Attachment-Based Appeals and Their Effect on Donations: Evidence from a Zoo Field Experiment

The experiment will be conducted in the hippo house, a building dedicated to hippo exhibits, over 14 consecutive days (September 28–October 11, 2025). The days will be randomly assigned into 7 treatment days and 7 control days.

During the treatment days, we will implement two attachment-based appeals designed to increase visitors’ emotional connection to the exhibited animal. Appeals in this study include both interactive (photo-taking) and visual (donation box design) interventions:

1) Interactive appeal (photo-taking event):
We will organize a photo-taking event encouraging visitors to photograph the hippo and submit their photos to the zoo. Posters inside the exhibit will also promote participation. Based on Diehl et al. (2016) and Hanisch et al. (2019), taking photos is expected to strengthen attachment to the subject of the photos. This intervention tests whether such engagement increases donations.

2) Visual appeal (donation box design):
The donation box inside the hippo house will be changed from a plain design to a hippo-themed decorated box. The aim is to test whether such visual design increases the feeling of donating “to the hippo directly” and thereby promotes giving.

References:
- Diehl, K., Zauberman, G., and Barasch, A. (2016). How taking photos increases enjoyment of experiences. Journal of Personality and Social Psychology, 111(2), 119.
- Hanisch, E., Johnston, R., & Longnecker, N. (2019). Cameras for conservation: wildlife photography and emotional engagement with biodiversity and nature. Human Dimensions of Wildlife, 24(3), 267-284.

2025-09-28

2025-10-11

1. Donation amount collected from the hippo exhibit donation box.
2. Donation amount collected from the two donation boxes located at the zoo’s entrances

*We will also record the number of contributions—coins/bills—where feasible.

1. Liking of the hippo, measured on a 5-point Likert scale in an iPad survey at the exhibit exit.
2. Number of QR code scans for photo submission (indicator of interest).
3. Number of photo submissions received via the QR code system (indicator of active participation).

The unit of randomization is the day. The field experiment will be conducted in the hippo house over 14 consecutive days (September 28–October 11, 2025). Seven days will be randomly assigned to the treatment group and seven days to the control group. The hippo house provides a suitable setting because the visitor flow is one-directional, from the entrance to the exit, which facilitates the consistent delivery of interventions along the path.

On treatment days, we implement two attachment-based appeals: (i) an interactive appeal consisting of a photo-taking event promoted by three posters displayed inside the hippo house, and (ii) a visual appeal, in which the donation box is decorated with hippo-themed designs instead of a plain box. In addition, a poster encouraging photo submission and the iPad survey are placed at the exit of the hippo house. Appeals in this study thus include both interactive (photo-taking) and visual (donation box design) interventions.

On control days, the photo-taking event is not held, and the donation box remains plain. However, the exit poster encouraging photo submission and the iPad survey are present on both treatment and control days, allowing us to measure secondary outcomes (liking of the hippo and photo submission interest) under comparable conditions.

Not available

We employ stratified randomization to ensure balance across key factors that may influence donations and visitor engagement. The stratification variables are:

1) Day type (weekday vs. weekend),
2) Calendar week (first vs. second week of the study: Sep 28–Oct 4 vs. Oct 5–Oct 11, 2025), and
3) Number of hippos on exhibit in the hippo house (recorded daily).
Within each stratum, days are randomly assigned to treatment and control groups in approximately equal numbers (e.g., 3 vs. 4 days), using computer-generated random numbers.

We further checked for balance in expected attendance across treatment and control days using visitor counts from the same calendar period in the previous year (September 29–October 12, 2024, also spanning two weeks starting on the last Sunday of September). Statistical tests showed no significant differences, suggesting that treatment and control days are comparable in terms of expected attendance.

Day.

No

N/A

The unit of randomization is the day. As the study covers 14 days in total (7 treatment and 7 control days), the number of randomized units is 14. However, the unit of our primary analysis is each monetary contribution (coin or bill) placed into the donation boxes. Therefore, the sample size for the analysis will be much larger than 14, corresponding to the total number of contributions collected during the study period.

Same as above

N/A