Only 12.3 percent of people in Spain reported buying transport services digitally — bus, train or flight tickets, taxi or ride-hailing bookings — in the three months before a major national survey, despite the country’s relatively strong digital skills. That striking gap between connectivity and actual digital transport spending is the starting point for a new study published in the European Transport Research Review, which dissects who makes transport-related digital purchases and why.
The research, led by Thais Rangel with Daniela Castaño-Herrera, Leise Kelli Oliveira and José Manuel Vassallo, analysed anonymised microdata from the 2022 Eurostat Survey on the use of Information and Communication Technology in households and by individuals. Although the survey is harmonised across all European Union member states, the team obtained disaggregated Spanish microdata directly from Eurostat under a formal request. Of 14,839 initial Spanish records, 13,061 remained after cleaning, covering individuals aged 16 or above.
The study’s central methodological move was to treat buying a ticket online as a distinct behaviour rather than a by-product of general technology adoption. Most earlier work on digital mobility has examined access to travel information or the uptake of transport apps, the authors note, without isolating the transaction itself. The Eurostat questionnaire, however, aggregates all such purchases into a single variable spanning public transport, air travel, taxi and ride-hailing, so the analysis captures a broad propensity to transact digitally rather than mode-specific behaviour.
To model this outcome, the team applied a Generalised Structural Equation Modelling approach, estimated with Diagonally Weighted Least Squares because the dependent variable is binary. Two latent behavioural constructs were built from survey items: Personal Internet Use, covering everyday activities such as social networking, reading online news, streaming video, seeking health information, booking medical appointments online and internet or mobile banking; and Learning-oriented Internet Use, covering online courses, digital learning materials, communication with educators or fellow learners through digital platforms, and internet use for educational, professional or personal learning. An exploratory factor analysis revealed the two-factor structure, and a confirmatory factor analysis validated it, with standardised loadings between 0.69 and 0.97, Cronbach’s alpha of 0.80 for Personal Internet Use and 0.84 for Learning-oriented Internet Use. Overall model fit was strong: a comparative fit index of 0.999, RMSEA of 0.031 and SRMR of 0.032.
The headline result is that Personal Internet Use is the variable most strongly associated with transport-related digital purchases, followed by Learning-oriented Internet Use, which shows a significant but smaller positive association. In other words, people who already live much of their lives online — banking on their phones, streaming, managing appointments through apps — are markedly more likely to buy their mobility digitally too. The learning-oriented link is, according to the authors, novel: prior literature does not appear to have explored that specific relationship. The two constructs remained positively correlated with each other even after controlling for sociodemographic characteristics.
Among observed variables, income was positively associated with the behaviour, with the largest estimated association for the highest-income group relative to the lowest. Age worked in the opposite direction: younger adults, particularly those aged 16 to 24, were more likely to buy transport digitally, while older people were less likely.
Those in the “other” employment category — a group that includes retirees, people unable to work because of long-standing health problems, and those engaged in domestic tasks — also showed lower levels of engagement than employed individuals. The authors suggest this lower engagement may be related to limited disposable income and reduced travel needs, and they note supporting evidence from Poland indicating that economically inactive groups, including retirees, tend to make less frequent use of mobile applications for public transport services. Student status, by contrast, did not show a significant association with transport-related digital purchases once other sociodemographic factors were accounted for.
Household characteristics produced more nuanced patterns. Larger household size was negatively associated with the behaviour, but this association reached statistical significance only for the total effect, not the direct effect. The number of children in the household was positively associated with both latent constructs, yet its direct effect on purchasing was negative and statistically significant while its total effect was not, indicating that the direct and indirect components operate in opposing directions and cancel out into no statistically significant overall association.
Place of residence also mattered. Individuals living in transition and more developed regions showed positive total effects, with the association stronger and more statistically robust for more developed regions, highlighting persistent regional heterogeneity in engagement. The authors connect this to broader concerns that rural communities are particularly vulnerable to the digitalisation of transport services because of limited information and communication technology infrastructure, which can hinder access to real-time information and online mobility tools.
One of the more intriguing findings concerns gender. Men exhibited lower levels of both Personal Internet Use and Learning-oriented Internet Use than women, yet the direct effect of being male on transport-related digital purchasing was positive and statistically significant. Because the indirect component ran the other way, the total effect was small and not statistically significant, meaning no clear overall gender association once both pathways are considered. The authors describe this result as interesting and, to some extent, contrary to what might be expected, noting that previous research shows women are generally more active in mediated communication and social internet use and tend to engage more frequently in online educational activities, while some gender differences may persist in certain forms of digitally mediated mobility behaviour such as ride-hailing. They suggest that women’s participation in digital mobility is increasingly comparable to men’s, and that this relationship is mediated by variables related to internet use.
Students displayed a contrasting pattern that the authors read as a caution about the limits of learning-oriented digital engagement. Compared with employed individuals, students showed lower levels of Personal Internet Use but higher levels of Learning-oriented Internet Use, consistent with their stronger exposure to online educational environments. Yet student status carried no significant association with the purchasing outcome, suggesting that educationally oriented forms of internet use may not be sufficient on their own to translate into digital transport transactions.
Placing the findings in a wider context, the authors draw on the EU-funded DIGNITY project, which combined workshops, interviews and a large-scale survey in the Barcelona Metropolitan Area in 2020 and found that the digitalisation of transport systems disproportionately disadvantages older adults, women, and individuals with low education levels or limited digital confidence. Recent estimates suggest that approximately 15 percent of the population in the metropolitan area of Barcelona may lack the competence to engage with digital mobility platforms. The European Commission has likewise identified the digital divide as an underlying barrier that may compound transport poverty, particularly among older adults, low-income groups and rural residents.
Several policy implications follow from the analysis. Because Personal Internet Use shows the strongest association with the outcome, the authors argue that policies aiming to increase engagement with digital mobility services should not focus exclusively on mobility systems themselves but also on broader digital engagement, with programmes promoting everyday digital practices plausibly supporting adoption by building confidence and familiarity — although the study’s cross-sectional design cannot confirm such an effect. The negative association with age points to the risk that digitalisation may disproportionately disadvantage older adults, and the authors propose age-inclusive strategies such as simplified ticketing interfaces, assisted digital services as alternative booking channels, targeted digital literacy programmes delivered through community centres or public transport operators, and the maintenance of hybrid digital and non-digital access channels during the transition. Income and education findings suggest digital mobility services may reproduce existing socioeconomic inequalities, prompting suggestions that transport authorities integrate digital inclusion objectives into mobility policies, potentially through partnerships between operators and public institutions supporting training in digital payment and online booking. Designing platforms that minimise cognitive and procedural complexity may also reduce barriers for those with lower digital literacy, and regional policies should consider connectivity, digital infrastructure and user support when implementing digital ticketing.
The study is explicit about its limitations. The Eurostat survey was not designed to capture transport behaviour, so the dataset contains no information on individuals’ main transport mode or overall travel frequency, which may themselves be associated with digital purchasing — people who primarily rely on private cars, for instance, may have less need for digitally mediated transport services. The cross-sectional design means the associations cannot be interpreted as causal effects, and reverse or simultaneous relationships cannot be ruled out; a broader propensity to transact online, not captured by the Personal Internet Use indicators, could jointly explain both internet use patterns and purchasing behaviour. The latent constructs were measured with available survey items covering only a subset of possible activities, and the dataset lacks information on attitudes towards transport modes, perceived ease of use, service quality and the availability of transport alternatives. The aggregation of public transport tickets, air travel, taxi and ride-hailing into a single variable further means the results capture a broad propensity rather than mode-specific behaviour. The authors also caution that the 87.7 percent who did not report such purchases does not exclusively represent people using non-digital channels; some may simply not have needed transport services during the reference period.
Looking forward, the authors propose two main directions: replicating the model across multiple national contexts using harmonised Eurostat microdata to enable cross-country comparisons, and disaggregating the purchasing outcome by transport mode or service to assess whether the identified determinants differ between train, bus, air travel, taxi and ride-hailing.
Subject of Research: Sociodemographic and behavioural factors associated with transport-related digital purchases in Spain
Article Title: Which sociodemographic and behavioural factors are associated with transport-related digital purchases? Application to the case of Spain
Article References: Rangel, T., Castaño-Herrera, D., Oliveira, L. K., & Vassallo, J. M. (2026). Which sociodemographic and behavioural factors are associated with transport-related digital purchases? Application to the case of Spain. European Transport Research Review, 18(1), Article 94. https://doi.org/10.1186/s12544-026-00855-9
Image Credits: AI Generated
DOI: 10.1186/s12544-026-00855-9
Keywords: digital divide, transport digital purchases, Spain, Eurostat ICT survey, internet use, mobility services, structural equation modelling, digital inclusion, online ticketing, transport inequality, digital mobility, sociodemographic factors
News Source: Denise Maddox. (October 28, 2026). Research report: Which sociodemographic and behavioural factors are associated with transport-related digital purchases? Application to the case of Spain. Scienmag.
