Moral and Emotional Drivers of Climate Beliefs and Behaviors: Experimental Evidence on Religious Framing and Guilt

We run an online survey-experiment in Italy. We recruit participants (age 18+) through a professional survey company (DEMETRA) to ensure that our sample is representative of the Italian population by age, gender and region.

Participants will first answer a set of background questions, including their gender, age, location, marital status, family size, level of education, trust in people and institutions, time use, altruistic behavior, household income, political attitudes, voting behavior, and personal religious affiliation. Following this, we will include an attention-check question to ensure engagement.
Next, participants in the treatment groups will be asked to complete an autobiographical recall task designed to induce guilt. Those in the control group will instead describe a typical day. After this emotional intervention, the relevant groups will receive either an informational treatment or a religious/moral treatment, while some groups will receive a neutral informational text about music history.
Following the treatment phase, we will include another attention-check question to confirm that participants have read the provided material.
Subsequently, they will answer several outcome questions, including their support for various pro-climate policies, their views on the prioritization of different areas of public intervention, and a real-stakes question involving the donation of a potential prize to different nonprofit organizations. We will also assess their beliefs about climate change, such as perceived costs, the role of individual behaviors, consequences for future generations, and the connection between religion and climate change.
Participants will be also asked to report on the emotions they were feeling at that moment. In addition, we will administer the Moral Foundations Questionnaire to identify the dominant moral values of each participant. After the Moral Foundation Questionnaire, participants complete a debriefing phase where they state their perceived study objective, then receive full disclosure of the actual research aims and access to psychological support if needed.

2026-02-15

2026-03-15

Support for climate mitigation policies and a real-stake behavioral task (donation to a charity) capturing a behavioral shift in environmental preferences.

Participants will be asked a series of policy support questions, rated on a 1 to 10 scale. These policies are designed to span a range of interventions: from carbon taxes and mobility restrictions to meat production limits and support for green infrastructure. Some questions explicitly introduce distributional trade-offs, such as whether the costs should fall more on the current generation or on future ones, or whether richer countries should bear more of the burden. This allows us to assess preferences around cost distribution, fairness, and political feasibility.
Additionally, participants are asked to prioritize different public policy areas, such as health, education, and the environment, to assess how climate concerns rank among other competing priorities.
Finally, a real-stakes donation task will offer a behavioral measure of climate concern. Participants who are randomly selected to win a cash prize will be asked how much of their prize (up to 50 euros) they would like to donate to different organizations, including environmental and humanitarian NGOs. This will allow us to observe actual willingness to part with resources, and whether pro-climate attitudes translate into real-world behavior.

To understand the mechanisms through which the narratives affect the aforementioned primary outcomes, we will explore the effects of these narratives on beliefs about climate change regarding costs, individual behavior, division of the burden and the connection between religion and climate change, priority of public interventions.

We also want to explore the heterogeneity in the outcomes with respect to different background characteristics of respondents and we might have a look at outcomes by sub-groups based on income, age, gender, political affiliation, educational level, employment status, religious attendance, moral values, trust and altruism. Moreover, we will also assess whether the different types of narratives trigger different emotional states.

As potential mechanisms, we aim to investigate how people's attitudes and beliefs toward climate change evolve in response to different framings and messages. To do so, we have developed a set of belief-based questions that capture various psychological, moral, strategic, and economic dimensions of climate change perceptions. These belief items will allow us to analyze which considerations are most salient for individuals and how these might shift when exposed to different informational treatments.
The first group of questions probes participants' beliefs about the material costs of climate change—both in terms of global economic and social damages and the personal or societal sacrifices needed to address it. Some questions are framed around strategic interdependence (e.g., the idea that climate action only works if many people cooperate), while others tap into atomistic beliefs, where individuals may feel their own actions are irrelevant.
Another key dimension involves intergenerational and interpersonal altruism. Some items focus on the anticipated consequences of climate change for future generations, while others explicitly measure people's willingness to make personal sacrifices for the benefit of others, whether contemporaries or posterity. In addition, moral and religious framings are used to explore how spiritual or ethical convictions (e.g., sin, reward, divine proximity) influence pro-environmental attitudes.
A complementary section focuses on individual efficacy and motivation, including items that tap into the warm-glow effect—the intrinsic satisfaction of doing the right thing—even when individual actions might seem inconsequential. These questions help capture the psychological mechanisms behind voluntary environmental action and moral commitment, independent of broader strategic calculations. We would like to use beliefs either as outcomes, helping us identify the underlying mechanisms, or as moderators, to examine the heterogeneity of treatment effects (assuming they were not affected by the treatments themselves).
Finally, we will also assess whether the different types of narratives trigger different emotional states, through participants’ self-reported responses to the PANAS questionnaire, which measures a range of positive and negative emotions experienced in the moment.

We randomized both the narratives provided to respondents and the task that should induce guilt. As a such, we have 5 groups with different treatment condition and 1 control group.

This study is based on a 3 [two climate change narratives; a neutral narrative about music] x 2 [guilt triggered; no emotion triggered] between-subject experimental design involving two types of treatments that interact with one another: an autobiographical recall task and a set of narratives. The autobiographical recall task is designed either to induce guilt or to serve as a neutral control (in which participants recall an ordinary day). Participants will also be exposed to one of three narrative conditions: an informational narrative that outlines the consequences and effects of climate change; a moral and religious narrative that frames climate change in religious terms; a neutral narrative about music history.
This results in six experimental subgroups: (1) a neutral recall with the neutral narrative—the “pure” control group; (2) a neutral recall with the informational narrative; (3) a neutral recall with the moral/religious narrative; (4) a guilt-inducing recall with the neutral narrative; (5) a guilt-inducing recall with the informational narrative; and (6) a guilt-inducing recall with the moral/religious narrative.
Following the treatments, participants will respond to a series of questions assessing their support for pro-climate policies, beliefs about the costs of climate change, the importance of individual behavior, who should bear the costs of mitigation, and the connection between certain religious concepts and climate change.

Random assignment of participants across the six experimental groups will be managed by the research team using computer-based randomization software.

Participants

No

1 country (no clusters)

6,000 respondents, with 1000 participants per condition.

Around 1,000 respondents for each of the 6 subgroups that will receive different treatment conditions (or no treatment condition).

MDE=0.15, with alpha=0.05 and power=0.8 on standardized outcome measures in within groups analysis