Reducing communication mismatch to foster conservation

How to cite

Uruena, Viviana; Ibanez, Marcela; Riener, Gerhard. 2019. "Reducing communication mismatch to foster conservation." AEA RCT Registry. AEARCTR-0004505. https://doi.org/10.1257/rct.4505-1.1

Download: BibTeX · RIS — for reference managers without a browser translator. Zotero and most others pick up this page's citation metadata directly.

Viviana Uruena* , Marcela Ibanez , Gerhard Riener

* Primary investigator

Registry ID
AEARCTR-0004505
DOI
https://doi.org/10.1257/rct.4505-1.1
First registered
Last updated
Status
Completed
Trial dates
2019-08-07 – 2020-04-30
Countries
Madagascar (SAVA region)
JEL codes
Q15, C93
Registry record
http://www.socialscienceregistry.org/trials/4505

Abstract

Conservation campaigns often struggle to persuade farmers to adopt pro-environmental practices. From field observations in the SAVA region of Madagascar, conservationists mostly use imperatives and nature-focus messages to encourage pro-environmental actions. Empirical evidence suggests that farmers and conservationists differ in the way they perceive the environment and wildlife, which reveals differences in their values. While conservationist emphasize biospheric or self-trascendence values in their messages, farmers consider the environment as a means to improve their own well-being, reflecting self-enhancement values or, in other words, an economic thinking. The communication mismatch between the two parties, might be hindering the success of promising environmental campaigns. We designed different message types varying form and content in order to investigate effective message types in persuading farmers towards pro-environmental behavior. We hypothesize that the way messages are communicated matters as farmers may identify themselves more with a certain type of message and thus, follow the desired behavior. By using a survey-based field experiment in the SAVA region of Madagascar, we focus on two forms of communication discussed in recent literature: narratives---a warm but not clear message---or imperatives---a cold but clear message. In each of these two forms, we modify the content making salient self-enhancement or self-trascendence values. While the effects on behavior from a value-mismatch have been previously studied, the novelty of this experimental design and our contribution to the literature lies in the different message types we employ to transmit the messages to incentivize a desired behavior. Madagascar is in need of effective strategies to improve pro-environmental behavior of its population. Hence, from a policy perspective, messages that are more effective are key to make real and long-lasting changes in the mindset of farmers and encourage collective actions in the villages.

Intervention

The intervention consists of four treatment groups and a control group, where participants will be randomly allocated to. In each of the 5 groups, participants will listen to one audio recorder. Participants in the control group will listen to a placebo audio. Participants in the first treatment will listen to an audio recorder with a narrative where self-enhancement values are salient at the end. Participants in the second treatment will listen to the same narrative, with the difference that self-transcendence values are salient at the end instead. Participants in the third treatment will listen to an audio recording using an imperative where self-enhancement values are salient. Finally, participants in the fourth treatment will play the same imperative, but where self-transcendence values are made salient .
Participants then will, in a real effort task, produce bookmarks that can either be used for a charitable pro-environmental donation campaign in Germany (they will have the color green) or bookmarks that will be bought by the experimenter for a defined price at the end.

Experimental design

The experiment follows a between-subject design and it takes the form of a survey consisting of 9 different sections, however, only 7 of them are part of the present project. Most of the sections are equal across participants with the exception of those related to the intervention. The intervention consists of 4 treatment groups and a control group, where participants will be randomly allocated in each of them. In each of the 5 groups, participants will listen to one audio recording. Participants in the control group will listen to a placebo audio. Participants in the first treatment will listen to an audio recording with a narrative where self-enhancement values are salient at the end. Participants in the second treatment will listen to the same narrative, where self-transcendence values are salient at the end instead. Participants in the third treatment will listen to an audio recording with an imperative where self-enhancement values are salient at the end. Finally, participants in the fourth treatment will play the same imperative, but self-transcendence values will be salient at the end instead.

After participants have listened one of the audio recordings, they will perform a real effort-task. The final product from this effort task could either be paid to the participants or they can decide to donate it to raise funds for an NGO that will plant trees in the SAVA region.
To study the effect of the different messages on participants' behavior, we will analyze how the elasticity of donation changes when the positive externality increases among the different treatments.

Design details

Randomization method
Treatment groups will be assigned (quasi-)randomly by a computerized draw from a uniform distribution by the software of the tablet at the moment of the survey, so the enumerator has no influence on the treatment shown. For the questionnaire we use the application of KoboToolbox. Moreover, households participating in the experiment will be selected randomly within the village. We will have four enumerators per village and the village will be split in 4 quadrants. Each of the enumerators starts at a random house in the quadrant and determines by a dice role whether the house will participate. Replacement strategy, if the household does not want to participate, the neighbouring house will be asked. Then she will go in a randomly determined direction and do this with the next house, until the required number of observations is reached. Within each household the male or female head of the household will be chosen to answer by a coin-flip.
Randomization unit
Individual randomization into 4 treatment arms and 1 control group.
Clusters
600 individuals
Observations
600 individuals
Arms
120 individuals in control group, 120 individuals in treatment 1, 120 individuals in treatment 2, 120 individuals in treatment 3, 120 individuals in treatment 4.
Primary outcomes
Our primary outcomes are the following: - Proportion of green bookmarks participant decides to make for donation - Elasticity of substitution of donation - Type of environmental concern: self-enhancement vs. self-trascendence
Secondary outcomes
- Relative inter-generational donation - Environmental concern

Cited by

No citation data available for this registration yet.

Keywords