JMIR Publications

ABSTRACT

Background:

Virtual reality (VR) serious games are increasingly used for environmental education, yet applications for children are often designed without their participation. Existing studies primarily evaluate completed applications, with limited reporting on how children’s contributions are translated into design decisions or remain appropriate after implementation.

Objective:

This study aimed to (1) characterize children’s perceptions, barriers, and motivations regarding recycling; (2) establish a traceable process for translating participatory contributions into design requirements; (3) derive design implications for immersive environmental education; and (4) audit the resulting requirements through formative evaluation of an implemented prototype.

Methods:

We conducted a two-phase study during the development of Immersive Recycling VR, a serious game addressing selective waste collection. In phase 1, 18 children aged 8-13 years participated in a 3-hour workshop at a public elementary school in PetrĂłpolis, Brazil. Activities comprised a collaborative card-based recycling game, empathy mapping, and sticker-based feature prioritization. Qualitative data underwent reflexive thematic analysis, with 2 researchers independently coding the corpus and resolving divergences through negotiated agreement; quantitative data were analyzed descriptively. Workshop findings were translated into traceable design requirements. In phase 2, the baseline prototype underwent mixed-method formative evaluation with 10 technically experienced adults using cognitive walkthrough, concurrent think-aloud, structured observation, and 5 standardized instruments. This phase served as screening before evaluation with minors.

Results:

All 6 groups correctly classified 98 waste cards (100% accuracy), and all 18 children reported familiarity with recycling; however, only 3 (16.7%) reported household waste separation, while 4 (22.2%) did not know whether their households separated waste. Exploration-oriented gameplay was selected by 14 children (77.8%), whereas actions directly involving waste handling received no votes. Unlockable characters were selected by 12 (66.7%), while score-based medals or trophies received no votes. These findings generated 15 traceable design requirements. All 10 adult evaluators completed 9 representative tasks. Usability was high (System Usability Scale mean 81.5, SD 8.6), and visually induced discomfort remained low (Virtual Reality Sickness Questionnaire total 0.40 pre-exposure vs 1.10 post-exposure). Two learnability defects met the 4-source convergence criterion, both involving signaling rather than interaction technique. Perspicuity was the lowest User Experience Questionnaire dimension (+1.03). Of 15 requirements, 4 were confirmed, 1 was weakened, and 10 remained untestable because corresponding features were not implemented.

Conclusions:

Children demonstrated strong recycling classification knowledge despite limited reported household recycling practice, shifting the educational target toward the knowledge-practice gap and ambiguous materials. Their preference for exploration suggests that recycling content was not rejected, but its framing as the primary activity was. Formative evaluation showed that gradual environmental restoration may pass unnoticed under a stylized visual aesthetic. Learning outcomes and evaluation with intended child users remain to be investigated in the planned child-centered study.