification by using the hierarchical ordered probit model, in addition to the balance statistic. We find that the quantified expectations from qualitative forecasts track the actual inflation rate better than the averages of the quantitative forecasts, highlighting the filtering role of qualitative tendency surveys. We also report estimates of the disagreement among households. The proposed approach is particularly suitable in emerging economies, where inflation tends to be high and volatile.

A note on accounting for underlying injury-severity states in statistical modeling of injury accident data

This study provides an empirical analysis of the severity outcomes of injury accidents (i.e., accidents that resulted in an injury-involved outcome) by exploring the possibility of two underlying injury-severity states: the minor-injury state and the ordered injury state. The former may reflect the generation mechanism of slight-injury accidents with limited potential to result in more severe injury outcomes, whereas, the latter may represent slight-injury accidents that share similarities with the mechanism of the more severe outcomes, in terms of their occurrence circumstances. To account for the possible presence of these two underlying regimes, a zero-inflated hierarchical ordered probit model with correlated disturbances is estimated using injury-severity data from single-vehicle accidents occurred in Scotland, UK between 2016 and 2017. The results of the empirical analysis provide statistical evidence on the possible presence of underlying states, with the identified effect of traditional driver-, vehicle-, roadway- and accident-specific determinants revealing the nuances and structural differences of such injury-severity states.

Analysis of accident injury-severity outcomes: The zero-inflated hierarchical ordered probit model with correlated disturbances

In accident injury-severity analysis, an inherent limitation of the traditional ordered probit approach arises from the a priori consideration of a homogeneous source for the accidents that result in a no-injury (or zero-injury) outcome. Conceptually, no-injury accidents may be subject to the effect of two underlying injury-severity states, which are more likely to be observed in accident datasets with excessive amounts of no-injury accident observations. To account for this possibility along with the possibility of heterogeneity stemming from the fixed nature of the ordered probability thresholds, a zero-inflated hierarchical ordered probit approach with correlated disturbances is employed, for the first time – to the authors’ knowledge – in accident research. The latter consists of a binary probit and an ordered probit component that are simultaneously modeled in order to identify the influential factors for each underlying injury-severity state. At the same time, the model formulation accounts for possible correlation between the disturbance terms of the two model components, and allows for the ordered thresholds to vary as a function of threshold-specific explanatory variables. Using injury-severity data from single-vehicle accidents that occurred in the State of Washington, from 2011 to 2013, the implementation potential of the proposed approach is demonstrated. The comparative assessment between the zero-inflated hierarchical ordered probit approach with correlated disturbances and its lower-order counterparts highlights the potential of the proposed approach to account for the effect of underlying states on injury-severity outcome probabilities and to explain more with the same amount of information.

Estimation on the health life expectancy of adults in Zhejiang province, 2016

Objective: To estimate the health-adjusted life expectancy (HALE) of adults in Zhejiang province and evaluate the health status of the adults. Methods: This study was based on the mortality data collected from Zhejiang Chronic Disease Surveillance Information and Management System, and mortality rates from the underreporting survey and self-reported health data in 2016. Hierarchical Ordered Probit (HOPIT) model was used to estimate the severity-weighted prevalence of disability. Sullivan's method was used to calculate the HALE. Results: After adjustment by HOPIT model, the severity-weighted prevalence of disability increased significantly with age (χ(2)=5 795.81,P<0.001), and it was higher in females than in males (χ(2)=5 353.27, P<0.001). The life expectancy and self-evaluated HALE were 59.08 years and 48.68 years, respectively, in those aged ≥20 years, the difference was 10.40 years due to disability. The proportion of HALE loss due to disability in the total life expectancy was 17.61%, and it increased with age. HALE was higher in males than in females (49.21 years vs. 48.14 years), and in urban residents than in rural residents (49.92 years vs. 47.43 years). Conclusion: The proportion of loss of HALE in the total life expectancy in adults was high in Zhejiang, and it higher in males than in females, in urban residents than rural residents. Programs on improving health care in women and rural residents should be promoted.

Internet use and job satisfaction

Does the use of Internet for professional purposes foster employees' job satisfaction? We focus on six main work characteristics – income, education, occupation type, autonomy, time pressure and social interactions – and we develop new hypotheses on how Internet use moderates the effects of these factors. We use data from the European Working Conditions Survey, and estimate a bivariate ordered probit model, and a hierarchical ordered probit model. The results point out that Internet technologies enhance job satisfaction by improving access to data and information, creating new activities, and facilitating communication and social interactions. However, these positive effects are skewed. Workers in some occupations, and with higher income and education levels, benefit relatively more from the Internet vis-a-vis workers in occupations that are more weakly related to ICTs activities.

Analysis of telecommuting behavior and impacts on travel demand and the environment

The discussion of whether, and to what extent, telecommuting can curb congestion in urban areas has spanned more than three decades. This study develops an integrated framework to provide the empirical evidence of the potential impacts of home-based telecommuting on travel behavior, network congestion, and air quality. In the first step, we estimate a telecommuting adoption model using a zero-inflated hierarchical ordered probit model to determine the factors associated with workers’ propensity to adopt telecommuting. Second, we implement the estimated model in the POLARIS activity-based framework to simulate the potential changes in workers’ activity-travel patterns and network congestion. Third, the MOVES mobile source emission simulator and Autonomie vehicle energy simulator are used to estimate the potential changes in vehicular emissions and fuel use in the network as a result of this policy. Different policy adoption scenarios are then tested in the proposed integrated platform. We found that compared to the current baseline situation where almost 12% of workers in Chicago region have flexible working time schedule, in the case when 50% of workers have flexible working time, telecommuting can reduce total daily vehicle miles traveled (VMT) and vehicle hours traveled (VHT) up to 0.69% and 2.09%, respectively. Considering the same comparison settings, this policy has the potential to reduce greenhouse gas and particulate matter emissions by up to 0.71% and 1.14%, respectively. In summary, our results endorse the fact that telecommuting policy has the potential to reduce network congestion and vehicular emissions specifically during rush hours.