Nature-based tourism has emerged as a pivotal domain within sustainable development debates due to its capacity to facilitate synergies between biodiversity conservation, local livelihoods and experiential learning. This integration of domains has the potential to generate simultaneous environmental and social benefits (Font et al., 2021; Donici and Dumitras, 2024). Ecotourism commonly defined as responsible travel to natural areas that conserves the environment, supports local people, and includes interpretation and education. Ecotourism is valued not only for its direct economic returns but also for its potential to build public environmental literacy and promote conservation behavior (Ballantyne et al., 2023; Donici and Dumitras, 2024). Recent systematic reviews underline that well-designed nature-based tourism positively affects protected area management, destination competitiveness, and community participation, but outcomes depend heavily on governance, product diversification and stakeholder education (Font et al., 2021; Donici and Dumitras, 2024).

Ecotourism encompasses a wide spectrum of activities from broadly accessible outdoor pursuits such as hiking, camping and rafting to more specialised pursuits such as birdwatching, botanical tourism, farm-based experiences and orienteering. Contemporary studies highlights that this diversity is important. Diversified offerings can broaden market reach, deepen visitor engagement with nature, and create differentiated livelihood opportunities for host communities; conversely, insufficient diversification or narrow activity portfolios may reinforce seasonality and uneven economic benefits (Samora-Arvela et al., 2020; Donici and Dumitras, 2024). Empirical work also shows that different activity types elicit different conservation attitudes and educational outcomes. For instance, interpretive birdwatching and botanical tours tend to foster species-specific knowledge and stewardship intentions more effectively than generic adventure activities (Ballantyne et al., 2023; Ren J. et al., 2021).

The modern ecotourism agenda has matured over several decades. Although early environmental awareness movements in the 1960s–1970s seeded interest in conservation tourism, international policy attention and scholarly consolidation accelerated through the 1980s and 1990s as ecotourism became framed explicitly within sustainable development discourses (Gössling et al., 2020). Since the 2000s, academic output and destination management efforts have expanded considerably, prompting more rigorous evaluations of ecotourism’s economic, social and ecological impacts (Font et al., 2021; Donici and Dumitras, 2024).

Türkiye’s climatic heterogeneity and biogeographical position confer high biodiversity and a rich portfolio of natural and cultural assets that are attractive for ecotourism development (Gössling et al., 2020; Kiper et al., 2022). Düzce province exemplifies a locality with varied potential including mountain plateaus, karstic caves, rivers suitable for white-water activities, coastal diving sites, and a network of trails and highlands that together create a multi-modal nature-tourism product. However, converting natural assets into sustainable ecotourism outcomes requires not only infrastructure and governance but also human capital (especially stakeholder groups who mediate conservation messages) and shape young people’s attitudes toward nature (Gültekin, 2022).

Education (particularly formal education delivered through schools) is a critical leverage point for long-term sustainability. A growing body of evidence shows that experiential and values-based interpretation embedded in tourism and complemented by school-based environmental education can increase pro-conservation attitudes and translate into more environmentally responsible visitor behavior (Ballantyne et al., 2023; Ren J. et al., 2021; Ren L. et al., 2021). Teachers function as both transmitters and interpreters of environmental knowledge. Their subject knowledge, personal engagement with nature-based activities, and pedagogical practices influence how sustainability concepts are delivered to successive cohorts of students (Rodríguez Pérez, 2024). Consequently, teachers’ own perceptions, experiences and preferences regarding ecotourism become relevant policy levers. Teachers can shape curricular emphasis, extracurricular programming (e.g., field trips, nature clubs, etc.), and the broader community discourse around nature-based development.

Despite the expanding literature on ecotourism, sustainability education, and forest governance, empirical research that explicitly conceptualizes and examines teachers as influential actors in the formation of sustainable ecotourism perceptions and governance processes remains notably limited. While recent studies have explored local residents’ attitudes and stakeholder perspectives on ecotourism in Türkiye and elsewhere (Upadhaya et al., 2022; Donici and Dumitras, 2024), comparatively few studies have systematically investigated how elementary and secondary school teachers conceptualize ecotourism, which activities they prioritise, and whether perceptual structures differ across teaching levels. This discrepancy underscores the necessity for tailored professional development and policy responses, given the heterogeneity observed between elementary and secondary educators in environmental literacy or activity engagement.

Against the backdrop of increasing emphasis on sustainable forest management, nature-based tourism, and education for sustainability, this study seeks to address an important empirical and conceptual gap concerning the role of teachers in ecotourism-related perception formation. While existing research has extensively examined ecotourism from the perspectives of tourists, local communities, and tourism operators, comparatively limited attention has been paid to teachers as socially embedded actors who mediate environmental values, learning processes, and long-term stewardship orientations.

Accordingly, the primary objective of this study is to examine the structure and determinants of ecotourism perceptions among primary and secondary school teachers in Düzce, a forest-rich region with significant ecotourism potential. Specifically, the study aims to (i) identify the underlying dimensions of teachers’ ecotourism perceptions, (ii) test the robustness and measurement equivalence of these dimensions across key social and professional groups, and (iii) examine whether the structural relationships among ecotourism-related constructs differ by gender, marital status, and teaching level using a multiple group structural equation modelling (MGSEM) approach.

Beyond its empirical aims, the study makes several original contributions to the literature. First, it conceptually positions teachers as intermediary actors within forest–society systems, thereby extending ecotourism and forest governance research beyond conventional stakeholder groups. Second, by establishing full measurement invariance and clarifying the logic of multigroup structural comparisons, the study contributes methodologically to the rigorous application of MGSEM in sustainability and environmental education research. Third, the findings offer policy-relevant insights by demonstrating how teachers’ ecotourism perceptions can inform interpretation planning, forest-based education initiatives, and participatory governance mechanisms in nature-based destinations.

Through these contributions, the study advances an integrated perspective that links ecotourism, education, and sustainable forest management, highlighting the importance of social learning and perception formation in addressing contemporary forest governance challenges under conditions of global environmental change.

The research area selected for this study is the province of Düzce and its surrounding region, chosen due to its significant potential for nature-based and ecotourism activities. The study population and primary material consist of teachers employed in schools affiliated with the Düzce Provincial Directorate of National Education (DPDNE). According to the DPDNE’s data, there were 5,437 teachers working in its affiliated schools at the time of the study.

An online questionnaire was administered to teachers in DPDNE. After obtaining official permission from the DPDNE, schools were visited and support was obtained from school administrations to conduct surveys with teachers. Teachers were invited via official provincial education mailing lists and online platforms. No stratified sampling was applied. Participation was voluntary. The instrument was deployed only after the research protocol received ethical approval from the Scientific Research and Publication Ethics Committee. Data collection occurred between February and May 2024. The number of surveys has been increased by sending participants a monthly reminder email. Teachers who did not respond were not replaced by other participants. Only follow-up reminders were issued to those who did not respond. By the end of the collection period, 320 valid responses were obtained. Approximate response rate is 5.9%. The average response rate is consistent with the expected participation level for this survey population (Bethlehem, 2010; Groves et al., 2011).

The questionnaire instrument was designed drawing on previous ecotourism perception related studies, and it followed typical online survey procedures (Gültekin and Gültekin, 2017; Sakıcı and Çelik, 2017; Chung-Shing et al., 2020; Gültekin, 2022; Huang et al., 2023) to ensure validity and reliability of responses. The translation and back-translation procedure followed established guidelines for cross-cultural scale adaptation (Brislin, 1970; Beaton et al., 2000). Content validity was assessed through expert evaluation in line with recommended practices (Lynn, 1986; Polit et al., 2007), while subsequent measurement analyses adhered to SEM quality criteria (Hair et al., 2019; Kline, 2011).

The survey instrument was adapted to the local context through a systematic translation and validation procedure to ensure linguistic accuracy, conceptual equivalence, and content validity. Initially, the original questionnaire items were translated into Turkish by two independent bilingual translators with expertise in environmental studies and educational research. Both translators were native Turkish speakers fluent in English and familiar with the terminology of ecotourism and sustainability.

Subsequently, the two translated versions were compared and synthesized into a single preliminary Turkish version through consensus, with particular attention paid to semantic consistency and contextual appropriateness. This version was then subjected to a back-translation process. Two different bilingual experts, who had not been involved in the initial translation stage and were blind to the original instrument, independently translated the Turkish version back into English. The back-translated versions were compared with the original English questionnaire to identify discrepancies, ambiguities, or potential losses in meaning.

Following the translation and back-translation stages, expert validation was conducted to assess content relevance, clarity, and cultural appropriateness. An expert panel consisting of five academics specializing in ecotourism, environmental education, tourism management, and measurement and evaluation reviewed the instrument. Experts were asked to evaluate each item in terms of conceptual adequacy, representativeness of the construct, and linguistic clarity, using a structured feedback form. Based on the experts’ recommendations, minor revisions were made to item wording to enhance clarity and ensure alignment with the Turkish educational and cultural context. This multi-step procedure followed established methodological guidelines for cross-cultural scale adaptation and contributed to the content validity and cross-language equivalence of the measurement instrument used in the study.

The survey comprised three main sections: (1) demographic and professional background of respondents (e.g., gender, age, teaching level); (2) two items assessing teachers’ self-reported knowledge of ecotourism and (3) a 17-item perceptual scale using a 5-point Likert format (1 = strongly disagree to 5 = strongly agree) to measure teachers’ perceptions of ecotourism. The online administration followed best-practices in web-survey design (Reips and Neuhaus, 2009) to reduce measurement error and non-response bias.

Prior to analysis, the dataset was screened for missing data and response completeness following established survey research guidelines (Kline, 2011; Hair et al., 2019). Given that missing values were minimal and randomly distributed, complete-case analysis was applied without imputation, consistent with recommended practices (Little and Rubin, 2019; Schafer and Graham, 2002).

For data analysis, a sequence of statistical procedures was employed: descriptive statistics to summarize respondent characteristics, exploratory factor analysis (EFA), confirmatory factor analysis (CFA) and MGSEM to identify the underlying perceptual dimensions (Kline, 2011; Hair et al., 2019). Prior to parametric testing, distributions of key variables (list of items) were evaluated for normality via skewness and kurtosis coefficients (Table 1). The obtained values met the commonly accepted thresholds (±3) and thus parametric methods were deemed appropriate (Ryu and West, 2009). In line with best practices in SEM research, results are reported using standardized path coefficients (β) alongside statistical significance levels. The magnitude of these standardized coefficients is interpreted as an indicator of effect size, allowing assessment of the substantive strength of estimated relationships beyond statistical significance alone. All analyses were performed using the IBM SPSS Statistics 22 and AMOS 22 software packages.

Multicollinearity was assessed at the latent construct level to measurement model loadings in covariance-based SEM. Accordingly, correlations among latent constructs were examined, and all correlations remained below conservative thresholds (r < 0.85), indicating the absence of problematic multicollinearity and supporting the distinctiveness of the modeled constructs (Kline, 2011; Hair et al., 2019).

Common Method Bias (CMB) is a particular concern in behavioral and perception research where self-reported measures are collected from the same respondents at a single point in time, as it can artificially inflate or deflate the observed relationships among variables (Podsakoff et al., 2003, 2012). To ensure the integrity of the study’s findings, both procedural and statistical remedies were employed.

Consistent with best practices in survey design (Podsakoff et al., 2003, 2012; Fuller et al., 2016; Kock et al., 2021), the following steps were implemented to minimize potential CMB:

These ex-ante procedural safeguards are widely recommended to reduce the likelihood of CMB prior to data collection (Fuller et al., 2016; Podsakoff et al., 2003). To complement procedural remedies, post-hoc statistical tests were conducted to empirically assess the potential influence of CMB on the data.

Harman’s single-factor test was performed using exploratory factor analysis (EFA) on all measurement items without rotation. Under this approach, CMB is considered problematic if a single factor accounts for the majority of the total variance (Podsakoff et al., 2003, 2012). In the present analysis, the first unrotated factor explained well below the conservative threshold, indicating that no single factor dominated the covariance structure. The threshold of concern is often cited as greater than 40–50% of variance explained by the first factor (Podsakoff et al., 2003; Kock et al., 2021). In this study, the first factor’s contribution was below this level, suggesting that CMB does not materially bias the data.

To further assess CMB beyond the limitations of Harman’s test (Podsakoff et al., 2003), a Common Latent Factor test was employed within the confirmatory measurement modeling framework. This technique introduces an unmeasured latent factor that loads on all observed indicators to capture any shared method variance. The results indicated no substantial change in key parameter estimates when the common latent factor was included, further supporting the conclusion that common method variance did not meaningfully distort the substantive relationships among constructs (Fuller et al., 2016; Kock et al., 2021).

Together, the procedural and statistical assessments provide convergent evidence that common method bias is not a substantive threat to the validity of the present findings. This comprehensive approach aligns with contemporary methodological recommendations in behavioral and tourism research, where mixed prevention and diagnostic strategies are advocated to enhance the credibility of self-report survey results (Fuller et al., 2016; Podsakoff et al., 2003; Kock et al., 2021).

To assess the potential influence of common method bias, both procedural and statistical diagnostics were applied. First, Harman’s single-factor test indicated that the first unrotated factor accounted for less than 40% of the total variance (32.7%), suggesting that no single factor dominated the covariance structure of the measurement items. Second, a common latent factor was introduced into the measurement model to capture any shared method variance. The inclusion of this factor did not result in substantive changes in standardized factor loadings, structural path estimates, or overall model fit indices (ΔCFI and ΔRMSEA < 0.01). Collectively, these results indicate that common method bias is unlikely to pose a significant threat to the validity of the findings.

To ensure the validity of MGSEM comparisons, measurement invariance was tested sequentially across gender and marital status groups. Configural, metric, and scalar invariance models were estimated using ΔCFI ≤ 0.01, ΔTLI ≤ 0.01 and ΔRMSEA ≤ 0.01 as decision criteria. Results supported metric invariance and partial scalar invariance, permitting meaningful comparison of structural path coefficients across groups (Cheung and Rensvold, 2002; Chen, 2007; Putnick and Bornstein, 2016). The goodness-of-fit cut-off criteria were set at 0.90 for the CFI and TLI and 0.08 for the RMSEA to indicate acceptable model fit, while values of 0.95 for the CFI and TLI and 0.06 for the RMSEA were considered indicative of excellent fit (Hu and Bentler, 1999).

Several procedural and statistical remedies were applied to mitigate common method bias. Procedurally, respondent anonymity was ensured and item wording was refined. Statistically, Harman’s single-factor test indicated that a single factor did not account for the majority of variance, and a common latent factor test revealed no substantive changes in path coefficients (Cheung and Rensvold, 2002; Chen, 2007; Putnick and Bornstein, 2016).

The structural models estimated in this study did not incorporate control variables such as age, years of professional experience, or self-reported knowledge level. This modeling decision was guided by the study’s primary theoretical objective, which was to examine whether the strength of structural relationships among ecotourism-related latent constructs varies across social and professional groups. In multigroup structural equation modeling, introducing multiple covariates may obscure theoretically meaningful relational mechanisms and complicate the interpretation of group-based differences in path coefficients, particularly when the focus is on comparative structural dynamics rather than causal estimation of net effects. Methodological literature emphasizes that control variables should be included only when they are theoretically central to the research question and do not undermine model parsimony or interpretability. Nevertheless, the exclusion of control variables is acknowledged as a methodological limitation. Future research may extend the present framework by incorporating covariates to test the robustness of structural relationships and to refine causal inferences regarding ecotourism perception processes. This modeling decision aligns with recommendations emphasizing parsimony and theoretical clarity in multigroup SEM (Kline, 2011; Cheung and Lau, 2011) and cautions against the uncritical inclusion of control variables that may obscure substantive relationships (Spector and Brannick, 2010; Bernerth and Aguinis, 2016).

This study contributes to the literature on ecotourism, environmental education, and forest governance by providing an integrated empirical examination of teachers’ ecotourism perceptions as a multidimensional and structurally robust construct. The findings demonstrate that teachers’ perceptions are coherently organized around a higher-order conceptual framework and operate consistently across key social and professional groups, as evidenced by full measurement invariance. Beyond documenting perception levels, the study advances theoretical understanding by showing that the relationships among ecotourism-related dimensions vary across groups, thereby highlighting the relevance of relational mechanisms rather than solely mean-level differences. Conceptually, these results position teachers as intermediary actors within forest–society systems who contribute to social learning, interpretation processes, and the long-term legitimacy of sustainable forest and ecotourism governance. By linking perception structures to governance-relevant mechanisms, the study extends existing ecotourism research and underscores the importance of education-based pathways in sustainability transitions under conditions of global environmental change.

The sample characteristics of the 320 teacher respondents are summarized in Table 2. The findings indicate that the participant teachers span a variety of age brackets, represent different institutional types (pre-school, elementary, secondary and high school settings), and report diverse levels of self-assessed knowledge regarding ecotourism. The presence of considerable heterogeneity in terms of age, professional context and knowledge is to be recommended, as recent research highlights the significant impact of educators’ demographic and institutional backgrounds on their engagement with sustainability and environmental education programmes (Zhou et al., 2024; Amézquita-Galindo and Losada Salgado, 2025). For instance, older teachers may draw on lived environmental experience while younger teachers may be more current with sustainability pedagogy, thereby contributing to richer variation in attitudes and perceptions.

Simultaneously, the sample shows relatively balanced distributions across gender and school type (institutional setting) in Düzce. The gender balance and institutional variety present here align with methodological guidance that a well-distributed sample enhances the robustness of subgroup comparisons and helps mitigate sampling bias (Colón-Flores et al., 2023). Balanced demographic representation supports the reliability of subsequent analyses such as factor analysis by reducing the risk of skewed subsamples dominating responses (Il’ko et al., 2025).

Altogether, the sample appears sufficiently diverse to explore variation in teachers’ perceptions of ecotourism while maintaining enough homogeneity to allow credible comparisons between subgroups. This foundational sample profile enables meaningful discussion of how knowledge, institutional context, and demographic variables may interact to shape ecotourism perceptions among teachers in Düzce. In the broader context of environmental education research, this sample structure is consistent with studies showing that both individual (age, teaching level, knowledge) and contextual (school setting, region of residence) factors drive teachers’ attitudes toward sustainability and nature-based tourism (Tuncer, 2009; Nyberg, 2020; Nguyen and Nguyen, 2023). The fact that participants came from different school levels and had balanced characteristics such as gender, age and marital status indicates that the survey was not subject to non-response bias or self-selection limitations (Bethlehem, 2010). Therefore, the results offer a credible empirical basis for further analysis of how these characteristics correlate with ecotourism perceptions, and how educational initiatives could be tailored for different teacher groups.

Exploratory factor analysis (EFA) was conducted to examine the underlying structure of the ecotourism perception scale (Table 3). Prior to extraction, sampling adequacy was assessed. The sample size (N = 320) met the recommended subject-to-item ratios for maximum-likelihood factor analysis (Costello and Osborne, 2005; Watkins, 2018). The Kaiser–Meyer–Olkin (KMO) value was 0.911, indicating excellent adequacy (Kaiser, 1974), and Bartlett’s Test of Sphericity was significant (χ²(190) = 4126.502, p < 0.001).

Using Maximum-Likelihood extraction and Promax rotation, a three-factor solution emerged, explaining 63.173% of the total variance surpassing the conventional 60% threshold (Hair et al., 2019). The factors were interpreted as Economic Benefits (ECOB), The Sustainability of Ecotourism (ECOS), and Environmental Effects (ECOE), consistent with contemporary ecotourism perception frameworks (Baral et al., 2012; Üzülmez et al., 2023). Item loadings ranged from 0.668 to 0.888 across the three components, indicating strong and clean factor structure (Tabachnick and Fidell, 2019). Reliability analysis showed high internal consistency for the overall scale (Cronbach’s α = 0.91) and satisfactory reliability for each factor (ECOB, Cronbach’s α = 0.935; CR = 0.943; AVE = 0.602; ECOS, Cronbach’s α = 0.731; CR = 0.846; AVE = 0.647; ECOE, Cronbach’s α = 0.723; CR = 0.827; AVE = 0.614), meeting the recommended threshold of α and CR ≥ 0.70 and AVE ≥ 0.50 (Nunnally and Bernstein, 1994; Kline, 2011). These results confirm that the three extracted dimensions form a psychometrically robust basis for subsequent CFA and SEM.

The second-order CFA conducted to reveal ecotourism perception (ECOP) of teachers (Figure 2). Overall model fit was acceptable: χ²/df = 2.67, CFI = 0.952, TLI = 0.941, SRMR = 0.048, and RMSEA = 0.072 (90% CI: 0.062–0.083). The RMSEA meets the widely used <0.08 threshold for moderate fit in complex models. The second-order structural model revealed that ECOE was by far the strongest predictor of overall ecotourism perception (ECOP). The standardized loading of ECOE on ECOP was 0.980, indicating a near-perfect association. Teachers who evaluated ecotourism as environmentally beneficial overwhelmingly rated ecotourism positively overall. “H₀₁: ECOP are positively associated with ECOE.” is supported by the data. In contrast, ECOB showed a moderate but meaningful influence on ECOP (β = 0.563). This suggests that economic, social, promotional, and well-being benefits matter but do not dominate teachers’ evaluations to the same extent as environmental outcomes. Therefore, “H₀₂: ECOP are positively associated with ECOB” is supported by the data. ECOS contributed modestly to ECOP (β = 0.304). Teachers appear less influenced by institutional sustainability mechanisms (monitoring, governance, environmentally responsible facilities) when forming their general ecotourism perceptions. These findings collectively indicate that teachers’ perceptions of ecotourism are environmentally driven, benefit-supported, and only weakly shaped by governance trust. Although the beta coefficient was average, “H₀₃: ECOP are positively associated with ECOS.” is confirmed by the data.

The confirmatory factor analysis (CFA) supported the three-factor structure underlying teachers’ ecotourism perceptions. All items loaded significantly (p < 0.001) on their respective latent constructs. The ECOE factor presented the strongest standardized loadings (λ = 0.839–0.936), indicating that teachers interpreted these items “ecotourism keeps the environment cleaner,” “supports nature conservation,” and “preserves historical and cultural assets” as highly coherent indicators. This is consistent with recent studies showing that stakeholders tend to evaluate ecotourism primarily through its perceived environmental performance and conservation credibility (Azinuddin et al., 2022; Gültekin, 2022).

The ECOB factor also demonstrated good reliability, with loadings ranging from 0.666 to 0.881. Items related to economic development (ECOB1, ECOB3, ECOB5, ECOB8), regional promotion (ECOB2), quality of life (ECOB10), well-being (ECOB6), and social/cultural interaction (ECOB7, ECOB9, ECOB11) all significantly contributed to the latent construct. This reflects the multi-dimensionality of benefits associated with ecotourism, which contemporary research conceptualizes as economic-social-psychological value bundles that influence local and visitor attitudes (Darda and Bhuiyan, 2022). Teachers, as local opinion leaders and transmitters of cultural knowledge, are well-placed to nurture the cultural revitalization and values the authors identify as decisive (Purnamawati et al., 2022). Teachers can integrate place-based cultural education, service-learning, and community partnerships to reinforce the cultural practices that support sustainable ecotourism.

The ECOS factor is represented the trust in institutions, monitoring, and environmental responsibility of tourism facilities yielded acceptable loadings (0.694–0.854). This suggests teachers perceive ECOS as somewhat coherent but not as salient as ECOE or ECOB. Similar findings are reported in regions where institutional capacity or transparency is perceived as limited, diminishing the explanatory power of governance-related indicators (Ren L. et al., 2021). Robust policies and active social-media engagement can positively affect on sustainable ecotourism development (Kansra et al., 2024). Education institutes can accelerate sustainable development through these activities.

Structural path estimates are reported using standardized coefficients (β) to facilitate interpretation of effect magnitudes across groups, consistent with established SEM reporting conventions (Kline, 2011; Hair et al., 2019). Beyond statistical significance, attention is given to the relative strength of relationships to identify substantively meaningful differences across gender, marital status, and teaching level groups, in line with contemporary MGSEM practices (Cheung and Lau, 2011; Preacher and Kelley, 2011).

According to the MGSEM result, “H₀₄: The ecotourism perception of teachers is different between females and males.” is supported by the data. In 2 different tests of the SEM model (ECOP → ECOB) between females and males, it is found that the p value is significant (p < 0.01) with 95% CIs (females, 0.59; males, 0.57). The positive effect of ECOP on ECOB is stronger for females than for males. In other words, the structural relationships between ecotourism perception dimensions (ECOP → ECOB) differ significantly between male and female teachers. This finding is consistent with a robust body of literature reporting stronger environmental concern and pro-environmental orientations among women, which is often attributed to gender socialization and an ethic-of-care orientation (Xiao and McCright, 2015; Zhao et al., 2021). Although the female–male gap in the ECOP → ECOB path is modest in absolute terms, it is statistically reliable in our sample.

Hypothesis 4 posits that female teachers are more aware of the ecotourism benefits. Similarly, the hypothesis “H₀₅: Teachers’ perception of ecotourism is different between married individuals and single individuals” has also been confirmed by the data. Accordingly, the difference tests conducted for 2 latent variables of the model (ECOP → ECOB, ECOP → ECOS) between married individuals and singles found that the p-value is significant (p < 0.01) with 95% CIs (ECOP → ECOB: married individuals, 0.57; singles, 0.49; ECOP → ECOS: married individuals, 0.29; singles, 0.25). The structural relationships between ecotourism perception dimensions differ significantly between married and single teachers. Prior resident-perception studies find married respondents place greater weight on the economic and socio-cultural benefits of tourism, likely because of household and family stakes (Timothy and Said, 2023; Uslu et al., 2023). We note that for valid interpretation of group differences it is essential to establish measurement invariance. Future work should confirm metric/scalar invariance and examine potential mediators (e.g., household dependence on tourism, environmental values) that may explain these demographic moderation effects.

Finally, the hypothesis “H₀₆: Teachers’ perception of ecotourism differs between teaching levels” was supported by the data. Specifically, the path from Ecotourism Perception (ECOP) to Ecotourism’s Economic Benefits (ECOB) was significantly stronger among elementary school teachers (β = 0.90) compared to secondary school teachers (β = 0.30) (p < 0.01, 95% CI). Likewise, the path from ECOP to Ecotourism’s Sociocultural Outcomes (ECOS) was slightly but significantly higher for elementary teachers (β = 0.28) than for secondary teachers (β = 0.26) (p < 0.01). The structural relationships between ecotourism perception dimensions differ significantly between primary and secondary school teachers. No significant differences were observed for other school levels, indicating that the elementary–secondary distinction captures the main variance in teachers’ ecotourism-related perceptions.

In the present study, hypotheses H04–H06 are explicitly operationalized as tests of group differences in structural path coefficients, rather than comparisons of latent mean levels. This analytical focus is consistent with current methodological standards in MGSEM, which distinguish between structural invariance tests that assess whether relationships among latent constructs vary across groups and latent mean comparisons that evaluate whether the average level of a construct differs between groups (Meredith, 1993; Cheah et al., 2023). Specifically, testing equality constraints on regression paths across groups enables examination of moderation in the strength and direction of hypothesized relationships—a core objective of relational analyses in MGSEM contexts—without conflating these with differences in absolute levels of latent constructs (Cheah et al., 2023; Perez Alonso et al., 2024). By contrast, latent mean comparisons require additional identification and invariance constraints (particularly scalar invariance) and are typically applied when the substantive interest lies in comparing levels of latent traits rather than the patterns of structural relations themselves (Dimitrov, 2006; Geiser et al., 2014). Because the primary aim of the current study is to investigate whether the strength of associations among ecotourism perceptions differs across gender, marital status, and teaching level—not whether groups differ in overall perception levels—latent mean comparisons were not undertaken.

A MGSEM assessed structural invariance across gender, marital status and teaching levels. Configural, metric, and scalar invariance were assessed across groups in Table 4. Partial scalar invariance was established by freeing two intercepts, allowing meaningful comparison of structural paths. Configural invariance was supported for all groups. Metric invariance was achieved (ΔCFI < 0.01). Establishing (partial) scalar invariance validates the comparison of structural coefficients, strengthening confidence in the MGSEM conclusions (see Table 5).

Multi-group measurement invariance was tested across gender, marital status, and teaching level using a sequential invariance framework. For all grouping variables, the configural models demonstrated acceptable model fit, indicating that the underlying factor structure of ecotourism perceptions was equivalent across groups. Subsequent comparisons showed that constraining factor loadings (metric invariance) did not result in meaningful deterioration in model fit, as changes in CFI, TLI, and RMSEA remained within recommended thresholds (Δ ≤ 0.01). The scalar invariance was also supported across all groups, suggesting equivalence of both factor loadings and item intercepts. These results indicate that teachers across different gender categories, marital status groups, and teaching levels interpret the ecotourism perception constructs in a comparable manner. Consequently, latent mean comparisons and structural path comparisons across groups are methodologically justified.

We tested alternative SEM models (correlated three-factor model, bifactor and constrained models) beside second-order DFA to obtain better results. Correlated three-factor model had better results. Correlated three-factor model fit was acceptable: χ²/df = 2.67, CFI = 0.952, TLI = 0.941, SRMR = 0.041, and RMSEA = 0.072 (90% CI: 0.062–0.083). Correlated three-factor model presented in Figure 3. While the correlated model showed similar fit, the second-order model was retained for theoretical parsimony.

The relatively high standardized loadings observed in the second-order measurement model should be interpreted not as a statistical deficiency, but as evidence of conceptual overlap and theoretical coherence among the first-order dimensions of ecotourism perceptions. In higher-order structural equation models, strong loadings indicate that the lower-order constructs share substantial common variance and are meaningfully organized by an overarching latent factor (Marsh and Hocevar, 1985; Brown, 2015). In the present study, this pattern suggests that the identified dimensions of ecotourism perceptions are not independent attitudinal fragments, but rather interrelated expressions of a broader, unified sustainability-oriented worldview.

From a conceptual standpoint, ecotourism inherently integrates environmental, social, and educational components. High second-order loadings therefore reflect the integrated nature of ecotourism as a social–ecological construct, particularly within educational contexts where conservation ethics, recreational use, and community responsibility are closely intertwined. This interpretation is consistent with contemporary research emphasizing that sustainability-related perceptions often exhibit strong internal coherence due to their normative and value-based foundations (Milfont and Duckitt, 2010; Mochizuki and Yarime, 2015).

Methodologically, high loadings in second-order models are expected when first-order factors are theoretically proximal and empirically correlated, provided that model fit indices remain within acceptable thresholds and discriminant validity is supported at the first-order level (Chen et al., 2005; Hair et al., 2019). In this study, the satisfactory global fit indices and the establishment of measurement invariance across multiple groups indicate that the higher-order structure is stable and interpretable rather than over-specified. Thus, the second-order factor functions as a parsimonious representation of teachers’ overarching ecotourism orientation, capturing shared variance without obscuring the distinctiveness of the underlying dimensions.

Importantly, conceptualizing high second-order loadings as indicative of theoretical integration aligns with forest governance and sustainability science literature, which increasingly frames environmental perceptions as holistic belief systems rather than discrete attitudes (Art and Visseren-Hamdkers, 2012; Turnhout et al., 2020). Within this perspective, the strong convergence of first-order ecotourism dimensions reflects how educators cognitively and normatively integrate forest conservation, nature-based tourism, and educational responsibility into a coherent framework that informs behavior and policy engagement.

Accordingly, the second-order model enhances the explanatory value of the analysis by capturing the systemic nature of ecotourism perceptions, rather than fragmenting them into artificially isolated components. This supports the use of higher-order modeling as both a theoretically grounded and methodologically robust approach for examining complex social–ecological constructs in forest and sustainability research.

These findings suggest that elementary school teachers translate their ecotourism perceptions more strongly into beliefs about local economic and social benefits, whereas secondary teachers maintain a weaker or more abstract connection between perception and outcome. One explanation lies in the pedagogical structure and community embeddedness of elementary education, which emphasizes experiential learning, environmental awareness, and direct engagement with local contexts. Prior studies report that elementary teachers often integrate outdoor and place-based learning, making them more receptive to linking ecotourism with local development (Liao et al., 2022). Similar evidence from intervention studies highlights that educational level and curriculum design shape attitudes toward sustainability and tourism-related entrepreneurship (Ambusaidi et al., 2024).

Cultural and contextual factors may also underpin these differences. As Purnamawati et al. (2022) demonstrate, cultural values and community participation play a pivotal role in shaping sustainable ecotourism development. Teachers at the elementary level—often deeply involved in parent–community networks—may therefore perceive ecotourism as a tangible driver of local economic resilience and cultural continuity. Cultural and contextual factors need to integrate local, regional and national forest management strategies. Thus sustainable forest governance can be enhanced.

The significance of these inter-group differences presumes that measurement invariance across teaching levels has been adequately tested, ensuring that observed variations reflect true structural divergence rather than measurement bias (Cheung and Rensvold, 2002; Vandenberg and Lance, 2000). The results underscore the need for differentiated teacher training programs: while elementary teachers can serve as community catalysts for sustainable tourism awareness, secondary teachers may benefit from curriculum-based initiatives that link abstract sustainability concepts with real-world ecotourism practices.

Kil et al. (2014) validates our use of latent-path modelling to disentangle how environmental attitudes and motivations jointly drive pro-ecotourism outcomes, reinforcing the robustness of the ECOP → ECOE/ECOB pathways we report in their SEM work. Høyem’s (2020) qualitative finding that reflection underpins environmentally responsible behavior helps explain why environmental beliefs (ECOE) dominate teachers’ overall perceptions. Teachers’ reflective capacities likely amplify the salience of ecological outcomes. Ajuhari et al. (2023) show that attachment and experiential benefits shape visitor motivations, which nuances our ECOB interpretation by suggesting that teachers’ life-stage and community ties (e.g., married status, elementary-level embedding) strengthen benefit-oriented appraisals. Arnegger et al. (2024) remind us that media and destination narratives frame stakeholder perceptions, indicating that coordinated messaging could augment teachers’ trust in governance (ECOS). Powers and Trauntvein (2024) demonstrate that interpretive learning mediates the recreation → citizenship pathway, supporting our policy recommendation to pair experiential outdoor programs with reflective, interpretive pedagogy in teacher training. Together, these studies corroborate our core findings and point to integrated interventions combining curricula, community attachment, and destination communication.

The results indicate that teachers’ perceptions of the environmental dimension of ecotourism constitute the principal determinant shaping their evaluations of associated economic and socio-cultural benefits, supporting sustainability research that identifies environmental value orientations as a core cognitive lens through which development outcomes are interpreted (Weaver, 2015; Gössling et al., 2009). Moreover, the MGSEM analyses reveal that the strength of these perception–benefit relationships varies systematically across gender and marital status, underscoring the moderating role of socio-demographic characteristics in the formation of sustainability-related attitudes. This pattern is consistent with empirical evidence showing that gendered socialization processes influence environmental concern and benefit attribution, while life-course factors such as marital status shape long-term orientations toward community welfare and resource stewardship (Xiao and McCright, 2015). Collectively, these findings highlight teachers as educational and social intermediaries who play a critical role in transmitting environmental values and shaping long-term awareness of ecotourism and forest governance. Through their dual influence on formal education and broader community discourse, teachers represent a strategically important yet underexamined actor group in advancing sustainable tourism practices and fostering public support for environmentally responsible forest management (Ballantyne and Packer, 2011; Gössling et al., 2009).

The results indicate that teachers in Düzce exhibit relatively high levels of engagement with accessible and experience-oriented nature-based activities, such as hiking, camping, and rafting, whereas participation in more knowledge-intensive forms of ecotourism (including birdwatching and botanical tourism) remains comparatively limited. From a forest governance perspective, this imbalance reflects a broader global challenge, whereby recreational use of forest landscapes tends to expand more rapidly than ecological understanding unless accompanied by targeted educational and interpretive interventions (Buckley, 2022). This pattern underscores the importance of strengthening interpretive and educational components within ecotourism strategies to ensure that increased visitation contributes to conservation-oriented outcomes rather than merely recreational consumption.

The establishment of full measurement invariance across gender, marital status, and teaching level further reinforces the robustness of these findings, indicating that ecotourism perceptions operate as structurally comparable and shared social constructs within the education sector. This result is consistent with forest governance scholarship emphasizing that shared norms, values, and interpretative frameworks constitute essential foundations for collective action, policy coherence, and the social legitimacy of forest management interventions (Art and Visseren-Hamdkers, 2012).

The structural model is interpreted by emphasizing effect magnitudes alongside statistical significance through the use of standardized path coefficients, enabling substantive comparison across groups. Although full measurement invariance indicates a shared conceptual understanding of ecotourism among teachers, variations in structural path strengths across gender, marital status, and teaching levels reveal differentiated relational dynamics. This finding underscores the analytical value of examining group differences in relationships rather than latent mean levels alone. The results suggest that social and professional positioning within the education system shapes how ecotourism-related knowledge and attitudes translate into behavioral and governance-relevant outcomes. By foregrounding relational mechanisms and effect sizes, the study advances understanding of teachers as intermediary actors linking environmental education, ecotourism development, and sustainable forest governance in the context of global environmental change.

Nevertheless, the absence of control variables is acknowledged as a methodological limitation. Future research may extend the present framework by incorporating theoretically relevant covariates to test the robustness of the observed structural relationships and to refine causal explanations of ecotourism perception processes.

Taken together, the findings suggest that teachers represent a scalable and institutionally embedded intervention point for enhancing forest-related sustainability outcomes. By strengthening teachers’ engagement with specialized forms of ecotourism and forest-based education, policymakers and forest authorities may indirectly influence visitor behavior, improve compliance with conservation measures, and foster long-term stewardship orientations. Such socially mediated pathways are increasingly critical under conditions of accelerating global environmental change, where adaptive forest governance depends on sustained social learning and value alignment across generations (see Table 6).

This study examined the differential pathways shaping teachers’ perceptions of sustainable ecotourism in Düzce, Türkiye, employing a Multiple Group Structural Equation Modelling (MGSEM) approach. The findings revealed that teachers’ perceptions of ecotourism (ECOP model) are most strongly influenced by environmental effects (ECOE), moderately by perceived economic benefits (ECOB) and the sustainability of ecotourism (ECOS). This pattern underscores that teachers primarily understand ecotourism as a nature-centered and educational construct rather than an economic or governance-driven domain.

Gender and marital status emerged as significant moderators of these perceptions. Female teachers displayed stronger associations between perceived benefits and overall ecotourism perception compared to males, aligning with the literature that links women’s environmental concern to empathetic, community-oriented values. Married teachers, likewise, showed higher recognition of the economic and cultural gains of ecotourism than single teachers, reflecting the family-centered appreciation of community welfare and local development. These results confirm that demographic and social experiences play crucial roles in shaping educators’ sustainability perspectives and that ecotourism-related training should be sensitive to such differences.

The validated three-dimensional structure (environmental outcomes, perceived benefits, and sustainability trust) confirms the conceptual robustness of ecotourism perception models. Teachers’ emphasis on ecological values indicates that their perception of ecotourism is embedded within a pedagogical ethic of environmental stewardship. This aligns with contemporary research recognizing educators as vital social agents who transmit sustainability values to students and communities. Hence, empowering teachers through ecotourism education represents a strategic mechanism for strengthening environmental literacy and advancing the United Nations Sustainable Development Goals, particularly SDG 4 (Quality Education), SDG 12 (Responsible Consumption and Production), and SDG 15 (Life on Land).

The findings of this study provide valuable guidance for policymakers, educational authorities, and tourism planners seeking to strengthen the link between education and sustainable tourism. Integrating ecotourism and sustainability content into teacher training curricula (both pre-service and in-service) can enhance educators’ ecological literacy through courses that blend local biodiversity, cultural heritage, and participatory pedagogy. Training approaches should also be gender- and family-sensitive, recognizing that female and married teachers display stronger pro-ecotourism orientations and therefore benefit from programs emphasizing intergenerational learning and care-based environmental education. Moreover, improving institutional trust and governance awareness is essential, as transparent resource management, environmental certification, and participatory decision-making can bolster teachers’ confidence in sustainable tourism governance. Strengthening collaboration between the Provincial Directorates of Education and Culture–Tourism can further institutionalize ecotourism initiatives in schools through field trips, ecoclubs, and service-learning projects that link classroom learning with local tourism development. Establishing formal certificate programs in “Ecotourism and Environmental Education,” jointly developed by universities and public institutions, would professionalize teacher competencies and position them as regional sustainability leaders. Finally, harnessing digital storytelling, virtual field experiences, and social media platforms can expand the reach of local ecotourism narratives and empower teachers as effective communicators of sustainability values within their communities.

The MGSEM findings confirm that teachers’ perceptions of ecotourism significantly differ across teaching levels, with elementary school teachers exhibiting a markedly stronger linkage between their ecotourism perceptions and perceived economic and sociocultural benefits. This suggests that elementary educators are more likely to view ecotourism as an avenue for local development and community enrichment, reflecting their closer engagement with experiential and place-based learning contexts. Conversely, secondary teachers appear to maintain a more conceptual understanding of ecotourism’s outcomes, indicating a need for curriculum strategies that better connect theoretical sustainability principles with local practice. This results highlight the importance of tailoring ecotourism education and professional development programs to teaching levels to enhance teachers’ roles as catalysts for sustainable community awareness and action.

These findings emphasize that teachers constitute a critical yet underutilized stakeholder group in ecotourism development. Their environmental orientation, moral leadership, and educational influence can catalyse a community-level transformation toward sustainable tourism practices. Policy frameworks that bridge education and tourism can thus foster “learning landscapes,” where ecological conservation, cultural continuity, and local prosperity reinforce each other.

In essence, this research highlights that enhancing teachers’ ecotourism literacy is both an educational and developmental priority. Düzce’s natural and cultural diversity provides a fertile context for integrating school-based environmental education with community-based ecotourism. Institutional collaboration, targeted teacher training, and policy coherence between education and tourism authorities can transform ecotourism into a participatory learning model—where schools become local anchors for sustainability and teachers serve as the primary agents of ecological citizenship.

This study also advances understanding of forest–society interactions by demonstrating that teachers’ perceptions of ecotourism represent a meaningful social dimension of sustainable forest management and nature-based tourism governance. By validating a multidimensional perception structure and establishing full measurement invariance across gender, marital status, and teaching level, the findings provide a robust empirical basis for interpreting teachers’ ecotourism perceptions as comparable and policy-relevant constructs.

The results indicate that teachers generally hold positive perceptions of ecotourism, with stronger engagement in accessible nature-based activities and comparatively lower familiarity with knowledge-intensive and conservation-oriented forms of ecotourism. From a forest governance perspective, this pattern suggests an opportunity to strengthen ecological literacy and stewardship-oriented practices through targeted educational interventions. Integrating forest-based learning, interpretation planning, and outdoor education into local tourism and forest management strategies may enhance the sustainability of nature-based destinations while mitigating pressures on forest ecosystems.

The findings further highlight the strategic role of teachers as intermediary actors linking education systems, local communities, and forest governance institutions. Recognizing teachers as stakeholders in forest and ecotourism policy processes can contribute to improved social acceptance of management measures, enhanced participatory governance, and long-term sustainability outcomes. Accordingly, forest administrations and tourism planners are encouraged to collaborate with educational institutions to develop forest school initiatives, interpretation programs, and curriculum-linked ecotourism activities that align with sustainable forest management objectives.

The study underscores that effective forest governance under conditions of global environmental change requires not only regulatory and technical approaches, but also sustained investment in social learning and education. Positioning teachers within integrated forest–tourism–education frameworks offers a viable pathway for strengthening sustainability transitions in forest-dependent regions.

Several limitations should be acknowledged. First, the study is based on cross-sectional survey data collected in a single province, which may limit the generalizability of the findings to other regions with different ecological or institutional contexts. Second, teachers’ perceptions were measured through self-reported data, which may be subject to social desirability bias. Third, while measurement invariance was established, the study did not examine causal relationships between perceptions and actual behaviors or policy engagement.

Future research could address these limitations by employing longitudinal designs to assess changes in ecotourism perceptions over time, incorporating qualitative methods to explore underlying motivations and narratives, and extending multi-group structural analyses to include other stakeholder groups such as forest managers, tourism operators, and local residents. Comparative studies across regions and countries would further contribute to understanding the role of education in forest governance and ecotourism under diverse socio-ecological conditions.