This study employed a basic qualitative research design to explore how Indonesian children in the early childhood phase expressed pro-environmental attitudes and how they justified their responses. A qualitative design was considered appropriate because the study was concerned not only with whether children selected environmentally favourable options but also with the meanings and value orientations reflected in their verbal explanations (Aluwihare-Samaranayake 2012; Edwards 2010; Léger-Goodes et al. 2023). In this study, children’s responses were examined in relation to everyday environmental issues, namely pollution, conservation, waste and saving resources, with particular attention to whether their reasoning reflected ecocentric or anthropocentric orientations.

The analysis was guided by qualitative content analysis, which enabled the researchers to identify recurring patterns in children’s responses while interpreting them through a theoretically informed framework (Meuleman et al. 2025; Naeem et al. 2023; Villamin et al. 2025). This approach was well suited to the study’s aims because it allowed children’s brief, concrete explanations to be systematically organised, compared across environmental topics and interpreted in relation to the ecocentric–anthropocentric distinction.

The study was conducted in four early childhood education centres in Pinrang, South Sulawesi, Indonesia: TK Islam Plus E-School, TK Al Barokah, TK Polewali, and TK Negeri Pembina Kecamatan Lanrisang. These settings were selected because they provided access to children in the targeted age group and enabled the researchers to collect data from diverse early childhood education contexts within the same regional setting.

The participants were 57 children, between 5 years and 6 years of age, comprising 31 boys and 26 girls. This age group was selected because children at this developmental stage are generally able to respond to simple, picture-supported interview questions and to provide short explanations based on familiar everyday situations.

Participants were recruited using purposive sampling. Children were included if they were: (1) between 5 years and 6 years of age; (2) enrolled in one of the participating early childhood education centres; and (3) able to participate in a short one-to-one interview in Bahasa Indonesia. Children were excluded if teachers or parents indicated they were unwell on the day of data collection or were unable to engage meaningfully in the interview process.

Written informed consent was obtained from parents or legal guardians before data collection, and verbal assent was obtained from each child prior to the interview. Each child participated individually and was assigned a unique identification code (A1-A57) to protect anonymity.

Data were collected through face-to-face, picture-based interviews adapted from the Preschool Children’s Pro-Environmental Attitude Scale developed by Li et al. (2024). The original instrument was adapted for interview use because the participants were young children and could not be expected to complete a self-administered questionnaire reliably. The interview format enabled the researcher to present each scenario orally, assess the child’s understanding and invite the child to explain their choice in developmentally appropriate ways. As illustrated in Figure 1, the adapted interview covered four dimensions of pro-environmental attitudes, each further divided into specific sub-dimensions relevant to young children’s everyday environmental experiences.

The adapted interview covered four dimensions of pro-environmental attitudes: pollution, conservation, waste and saving resources. Across these dimensions, the instrument addressed 10 sub-dimensions, including air, water and noise pollution; conservation of animals, plants and soil; waste disposal and sorting and energy and paper conservation. Each sub-dimension presented the child with a pair of contrasting visual scenarios, one reflecting a more environmentally supportive option and the other a less environmentally supportive option. For example, in the waste-disposal item, children were shown one image depicting littering and another showing rubbish being placed in the appropriate bin. Children were then asked which situation they preferred and why.

The interview prompts were translated from English into Bahasa Indonesia and linguistically refined to ensure conceptual equivalence and age-appropriate wording. To support children’s comprehension of the interview scenarios, each prompt was accompanied by a visual illustration (see Figure 2). The visual materials were designed using Canva and reviewed by two early childhood education experts and two environmental education experts to assess clarity, relevance and developmental appropriateness. Based on their feedback, minor revisions were made to the wording and image presentation to improve children’s understanding of the scenarios.

Before formal data collection, the adapted prompts and illustrations were pilot-tested with five early childhood educators of a similar age who were not part of the main sample. The pilot phase was used to identify ambiguous wording, confusing visuals and interview pacing issues. Minor revisions were subsequently made to improve question clarity and reduce potential misunderstanding during the main study.

Data collection was conducted individually in a quiet and familiar space within each child’s educational setting. Each child participated in a one-to-one interview with the researcher. At the beginning of the session, rapport was established through a brief, informal conversation to help the child feel comfortable in participating.

The researcher then presented each picture pair one at a time and asked the child to indicate which situation they preferred. After the child made a choice, the researcher asked a follow-up question, such as ‘Why do you prefer that?’ to elicit the child’s reasoning.

The interviewer used simple, non-leading language and avoided evaluative responses that might influence the child’s answers. When necessary, neutral prompts such as: ‘Can you tell me more?’ or ‘Why is that?’ were used to encourage fuller responses without directing the content of the explanation. Each interview lasted approximately 10 min.

Children’s responses were documented during the interviews through handwritten field notes. The responses were then digitised and organised into a textual dataset for analysis. All identifying information was removed, and each response was linked only to the participant code.

Data were analysed using qualitative content analysis, focusing on both children’s behavioural choices and the reasons they provided for those choices. Each child responded to picture-based interview items by selecting one of two contrasting options and then explaining why that option was preferred. Accordingly, the analysis proceeded in two complementary stages.

Firstly, children’s selections were tabulated descriptively for each item to identify whether they chose the more pro-environmental or less pro-environmental option within the four dimensions of the study: pollution, conservation, waste and saving resources. These frequencies were used to describe the general direction of children’s environmental preferences across sub-dimensions.

Secondly, children’s verbal explanations were analysed interpretively using a combination of deductive and inductive qualitative content analysis. Deductively, the explanations were coded using the framework of Thompson and Barton (1994). A response was coded as ecocentric when the child’s reason referred primarily to the well-being, protection or intrinsic value of nature, animals, plants, ecosystems or the environment itself. A response was coded as anthropocentric when the child’s reason referred mainly to human-centred concerns, such as personal benefit, convenience, safety, comfort, cleanliness for people, avoidance of punishment or usefulness for human purposes.

In addition to these two theoretically derived categories, a third category, do not know, was used when a child’s explanation could not be meaningfully interpreted within either orientation. This category was assigned when the child explicitly said that they did not know, remained unable to provide a reason after prompting, gave an irrelevant answer or produced a response that was too unclear to classify reliably. Thus, the reasoning analysis distinguished among ecocentric, anthropocentric and unclassifiable responses.

After the initial coding of value orientation, the explanations were further examined inductively to identify recurring factors that appeared to shape children’s responses across cases. Through repeated reading and comparison of responses, themes emerged regarding influences such as cognitive development, religious beliefs, parenting practices, parental environmental knowledge and anthropomorphic thinking. This stage of the analysis primarily addressed the second research question, which examined factors associated with children’s pro-environmental reasoning in the Indonesian early childhood context.

All responses were organised by participant code (A1-A57), dimension and sub-dimension. The coded data were then compared across the four environmental dimensions to examine similarities and differences in children’s reasoning. This analytic procedure enabled distinguishing between children who made similar pro-environmental choices but justified them on different value bases.

Ethical clearance to conduct this study was obtained from the Ethics Committee of the Doctoral Programme in Early Childhood Education, Universitas Negeri Jakarta (No. 89/UNJ-KE/2025). Permission was also obtained from the participating schools. Written informed consent was obtained from parents or legal guardians, and verbal assent was obtained from each child before participation. Participation was voluntary. All responses were anonymised using participant codes, and the data were used only for research purposes.

The pollution dimension comprised three sub-dimensions: air pollution, water pollution and noise pollution.

As shown in Table 1, most children selected the more environmentally supportive option in all three sub-dimensions. However, the reasons behind these choices varied. Explanations in the air pollution and noise pollution sub-dimensions were predominantly anthropocentric, whereas the water pollution sub-dimension showed a markedly stronger ecocentric pattern.

As shown in Table 1, the results for the pollution dimension indicate that most children rejected burning the forest (n = 55), rejected throwing rubbish into the sea (n = 57) and rejected using a device at a high volume (n = 56). However, their explanations for these choices were more anthropocentric in air pollution (n = 44) and noise pollution (n = 55), while the water pollution item elicited mostly ecocentric explanations (n = 51).

Examples of anthropocentric reasoning included:

Examples of ecocentric reasoning included:

The conservation dimension encompasses animal, plant and soil conservation. As shown in Table 2, children again overwhelmingly selected the more environmentally supportive option. Nevertheless, the orientation of their explanations differed across the three sub-dimensions. The animal conservation item elicited mainly ecocentric reasoning, whereas plant and soil conservation elicited mainly anthropocentric explanations.

As shown in Table 2, the results for the conservation dimension indicate that most children said that they would not kill or harm animals (n = 56), chose to water or plant trees or plants (n = 54) and said they would not dispose of leftover cooking oil on the ground (n = 55). When asked to explain these choices, animal conservation showed predominantly ecocentric reasoning (n = 49), while plant conservation (n = 43) and soil conservation (n = 40) were more anthropocentric.

Examples of anthropocentric reasoning included:

Examples of ecocentric reasoning included:

The waste dimension included waste disposal and sorting. As shown in Table 3, nearly all children chose the environmentally supportive option in both sub-dimensions. Nevertheless, their explanations were overwhelmingly anthropocentric.

As shown in Table 3, the results for the waste dimension indicate that most children chose to throw rubbish in the appropriate place (n = 56) and to dispose of plastic bottles in the correct bin (n = 56). However, their reasons were mainly anthropocentric in both waste disposal (n = 45) and waste sorting (n = 54).

Examples of anthropocentric reasoning included:

Examples of ecocentric reasoning included:

The saving resources dimension comprised energy saving and paper saving. As shown in Table 4, most children selected the more environmentally supportive option in both sub-dimensions. However, the reasons underlying these choices were again predominantly anthropocentric.

As shown in Table 4, the results for the saving resources dimension indicate that most children reported turning off the lights when leaving the room (n = 51) and using both sides of the paper when writing or drawing (n = 54). However, their explanations were mainly anthropocentric in both energy saving (n = 49) and paper saving (n = 53).

Examples of anthropocentric reasoning included:

Examples of more ecocentric reasoning included:

Taken together, the results show a clear overall pattern. Across all four dimensions, children generally selected the more environmentally supportive option, indicating that pro-environmental preferences were already present at the behavioural level. However, the explanations they provided were more often anthropocentric than ecocentric. The main exceptions appeared in sub-dimensions involving animals, especially water pollution and animal conservation, where children more frequently referred to the well-being and survival of living beings.

This study has several limitations that should be acknowledged. Firstly, the study involved 57 children aged 5–6 years from four early childhood education centres in Pinrang, South Sulawesi. As the study was conducted in one regional context only, the findings should be interpreted as context specific and should not be taken as statistically representative of all Indonesian children. Secondly, the data were generated through brief picture-based interviews and documented using handwritten field notes. Although this approach was appropriate for young children, it may not have captured the full richness, nuance or spontaneity of children’s verbal reasoning. Thirdly, the factors discussed in relation to children’s responses, such as cognitive development, religious language, parenting practices, parental environmental knowledge and anthropomorphic thinking, were identified through interpretive analysis of children’s explanations rather than measured directly. Accordingly, these factors should be understood as plausible influences rather than as variables that were independently tested in this study.

Despite these limitations, the findings offer several directions for practice and future research. In educational practice, greater attention should be given not only to encouraging environmentally appropriate behaviour, but also to strengthening children’s ecocentric reasoning, especially in relation to air and noise pollution, plant and soil protection, waste management and the saving of electricity and paper. Environmental learning materials may also be improved by presenting not only animals but also other natural elements, such as trees, soil, water and energy, in concrete, meaningful and child-friendly ways. Illustrated storybooks, visual media and audio-visual learning resources may be particularly useful for helping children connect everyday actions with broader ecological consequences. For future research, studies should involve a wider range of geographical settings and social backgrounds, use audio-recorded interviews where possible and examine more directly how parental knowledge, family communication and teacher practices shape children’s environmental reasoning.