Increasing Environmental Giving at Airline Checkout

Study record. Increasing Environmental Giving at Airline Checkout: Evidence from Three Simulated Booking Experiments reports three experiments with confidential airline partners (N = 1,752). We label the public study arms Airline A, Airline B, and Airline C. In Phase 1, participants completed simulated booking flows and could keep their study payment or donate part of it to charity. Phase 2 varied donation amounts after that choice, but those later donations were hypothetical.

What we tested

Checkout donation requests are common, yet a plain, low-information request leaves people unsure what their contribution supports. Each experiment compared an airline's baseline request with an integrated redesigned offer. The packages varied by airline, but combined clearer destination information, visual cause context, social-proof wording, and donation-choice design.

The experiment estimates the effect of that integrated package. A dedicated component experiment can estimate the contribution of an image, social-proof message, suggested amount, or active-choice prompt.

Phase 1: real-money participation rose

In Phase 1, a donation was deducted from the participant's study payment and transferred to charity. Baseline uptake was tightly grouped across the three flows, between 17.8% and 19.3%. Redesigned offers reached 26.4% to 35.8%.

• Airline A (N = 485): 18.6% in the baseline flow and 29.7% with the redesign, a +11.1 percentage-point difference (p = .005).
• Airline B (N = 774): 17.8% in the baseline flow and 26.4% in the comparable redesigned flow with a voluntary response, a +8.5-point difference (p = .025). A separate forced-choice version reached 41.8%; that version estimates the combined effect of the redesign and response requirement.
• Airline C (N = 493): 19.3% in the baseline flow and 35.8% in the redesigned offer, a +16.5-point difference (p < .001). This comparison also changed the named charity, so it estimates an offer-and-cause package.

Composite brand favorability was statistically stable in every Phase 1 experiment. The ratings capture one simulated exposure; a live request test can measure ticket completion, abandonment, and long-term brand behavior.

Phase 2: higher asks raised projected donations

Phase 2 changed the donation amount after participants had already completed Phase 1. It used hypothetical choices and reports projected-donation results.

• Airline A's high-versus-low amount comparison projected 5.2 times as much donation per 100 participants while uptake remained statistically stable.
• Airline B projected 2.7 times as much per 100 with statistically stable uptake.
• Airline C projected 7.5 times as much per 100, while a single higher fixed request reduced uptake by 10.9 percentage points.

The pattern guides amount design: the two menu-style experiments retained a smaller option, while Airline C's fixed higher request omitted it. The tested amount ladders generated the highest projected totals in these experiments.

What the evidence supports

The most defensible conclusion is straightforward: a well-built checkout request can raise consequential donation participation in a simulated booking flow. The Phase 1 results replicate that direction across three airlines. A live test can establish the carrier-specific percentage lift, production revenue, and ticket outcomes, while a component study can isolate individual elements.

The study also supports treating a larger amount request as a live-test question. The Phase 2 projections were higher at larger amounts, and the Airline C result shows the value of keeping a smaller amount available.

Responsible deployment

An effective request should preserve a real choice. Make the acceptance and decline options equally visible, begin with no donation preselected, and use accurate, verifiable social proof. A live pilot should make the destination of funds clear and measure donation participation, the amount given, booking completion, abandonment, and customer response together.

Methods and limitations

Participants were recruited through Prolific and randomized within each experiment. The booking flows were simulated, Phase 1 donations were consequential, and Phase 2 choices were hypothetical. The three airlines used different flows and currencies, and the Airline C Phase 1 comparison changed both the charitable cause and the presentation. A preregistered live airline A/B test is the next step for ticket-level outcomes.

Data availability

This page summarizes the current working paper. Public links to underlying data and materials will be added here when they are available for release.