Dishonesty, Norms, and Fare Evasion on Regional Trains: Preparatory Surveys and Planned Randomized Control Trials

The current phase consists of two anonymous passenger surveys preparing subsequent randomized field interventions to reduce fare evasion. All survey respondents complete the same core questionnaire; there are no treatment and control groups in this phase.
The preregistration specifies in advance a conditional framework through which survey findings will guide the prioritization of candidate interventions for the subsequent randomized trials. Final treatment selection will also depend on technical and operational feasibility. The treatment and control conditions will be specified and preregistered before the interventions begin.

2026-10-03

2027-01-31

Personal normative strictness toward fare evasion.
Normative-expectation strictness, defined as respondents’ beliefs about how other passengers judge fare evasion.
Relative leniency toward fare evasion on short journeys compared with long journeys.
Personal endorsement of material, intrinsic, reputational, and social motivations for compliance.
Perceived social endorsement of these motivations among other passengers.
Perceived probability of ticket inspection.
Perceived prevalence of fare evasion.
Perceived effect of fare evasion on regional-train ticket prices.

(Additional details are provided in the attached preregistration document.)

Personal normative strictness is constructed from four scenarios concerning long and short journeys with and without a valid ticket. Responses to the two ticket-purchase scenarios are coded so that higher values indicate stronger approval of purchasing a ticket. Responses to the two fare-evasion scenarios are reverse-coded so that higher values indicate stronger disapproval of traveling without a valid ticket. The four responses are then averaged.

Normative-expectation strictness is constructed in the same way, using respondents’ predictions of how other regional-train passengers answer the four scenarios.

Relative leniency compares the perceived acceptability of traveling without a valid ticket on a short journey with its perceived acceptability on a long journey. A positive value indicates greater leniency toward fare evasion on short journeys.

The motivational questions cover material, intrinsic, reputational, and social reasons for compliance. Items are recoded where necessary so that higher values consistently indicate a stronger perceived cost of fare evasion or a stronger motivation to comply.

Perceptions of inspection risk, fare-evasion prevalence, and effects on ticket prices are measured using ordered response categories. These responses are compared with benchmarks documented by the transport operator. Responses above the benchmark indicate overestimation, while responses below the benchmark indicate underestimation.

Responses to each individual normative-judgment question.
Responses to each individual motivational statement.
Differences between respondents’ personal answers and their predictions about other passengers.
Self-reported frequency of traveling without a valid ticket during the previous 12 months.
Self-reported possession of a valid ticket on the current journey.
Differences by regular or irregular situation, age, gender, employment status, travel purpose, travel frequency, subscription ownership, and mobile-ticketing use.
Differences between the on-board and mobile-application survey channels.
Exploratory passenger profiles based on norms, motivations, beliefs, travel characteristics, and fare-evasion status.

(Additional details are provided in the attached preregistration document.)

Individual questionnaire items will be described using their complete response distributions, modes, medians, and interquartile ranges.
Within-person differences will compare respondents’ own judgments and motivations with their beliefs about the judgments and motivations of other passengers.
Subgroup analyses will examine whether the primary outcomes differ between passengers in regular and irregular situations and across prespecified demographic and travel-related groups.
Exploratory clustering methods may be used to identify passenger profiles characterized by similar combinations of normative judgments, perceived norms, motivations, beliefs, travel patterns, and fare-evasion status. These analyses are exploratory and will not replace the prespecified primary analyses.

The project comprises a preparatory survey phase followed by subsequent randomized field trials. This registration currently documents the surveys and the prespecified framework linking survey findings to the prioritization of candidate interventions.
The surveys recruit regional-train passengers through an on-board channel and a mobile-ticketing application. Passengers’ regular or irregular travel status and application users’ high or low fare-evasion suspicion classifications are observed characteristics, not randomly assigned treatment conditions. All respondents complete the same core questionnaire.
Personal judgments and motivations are elicited before predictions about other passengers. Question order within the motivational blocks is randomized to mitigate order effects. Random selection for accuracy-based payment provides incentives for selected survey questions. Neither procedure is intended to identify the causal effect of a fare-evasion intervention.
Survey findings will be interpreted using the conditional framework specified in Section 6 and Appendix E of the preregistration. This framework identifies candidate interventions and potentially inappropriate messages across the full sample and prespecified subgroups. Final intervention choices remain subject to technical and logistical feasibility.
There are no treatment and control groups in the current survey phase. The subsequent trials’ treatment and control arms, assignment procedures, randomization units, sample sizes, outcomes, and analysis plans will be specified and preregistered before the interventions begin.

Not available

Randomization is performed automatically by the survey software.
For each respondent, the seven personal-motivation statements are presented in a randomized order. The seven corresponding questions about the perceived motivations of other passengers are also presented in a randomized order.
In addition, each valid respondent is randomly assigned one unused outcome from a pregenerated payment lottery. Every outcome has a one-percent probability of payment eligibility.
There is no random assignment to different behavioral interventions or different substantive versions of the questionnaire.

The individual survey respondent is the unit of randomization for question order and lottery selection.
The passenger’s regular or irregular situation, recruitment channel, and mobile-application suspicion classification are observed characteristics and are not randomly assigned.

No

At least 2,000 individual respondent-level randomization units. The design is not clustered.

Between 2,000 and 6,000 completed individual survey responses. The study will target a minimum of 2,000 valid completed responses, with recruitment continuing up to a maximum of 6,000 responses, depending on field access and response rates.

Current survey phase: no treatment or control arms; all respondents complete the same core questionnaire. Recruitment targets at least 2,000 completed responses, including at least 100 respondents externally classified as travelling in an irregular situation or as high-suspicion application users. Treatment-arm allocation and sample sizes for the subsequent randomized trials have not yet been determined and will be preregistered before the interventions begin.