The Impact of Reference-Dependent Preferences on Technology Adoption

2026-04-01

2027-01-31

The primary outcome of the study is a binary variable indicating the decision to buy the technology. The variable equals '1' if the participant chooses to buy/keep the technology. On the other hand, it equals '0' if the participant chooses not to buy/return the technology.

The primary outcome is directly constructed using the decisions of the participants. For each round in the experiment, the outcome variable takes the value of 1 if the participant chooses to buy (or keep) the technology, and 0 if the participant chooses not to adopt (or return) the technology. No transformation or aggregation of this variable are used in the empirical analysis.

At the beginning of the experiment, participants are randomly assigned to treatment groups, after which they are shown the corresponding instructions. Participants will complete a comprehension test after reading the instructions to ensure understanding of the experimental procedures. The test will consist of 10 questions based on the information given in the instructions. Following the quiz, the participants proceed to the practice rounds, which allow them to become familiar with the interface of the rounds and all the relevant parameters. Thereafter, the participants will proceed to the main rounds, wherein the rounds they encounter will determine payments. The first main round is a no-policy round. Subsequently, they first encounter the insurance policy, followed by the reward policy. Finally, the participants will end the experiment after completing two additional tasks meant to elicit their loss- and risk-aversion.

Not available

The randomization is determined by a computer using a Python-based function.

Randomization is conducted at the individual level. Participants are randomly assigned to treatment groups by the computer. The experiment involves multiple treatment dimensions, all of which are randomized at the individual level.

No

The planned number of clusters is 150-175 individuals.

The planned number of observations is 300–350 individual-intervention observations, corresponding to 150–175 participants each contributing one observation per intervention (two interventions in total).

The experiment uses 4 treatment arms. With a total sample of 150–175 participants, this yields approximately 37–44 individuals per treatment arm, assuming equal allocation. The treatment arms correspond to all combinations of information and endowment.

The minimum detectable effect (MDE) is 0.20 (i.e. 20 percentage points) for the primary binary outcome. This is based on a baseline probability of 0.35 and assumes balanced assignment across treatment arms. The standard deviation of the outcome is approximately 0.5. With a planned sample size of 150–175 individuals, this corresponds to approximately 81–86% statistical power at the 10% significance level.