ARE WE READY TO EMBRACE CONNECTED AND SELF-DRIVING VEHICLES? A CASE STUDY OF TEXANS

Automated and autonomous vehicles (AVs), connected vehicles (CVs), and connected-autonomous vehicles (CAVs) promise major advances in transportation, with the potential to eliminate a large share of crashes attributable to human error. However, the convenience they offer may induce additional travel demand, complicating safety and congestion outcomes and motivating consideration of smart congestion-pricing strategies. Emerging shared autonomous vehicles (SAVs) could alter residential location choices by making both central living (due to dense, low-cost fleets) and suburban living (due to productive travel time) more attractive, raising policy questions regarding land prices and pricing of SAV services. Public opinions will heavily influence how this future unfolds. Prior surveys generally show cautious attitudes toward AVs, with concerns focused on price, safety, and security. Building on earlier work, this study surveys 1,088 Texans to assess WTP for CAV technologies, perceived benefits and concerns, SAV adoption under different prices, potential home-location shifts, adoption timing, and views on tolling policies, and estimates econometric models to identify multivariate relationships with demographics, travel behavior, crash history, and built environment.

A broad set of academic, professional, industry, and web-based surveys has explored public perceptions of AVs/CAVs. Findings commonly indicate caution toward AVs, with safety, affordability, and information security as primary concerns. Academic studies: Casley et al. (2013) found safety and legislative issues as top concerns and a gap between expected cost and WTP. Shoettle and Sivak (2015) reported preferences for non- or partially autonomous modes, with younger motorists and men more accepting of higher automation. Bansal and Kockelman (2015) found average WTP of $67 for connectivity and $5,857 for Level 4 automation, and mixed comfort with data sharing. Bansal et al. (2015) (Austin) identified equipment failure as a key concern. Underwood (2014) reported experts citing legal issues and technological limits as main barriers and high safety thresholds for AV acceptability. Howard and Dai (2013) noted safety as the most attractive AV feature. Kyriakidis et al. (2015) highlighted concerns about information security and liability and a wide WTP distribution. Industry and web surveys (Accenture, J.D. Power, Puls Marktforschung, Cisco, Goldman Sachs/Motor Fan, Ipsos MORI, Continental, KPMG, Insurance.com, Open Robotics, NerdWallet) echo these themes: mixed enthusiasm, higher acceptance among men/younger respondents, significant concern about reliability and hacking, and sensitivity to monetary incentives. This study extends prior work by incorporating additional aspects (e.g., home relocation and SAV pricing scenarios) and estimating ordered probit and interval regression models for multivariate inference.

A statewide Texas survey (93 questions across 7 sections) was administered via Survey Sampling International’s panel using Qualtrics in June–July 2015. Topics included perceived AV/CV benefits and concerns, crash and moving-violation history, opinions on speed regulation, WTP for Level 1/2 features and automation/connectivity, interest in CV features, SAV and TNC usage, SAV adoption under per-mile pricing ($1, $2, $3), potential home-location shifts in an AV/SAV future, opinions on congestion pricing policies, travel patterns, and demographics. From 1,297 completes, 1,088 remained after excluding speeders (<15 minutes), those not understanding NHTSA automation definitions, under-18s, inconsistent responses, and other sanity checks. Population correction weights were computed to address sample biases: person-level weights by gender, age, and education (based on 2013 ACS PUMS for Texas), and household-level weights by household size, number of workers, and vehicle ownership. Respondents’ home locations were geocoded (Google Maps API), with IP-based proxies when necessary, and linked to census tract data (population density, poverty, employment density). Interval regression (IR) was used for WTP outcomes because respondents selected WTP intervals with right-censoring (e.g., “$3,000 or more”). Ordered probit (OP) models were used for ordinal outcomes: interest in connectivity, adoption timing (never, when 50% friends adopt, when 10% adopt, as soon as available), SAV adoption frequencies under price scenarios, home-location shift choices (closer, same, farther), and support for tolling policies (five-point Likert). Model estimation proceeded by including a subset of explanatory variables and removing the least significant iteratively until all remaining variables had p-values < 0.32; practical significance was assessed by standardized coefficients (>0.2 for IR) or probability shifts (>40% for OP). All estimates are population-weighted/sample-corrected. Software: Stata 12.

• Average WTP: Level 2 = $2,910; Level 3 = $4,607; Level 4 = $7,589; connectivity = $127. For L2/L3/L4, 54.4%/31.7%/26.6% would not pay more than $1,500.
• Adoption timing: about 47% would time AV adoption with friends’ adoption rates.
• SAV adoption rates: If SAVs existed today, 41.0% would not use at $1/mile, and only 7.3% would rely entirely on an SAV fleet (4.6% at $2/mile; 3.9% at $3/mile).
• Home-location shifts: 81.5% would stay at current locations in an AV/SAV future; 7.4% would move closer to the city center; 11.1% would move farther.
• Concerns: Affordability (64.5% very worried) and equipment failure (61.4% very worried) are top concerns; legal liability (52.9%) and hacking (55.1%) also rank high; privacy breaches and learning to use AVs are least concerning (34.7% and 24.7% very worried, respectively).
• Expected benefits: 53.9% expect very significant improvements in fuel economy; 53.1% expect very significant crash reductions.
• Interest in Level 4 AVs (if affordable): 71.5% at least slightly interested (21.9% very, 28.6% moderately, 21.0% slightly); 28.5% not interested. Preferred AV contexts: freeways (60.9%) and scenic areas (58.6%); least comfortable on congested streets (36.1%). Top anticipated in-vehicle activities: talking to passengers (59.5%) and looking out the window (59.4%).
• Connectivity features: Highest interest in automatic emergency notification (71.5%) and vehicle health reports (68.5%). Most adopted currently: real-time traffic info (15.6%) and steering-wheel smartphone controls (13.4%). Only 39% interested in adding connectivity even if affordable; 29.3% would pay $0.
• Policy support: Among tolling options, most support for tolling congested highways if revenues reduce property taxes (37.3% support). Texans are most supportive of adaptive traffic signal timing (64.0%) and least supportive of real-time parking price adjustments (20.5%) when 80% of vehicles are connected.
• Model-based relationships: Older respondents value automation less and are slower adopters; more experienced licensed drivers have greater interest in and higher WTP for connectivity; awareness (e.g., having heard of Google’s self-driving car) and support for speed governors are associated with higher WTP for automation; higher income increases WTP. Greater distance to transit correlates with lower WTP for L3/L4 and lower SAV usage; higher annual VMT associates with greater WTP for L4 and more frequent SAV use (especially at $1/mile). Those with more moving violations show higher WTP for connectivity and higher SAV adoption. Caucasians, licensed/experienced drivers, and those farther from transit stops are less supportive of tolling policies and tend to use SAVs less frequently. Households owning at least one Level 2-equipped vehicle and those farther from city center are more likely to move closer to the center in an AV/SAV future, while those farther from transit, making more social trips, and familiar with UberX tend to shift farther out.

The study addresses how Texans’ perceptions, preferences, and concerns shape the likely adoption and diffusion of CAV technologies and SAV services. By linking detailed demographics, travel behavior, safety history, and built-environment measures to WTP, adoption timing, usage frequency, and policy support, the findings identify groups more or less inclined to adopt and pay for automation and connectivity. This enables inference on where and among whom CAV technologies may penetrate first, informing targeted outreach and infrastructure planning. For instance, older individuals’ lower valuation of automation and greater dependence on peers’ adoption suggest that outreach and demonstrations may be crucial to accelerate acceptance. Awareness and pro-safety attitudes (support for speed enforcement/governors) correlate with higher WTP and earlier adoption, suggesting that safety messaging could be effective. The results also highlight potential induced demand and the need for pricing strategies; while Texans show mixed support for tolling, adaptive signals and variable speed limits receive relatively higher support. Anticipated limited changes in home location for most respondents imply gradual land-use impacts, though subsets may move closer to centers to leverage SAV availability. Overall, these relationships are pertinent for forecasting adoption trajectories, planning charging/parking for potential electric SAV fleets, and designing equitable, acceptable congestion management policies.

Ordered probit and interval regression models were used to quantify how demographics, built environment, travel and safety histories, and attitudes relate to Texans’ CAV valuations and prospective behaviors. Key contributions include population-weighted WTP estimates ($2,910 for L2, $4,607 for L3, $7,589 for L4; $127 for connectivity), characterization of SAV usage at various price points, identification of top concerns (affordability and equipment failure), and multivariate determinants of adoption timing, home-location shifts, and tolling-policy support. Practically significant predictors (e.g., age, experience, income, transit proximity, safety-policy attitudes, and awareness) provide actionable insights for targeting education, deployments, and supportive infrastructure. Future work should collect more data over time and across regions to capture evolving opinions and behaviors, refine forecasts of technology adoption and VMT, and guide policies (e.g., credit-based congestion pricing) and infrastructure (e.g., charging and parking for electric SAVs) to maximize public benefits and manage potential negative externalities.

Findings reflect a Texas-wide, cross-sectional web-panel survey conducted in mid-2015; opinions are likely to evolve as technologies mature and public awareness changes. Despite population weighting and data cleaning, generalizability beyond Texas may be limited. Some models excluded safety/tech predictors (e.g., in tolling-policy support), and certain measures (e.g., residential location via geocoding with occasional IP proxies) may introduce measurement error. The study notes the need for longitudinal data and replication in diverse locations to improve inference and forecasting.