Climate Game ClimarisQ Teaches Trade-Offs, Not Facts, Study Finds

What happens when you hand a climate scientist’s dilemma to a player with three flashing gauges on screen? A new study of the free online game ClimarisQ suggests the answer is not a surge of new scientific knowledge, but something subtler and arguably more valuable: a felt understanding of why climate action is so hard. The game, developed by the Institut Pierre-Simon Laplace in France, puts players in the shoes of a public decision-maker juggling ecology, money, and popularity under escalating extreme weather risk. An exploratory evaluation published in the journal Geoscience Communication by Davide Faranda, Lucas Taligrot, Pascal Yiou, and Nada Caud finds that players rate the game highly for usability and scientific credibility, and that its main educational payoff lies in reflection on complexity, trade-offs, and uncertainty rather than in the transmission of facts.

The design of ClimarisQ is deceptively simple. It is a turn-based digital game in which the player must keep three coupled parameters from reaching zero: ecology, representing environmental sustainability; money, representing economic capacity; and popularity, representing social acceptance. At each turn, policy proposals arrive from different advisory domains, and the player decides whether to implement them. The game ends the moment any one parameter hits zero. Measures that improve environmental outcomes may drain budgets or provoke public resistance, while cheap or politically popular options can quietly raise long-term climate risk. Feedback arrives through shifts in the parameters and short in-game news items describing societal or environmental consequences, with no indication of what the optimal choice might have been.

Underneath this abstraction sits a probabilistic climate engine. The climate state is summarized by atmospheric CO2 concentration, displayed on a color-coded gauge that controls the likelihood of extreme events such as heatwaves, floods, and droughts. The timing of any given event remains uncertain, but occurrence becomes more probable as cumulative emissions climb, mirroring the statistical logic of event attribution under climate change. The designers deliberately avoid naming specific countries or governance systems, using a generalized representation that can travel across classrooms, workshops, and science festivals. The learning goals, the authors stress, differ from the game objectives: ClimarisQ is not built to transmit factual knowledge or score decision-making performance, but to expose players to the feedbacks, delays, and uncertainty inherent in climate governance.

Launched publicly in mid-2022 as a free app on the Google Play Store and Apple App Store, and playable on a dedicated website, the game was disseminated through outreach events such as the Fête de la Science, Forum Météo-Climat, and Trieste Next, as well as through online climate forums, social media, and the European H2020 project XAIDA. Beyond the formal evaluation, the game has been tested in dozens of facilitated sessions with thousands of non-specialist participants, including hundreds of primary and middle school pupils at science festivals, with consistently positive feedback. Yet the first pool of voluntary users, and therefore the survey respondents, turned out to be largely adults already familiar with climate topics, a skew that shapes what the study can and cannot claim.

That evaluation took the form of an anonymous online questionnaire, administered roughly six to twelve months after launch and available in English and French. Seventy-seven users responded, with no incentives offered. The survey combined demographic questions, five-point Likert-scale ratings of ergonomics, scientific content, clarity, difficulty, and aesthetics, yes-or-no questions about perceived learning and everyday-life impact, and open-ended comments. Responses were collected between June and December 2022 and analyzed descriptively; given the modest sample size and the exploratory scope, the authors performed no inferential statistical testing, in line with guidance against overstating results from small datasets. Free-text answers were coded inductively by two authors working independently, who resolved differences through discussion.

The respondent profile confirms the self-selection effect. Participants ranged in age from 8 to 56 years, with a mean of roughly 30 and the most common age band around 25 to 35. Most were highly educated, many holding graduate-level qualifications, and a substantial fraction identified with higher education or academia, including university students, PhD candidates, and postdoctoral researchers. Respondents spanned multiple continents, from France, Germany, Italy, Romania, the Netherlands, and Belgium to the USA, India, China, Tunisia, Mexico, Brazil, and Egypt. Most discovered the game through academic or professional networks and online climate communities, with far fewer arriving via app store browsing or school channels.

The ratings were strikingly positive. Scientific content and aesthetics tied for the highest marks, each averaging about 4.4 out of 5, with respondents praising the credibility of the information and its link to real climate data and scenarios, and noting that the content was scientifically sound without being oversimplified. Ergonomics averaged about 4.3, with an interface most found intuitive, though a few mentioned a short adaptation period on mobile devices. Clarity of the in-game questions scored around 4.2. Perceived difficulty was the most variable measure, averaging roughly 3.6: some players found the challenge of keeping all indicators high frustrating, while others considered it manageable after several plays, a dispersion the authors attribute to differences in prior knowledge and to a design that emphasizes constrained trade-offs over easy success. A gender comparison revealed no significant differences for ergonomics, clarity, difficulty, or aesthetics, but women rated the scientific content significantly higher than men, with a median of 5 versus 4 and a small-to-moderate effect size.

The educational findings are the most thought-provoking part of the study. A majority of respondents reported that the game did not substantially change their everyday lives, and many said they acquired no new factual knowledge, often because they were already familiar with the science. Yet even these respondents emphasized that the game helped them appreciate the complexity of climate decision-making, particularly the need to balance environmental, economic, and social constraints. Those who did report learning something new most often cited a clearer grasp of trade-offs, of how ecological objectives can conflict with financial resources or public acceptance, and of the role of chance in extreme events despite prudent choices. The authors interpret this as reflective or affective learning rather than information transfer, a form of learning they note may be longer lasting as facts evolve. Encouragingly, the vast majority said they would recommend the game, frequently as a teaching tool for students or non-specialist audiences, even when they had not learned new content themselves. A minority withheld recommendation, citing perceived difficulty, limited replayability, or mismatched expectations.

The authors are candid about the limits of their evidence. The sample was small, self-selected, and skewed toward educated, climate-aware adults, so the findings cannot be generalized to younger or less climate-literate audiences. The evaluation relied entirely on self-reported perceptions, without pre-tests, control groups, or validated instruments, so references to learning mean perceived learning or reflection, not measured knowledge gains or behavioral change. Respondents were also not asked to document how long they played or whether they played alone or in a group. The authors, who are also the game’s designers and developers, frame the study as a formative assessment meant to inform future development rather than an external validation, and they disclose that they have no commercial interests in the freely available game.

What emerges is a nuanced picture of what serious games can and cannot do for climate education. ClimarisQ works best not as a standalone lesson but as a complement to facilitated sessions, where gameplay is followed by debriefing discussions that connect decisions, trade-offs, and outcomes to real-world climate governance, addressing uncertainty, feedbacks, ethics, and the limits of individual versus collective action. In the authors’ experience, learning is often supported through ongoing dialogue during play, with scientific explanations introduced progressively at each decision point, and through formats in which small groups represent institutional actors who must deliberate and justify collective choices. The results align with a broader literature, including evaluations of World Climate, KEEP COOL, and the CAULDRON attribution game, showing that games can build trust, enthusiasm, and systems-oriented reasoning. The team’s conclusions point toward classroom-based studies, validated assessment tools, structured in-game facilitation support, and training for scientists and educators as the next steps. For now, the message is clear: a game cannot teach the climate in a few turns, but it can make the agonizing arithmetic of climate policy feel real, and that may be the lesson that sticks.

Subject of Research: Evaluation of the ClimarisQ serious game for climate education and decision-making

Article Title: ClimarisQ: What can we learn from playing a game for climate education?

Article References: Faranda, D., Taligrot, L., Yiou, P., & Caud, N. (2026). ClimarisQ: What can we learn from playing a game for climate education?. Geoscience Communication, 9(1), 115-125. https://doi.org/10.5194/gc-9-115-2026

Image Credits: AI Generated

DOI: 10.5194/gc-9-115-2026

Keywords: ClimarisQ, serious games, climate education, climate communication, extreme weather, decision-making, uncertainty, trade-offs, public engagement, Geoscience Communication, climate risk, experiential learning