Historic gardens represent a unique intersection of cultural heritage and environmental protection. The International Charter of Florence defines them as architectural and horticultural compositions of historical or artistic interest, emphasizing their conservation as cultural monuments. However, their ecological function, providing habitat structures and supporting biodiversity, is equally significant. This dual nature—cultural and ecological—presents opportunities for interdisciplinary learning, integrating environmental education with historical and aesthetic content. Despite this potential, didactic approaches utilizing historic gardens remain largely unexplored. Understanding the impact of such educational programs on psychological constructs, particularly attitudes, is crucial. Previous research has shown positive effects of nature-based education on attitudes and behavioral intentions, but these findings haven't been extensively investigated within the context of historic gardens. A key barrier to such research is the lack of suitable psychometric tools. This study addresses this gap by developing and validating the Garden Heritage Scale (GHS), a novel instrument designed to measure attitudes toward historic gardens, incorporating affective, cognitive, and behavioral components based on established attitude theory. This scale will enable a more nuanced understanding of the effectiveness of educational interventions in these unique settings, paving the way for evidence-based improvements in environmental education within historic garden environments. Existing research on environmental attitudes often focuses on ecological aspects, neglecting the specific values associated with cultural heritage sites. While qualitative studies have examined park visitor perspectives, the lack of appropriate quantitative scales hinders systematic investigation. This research aims to fill this void, providing a tool to measure the affective, cognitive, and behavioral dimensions of attitudes towards historic gardens, ultimately contributing to a richer understanding of how educational interventions can shape environmental attitudes and actions within these spaces.

The theoretical underpinning of the study rests on established social psychology theories of attitudes, which define them as evaluative beliefs influencing behavior. These are considered latent constructs, not directly observable. A widely accepted three-component model posits affective (emotional), cognitive (belief), and behavioral (intentional) dimensions. While this model enjoys support, alternative perspectives suggest simpler structures. Research on environmental attitudes often focuses on preservative and utilitarian perspectives. The relationship between attitudes and behavior remains a topic of ongoing debate; correlations can be weak, suggesting mediating factors such as social norms influence behavior. This study adopts the three-component model, recognizing the interplay of emotion, cognition, and intention in shaping attitudes toward historic gardens. Existing scales for measuring environmental attitudes largely focus on ecological concerns and aren't fully applicable to cultural heritage sites. Limited research specifically addresses attitudes toward historic greenery, with qualitative methods dominating. This study addresses this research gap by developing a quantitative instrument (the GHS) explicitly addressing the three-component attitude model within the context of historic gardens.

The development and validation of the GHS followed a rigorous procedure involving multiple stages. First, an initial item pool of 18 items was generated, six items for each dimension (affective, cognitive, behavioral). Seven items were adapted from an existing scale (Sieg et al., 2018). An exploratory factor analysis (EFA) was conducted on data from the first sample (n=233) to assess item suitability and identify underlying factors. This sample included 143 adults from the general population and 90 high-school students. Data were collected through online and paper-pencil surveys. Principal axis factoring with promax rotation was used, aiming for a three-factor solution. Items with cross-loadings less than 0.20 were eliminated. The scale was refined to achieve balanced subscales. A confirmatory factor analysis (CFA) was then performed on a second independent sample (n=183) to confirm the factor structure. This sample consisted primarily of adult students from Goethe University Frankfurt. Maximum likelihood estimation was employed, and model fit was assessed using CFI, TLI, RMSEA, SRMR, and χ²/df. Three models were compared: a three-factor model (Model 1), a two-factor model (Model 2), and a one-factor model (Model 3). Model fit was evaluated using several fit indices, and a likelihood-ratio test was conducted to compare the models. Criterion-related validity was evaluated by correlating the GHS with a convergent measure (modified Nature Interest Scale focusing on historic gardens) and a discriminant measure (adapted scale assessing attitudes towards wilderness). Pearson correlations were used to assess the relationships between the GHS subscales and these external references. Finally, reliability was assessed using Cronbach's alpha for internal consistency and test-retest reliability with a panel survey of biology students (n=41 initial, n=31 retest). All analyses were performed using SPSS 29 and Amos 29.

The KMO test for the EFA (n=233) yielded a coefficient of 0.904, indicating the data's suitability for factor analysis. The initial EFA suggested a three-factor structure, although the eigenvalue for one factor fell slightly below the Kaiser criterion. Several items were excluded based on cross-loadings and logical considerations. The final EFA resulted in a three-factor structure (affect, behavior, cognition) with high factor loadings (>0.590). The CFA (n=183) using three different models revealed that Model 1 (three-factor structure) exhibited the best fit, outperforming both a two-factor and a one-factor model in terms of fit indices (CFI, TLI, RMSEA, SRMR, χ²/df). The likelihood-ratio test confirmed that Model 1's superior fit was highly significant (p<0.01) compared to the other models. The path diagram of Model 1 showed moderate-to-high regression weights, indicating a good model fit. Correlations with the convergent measure (modified Nature Interest Scale) were moderate to high (average r = 0.494), supporting convergent validity. Correlations with the discriminant measure (adapted wilderness attitude scale) were low (average r = 0.114), indicating discriminant validity. Cronbach's alpha for internal consistency across all three subscales and the total scale ranged from 0.735 to 0.871, indicating good internal consistency. Test-retest reliability using a panel survey design showed a high correlation (r = 0.782) over an eight-week interval, demonstrating acceptable test-retest reliability.

The findings strongly support the validity and reliability of the GHS as a tool for measuring attitudes toward historic gardens. The three-factor structure, consistent with the three-component model of attitudes, was robustly supported across both EFA and CFA. The relatively lower correlation between the behavioral and affective components in the CFA warrants further investigation. The scale effectively captures the affective, cognitive, and behavioral dimensions of attitudes. The successful validation of the GHS fills a crucial gap in the research landscape. It provides a robust instrument for systematically evaluating educational programs in historic gardens, allowing for quantitative assessments of program effectiveness. The findings can inform the design and development of more effective environmental education programs in these unique learning environments. The results contribute to both environmental psychology and heritage studies, integrating these interdisciplinary fields.

This study successfully developed and validated the Garden Heritage Scale (GHS), a reliable and valid instrument for measuring attitudes toward historic gardens. The GHS is grounded in established attitude theory and provides a valuable tool for future research exploring the effectiveness of environmental education programs within historic settings. Future research could explore the GHS's applicability across diverse populations and geographic locations, further refining its use and generalizability. Investigating the relationship between attitudes and actual behavior in response to environmental education interventions would provide a richer understanding of the impact of these programs.

The study's limitations include the use of slightly different questionnaires for the EFA and the internal consistency analysis and some variations in the questionnaires used for criterion-related validity. These differences, however, primarily involved socio-demographic data or initially included all 18 items. The samples, while sufficient for the analyses, were relatively small and comprised a disproportionate number of young female adults, limiting generalizability. Further research is needed to assess the GHS's applicability across diverse age groups and genders and to ensure its robustness across different socio-cultural contexts. The use of self-reported behavioral intentions instead of observed behaviors is another limitation.