Virtual tree, real impact: how simulated worlds associate with the perception of limited resources

Environmental perception of natural resource scarcity is a cross-disciplinary concern spanning economics, public policy, sustainability education, and environmental studies. Human-driven environmental change (e.g., biodiversity loss, climate change, deforestation) signals that exploitation is exceeding the Earth’s regenerative capacity, and recovery may be slow or incomplete. This underscores the need for sustainable resource use, yet implementation faces challenges due to mixed perceptions about the Earth’s finite resources. Video games, increasingly complex virtual worlds, offer a flexible platform to study behavior and cognition, with resource management a common mechanic across genres. Games can serve as educational tools for sustainability and resource management, offering constrained scenarios that require strategic behavior. ACNH, a popular life simulation blending production, trade, and consumption, presents limited, replenishing resources and open-ended play, enabling examination of associations between real-world perceptions of scarcity and in-game behaviors. Leveraging ACNH’s broad appeal, this study explores whether perceptions about limited resources correlate with virtual planting and exploiting behaviors, using a global sample of 640 players. The paper proceeds with a literature review, methodology (survey and Bayesian multilevel models), results, and a discussion on how game design may motivate pro-environmental behaviors and perceptions.

The review covers three themes. (1) Exposure to nature, perception, and behavior: Perception and behavior are interlinked; diminished contact with nature is linked to weaker environmental concern, while exposure can strengthen environmental understanding and pro-environmental behavior. Experience and cultural background shape reasoning about nature, with utilitarian versus taxonomic categorization illustrating how interaction patterns reflect perception. (2) In-game behavior and real-world perception: Research on violent games and aggression is mixed; some models (e.g., GAM) suggest effects, but critiques and null findings highlight human ability to distinguish virtual from real. Conversely, digital game-based learning shows promise: immersive virtual environments and rule-based goals can foster environmental knowledge and behaviors (e.g., reduced energy use, increased recycling). Outcomes depend on mechanics, necessitating investigation of motivations behind pro-environmental behaviors in games. (3) Economic rationales in ACNH: ACNH’s economy centers on producing and trading limited, time-replenishing resources (fruits, wood, fish, bugs). Over-exploitation risks resource loss (e.g., tree removal), whereas planting (trees, flowers) is rewarded. Activities span planting and exploiting for profit, crafting, aesthetics, and crossbreeding. Given limited resources and multiple needs, players may enhance resources while learning about scarcity through game constraints. The study asks: Is perception of limited resources correlated with creating (developing) more resources?

Design and data collection: A global online survey (May 15–30, 2020) using convenience sampling recruited ACNH players via Facebook, Reddit, and Discord communities. A three-step procedure included a pilot with 15 players (Japan, Singapore, USA, Vietnam), admin approvals in communities, and informed consent. Incentives: $5 Amazon gift cards for the first 100 respondents, $2 for the next 200. Sample: N=640 from over 30 countries; 55% from Canada/USA; 28.13% Asia, 14.38% Europe, 2.50% South America/Oceania; mean age 26.1; 64.38% female. Measures: Outcome—Perception of limited resources, measured via two items (C1, C11) from the revised New Ecological Paradigm Scale (NEPS). EFA and CFA indicated issues with the original limits-to-growth items (C1, C6, C11); C6 showed poor fit; C13 loaded empirically but lacked theoretical fit. The final outcome used the mean of C1 and C11. Predictors—Self-reported frequencies (1=never to 4=often) of in-game activities grouped into: Plant Tree; Plant Flower; Develop Resources (average of planting tree/flower); Exploit Tree (mean of taking wood, cutting down, selling for profit, reserving for bugs); Exploit Flower (mean of crossbreeding, showing off, selling for profit, sending as gift). Ethnicity (7 categories) used as a grouping factor. Analytic approach: Bayesian multilevel models (partial pooling) with ethnicity-level variation were fit using R (v4.0.2) and the rethinking package (v2.13). Priors for ethnicity effects were Normal(0,1.5). Five models predicted Perception from each predictor (Plant Tree; Plant Flower; Develop Resources; Exploit Tree; Exploit Flower), with multilevel structure by ethnicity. Model selection used WAIC; MCMC diagnostics included 4 chains × 2000 samples, effective sample sizes mostly >1000, and Rhat ≤ 1, indicating convergence. Additional diagnostics (trace plots, rank plots, pairs plots) supported well-mixed, stationary chains.

- Across models, both developing (planting) and exploiting actions generally showed positive correlations with perception of limited resources across ethnicities. - Model comparison (WAIC): Exploit Tree had the best predictive performance (WAIC=1536.1; weight=1.00), followed by Exploit Flower (WAIC=1669.9), Develop Resource (WAIC=1757.9), Plant Flower (WAIC=6450.9), and Plant Tree (WAIC=11588.7). - In the Exploit Tree model, most ethnicity-specific effects were positive, except Black or African American, which showed a negative correlation (mean = -0.27, s.d. = 0.25, 95% CI), suggesting that more tree exploitation was associated with lower perception of resource limits for this group. - Notable posterior means (Exploit Tree): Caucasian mean = 1.45 (s.d. = 0.20); Asian mean = 1.91 (s.d. = 0.40); Hispanic/Latin mean = 0.85 (s.d. = 0.41). In Exploit Flower: generally positive means (e.g., Caucasian 2.26, s.d. = 0.39; Asian 2.01, s.d. = 0.42). - Developing actions (Plant Tree/Flower, Develop Resources) also showed positive correlations, with Asian and Caucasian groups exhibiting higher means (e.g., Plant Tree: Asian 7.01, s.d. = 0.57; Caucasian 7.51, s.d. = 0.53). - MCMC diagnostics indicated convergence and reliable posterior estimation across models.

Findings indicate that ACNH players reporting higher frequencies of both planting and exploiting in-game resources tend to also report higher perceptions of real-world resource limits, with variation by ethnicity. The strongest predictive model involved tree exploitation, suggesting mechanics around extraction and regeneration may be salient for perceptions. A notable exception is the negative association for Black or African American respondents in the Exploit Tree model, warranting further investigation into cultural, experiential, or contextual factors. The broad positive associations align with literature showing virtual nature exposure and goal-driven game mechanics can influence pro-environmental attitudes and behaviors. However, the relative indifference across action types (planting versus exploiting) may reflect ACNH’s permissive mechanics (e.g., low time pressure, relatively low value of flowers), leading players toward similar behaviors regardless of perception. This points to opportunities for game design to incorporate constraints, values, and events (e.g., environmental disturbances, enhanced ecological functions and rewards for biodiversity) to better align play with pro-environmental learning outcomes and awareness. Collaboration between developers and environmental organizations could extend educational impact and reach.

The study provides early evidence of correlations between real-world perceptions of resource limits and in-game planting/exploitation behaviors in ACNH, suggesting that popular entertainment games can serve as avenues to promote pro-environmental awareness. To maximize impact, stakeholders—developers, NGOs, and educators—should integrate pro-environmental mechanics and content into game design. Future work should expand samples, include diverse communities and multiple game genres with resource management, and employ designs capable of probing causal relationships between perception and behavior.

- Sampling and generalizability: N=640 is modest relative to the global player base; convenience sampling via English-speaking online communities limits representativeness. - Self-report bias: Behaviors and perceptions were self-reported, introducing potential biases. - Measurement: NEPS limits-to-growth items required modification (dropping C6), and reliance on two items (C1, C11) may limit construct breadth. - Scope: Single-game focus (ACNH) limits external validity to other mechanics/genres. - Causality: Cross-sectional correlations cannot establish causal directions between perceptions and behaviors; longitudinal/experimental designs are needed. - Literature base: Limited theoretical and empirical foundations for this specific topic constrain interpretation.