The impact of Resilience Training on risk and loss aversion

The intervention consists of an 80 minute long resilience training program designed to enhance participants' capacity to cope with stress, uncertainty, and setbacks. Adapted from established positive psychology frameworks, the training focuses on helping individuals reflect on personal challenges, identify inner strengths, and develop adaptive coping strategies. It is delivered through a video-based format complemented by a structured worksheet that guides participants through seven steps of self-reflection and goal setting. The intervention is designed to be practical, engaging, and suitable for experimental settings, with the objective of fostering psychological resilience that may influence economic decision-making under risk and loss.

2025-05-02

2025-10-31

Risk Aversion and Loss Aversion

The key outcomes—risk aversion (ρ) and loss aversion (λ)—will be constructed based on participants’ responses to a set of Multiple Price List (MPL) tasks designed to elicit preferences under risk and over mixed outcomes. These include both gain-only lotteries and mixed lotteries involving potential losses. We will use the DOSE (Dynamically Optimized Sequential Experimentation) approach to estimate these parameters. DOSE uses a Bayesian hierarchical modeling framework to update beliefs about participants’ preference parameters in real time, based on observed choices. This allows for efficient learning from a limited number of responses by adaptively selecting questions that are most informative. The resulting posterior distributions provide individualized and group-level estimates of risk and loss aversion, capturing both central tendencies and uncertainty in a principled way.

This is a lab-based experimental study conducted in two stages: an in-person laboratory session and a follow-up online session. The objective is to examine whether resilience training influences participants’ risk and loss aversion in monetary decision-making tasks. In the laboratory session, participants are randomly assigned by session to either a treatment group (Resilience Training) or a control group (PowerPoint skills). All participants complete an initial psychological survey, watch a video lesson with an activity sheet, complete a post-video survey, and then make incentivized lottery-based decisions involving trade-offs between risky and certain options. Approximately two weeks later, participants are invited to complete a second session online, which includes another lottery task and a final survey. One decision from each session is randomly selected for real monetary payment, allowing estimation of individual preferences. The study uses the DOSE method—a Bayesian hierarchical modeling framework—to estimate risk and loss aversion parameters from participants’ observed choices.

computer using STATA set seed commands

Lab sessions

No

130 participants

130

130 (65 treatment ; 65 control)

Power calculations are based on estimates from Chapman et al. (2024), which our experimental design closely follows in both task structure and outcome measures. The study reports mean values and standard deviations for risk aversion (ρ) and loss aversion (λ) across both representative and student samples. Using these values, we conducted two-sample t-tests in STATA assuming a significance level of 5% and 80% power, targeting a moderate effect size of Cohen’s d = 0.5. Across tasks, the Minimum Detectable Effect Sizes (MDES) are as follows: Main Sample – Lottery Equivalent Tasks: Risk Aversion (ρ): Mean = 0.88, SD = 0.39 → MDES = 0.20 (≈ 22.6% of control mean) Loss Aversion (λ): Mean = 1.18, SD = 0.98 → MDES = 0.49 (≈ 41.5%) Main Sample – Certainty Equivalent Tasks: Risk Aversion (ρ): Mean = 1.09, SD = 0.29 → MDES = 0.15 (≈ 13.9%) Loss Aversion (λ): Mean = 2.16, SD = 1.27 → MDES = 0.63 (≈ 29.2%) Student Sample – Lottery Equivalent Tasks: Risk Aversion (ρ): Mean = 1.05, SD = 0.30 → MDES = 0.15 (≈ 14.3%) Loss Aversion (λ): Mean = 1.30, SD = 0.69 → MDES = 0.35 (≈ 26.9%) Student Sample – Certainty Equivalent Tasks: Risk Aversion (ρ): Mean = 1.06, SD = 0.30 → MDES = 0.15 (≈ 14.2%) Loss Aversion (λ): Mean = 2.26, SD = 0.86 → MDES = 0.43 (≈ 19.0%) These MDES values indicate the smallest change in parameters (in original units and as a percentage of the control group mean) that the study is adequately powered to detect under standard assumptions. All calculations were made using power twomeans in STATA. Clustering is not a major concern given the random assignment at the individual level within sessions. Reference: Chapman, J., Snowberg, E., Wang, S.W., & Camerer, C. (2024). Looming large or seeming small? Attitudes towards losses in a representative sample. Review of Economic Studies, p.rdae093.