Objective Biodiversity perception is an important way for the public to achieve psychological restoration and well-being in urban green spaces. Existing studies have shown that compared to objective biodiversity indicators, perceived biodiversity has a more direct predictive effect on restorative experiences, positive emotions, and place satisfaction, and may play a mediating role between objective biodiversity and psychological well-being. However, the existing measurement tools have multiple limitations: in terms of measurement methods, early studies mostly required respondents to estimate the number of species, which actually measured ecological knowledge and counting ability rather than actual perceptual experience, with high cognitive load and limited data validity; in terms of measurement tools, domestic studies mostly used a small number of items for measurement, with inconsistent dimension divisions, insufficient reliability and validity tests, and lack of comparability among studies, and direct use of foreign scales faces the risk of cultural adaptation. Overall, there is still a lack of a localized perceived biodiversity scale that has been systematically developed and tested with a clear dimensional structure in China. This study aims to develop and validate the Perceived Biodiversity Scale in urban green spaces of China, with the goal of shifting the measurement focus from the judgment of species numbers to the evaluation of on-site experience based on perception cues, providing standardized tools for the assessment of urban green space ecological experiences, health and well-being research, and public natural perception evaluation.

Methods This study is based on the attention restoration theory and stress reduction theory, defining perceived biodiversity into four dimensions: perceived plant diversity, perceived animal diversity, perceived habitat heterogeneity, and perceived naturalness. Firstly, an initial item pool was constructed through literature deduction, and 7 interdisciplinary experts with backgrounds in urban ecology, landscape architecture, and environmental psychology were invited for content validity assessment. Items with ambiguous expressions or overlapping concepts were eliminated based on the item-level content validity index (I-CVI). Subsequently, a pre-survey was conducted among 36 participants through focus group interviews to identify semantic ambiguities and simplify professional terms, resulting in a pilot scale consisting of 22 items. The empirical investigation selected 6 typical urban parks covering different urban zones and biodiversity gradients in Chengdu as sample sites, and conducted an on-site questionnaire survey among 299 park users. Items were rated on a 7-point Likert scale. In the data analysis stage, exploratory factor analysis (EFA), confirmatory factor analysis (CFA), and reliability, convergent validity, and discriminant validity tests were used to systematically verify the scale structure.

Results Exploratory factor analysis showed that the KMO value was 0.930, the Bartlett’s test of sphericity was significant (p < 0.001), and four factors were extracted, with a cumulative variance explanation rate of 71.31%. The factor loadings of each item ranged from 0.574 to 0.831. In the confirmatory factor analysis, after eliminating two items with standardized factor loadings below the 0.500 threshold (VC4 “places with vegetation cover” and VD3 “traces of neatly trimmed artificial landscape”), the model demonstrated an acceptable fit (χ²/df = 3.070, RMSEA = 0.079, CFI = 0.924, TLI = 0.910). The final scale consists of 20 items from four dimensions: perceived plant diversity, perceived animal diversity, perceived habitat heterogeneity, and perceived naturalness. In terms of reliability, the Cronbach’s α coefficient of the full scale was 0.937, and the α coefficients of each dimension ranged from 0.775 to 0.929. In terms of convergent validity, the average variance extracted (AVE) for each dimension ranged from 0.65 to 0.79, and the composite reliability (CR) ranged from 0.78 to 0.93, all meeting the recommended standards. In terms of discriminant validity, according to the Fornell-Larcker criterion, the square roots of the AVE for each dimension were all greater than their correlation coefficients with other dimensions, and the four dimensions had good statistical independence.

Conclusion The Perceived Biodiversity Scale in urban green space constructed in this study has a clear four-dimensional structure and good psychometric performance, and can effectively measure the public perception of biodiversity. The research results reveal that the public’s judgment of urban green space biodiversity mainly unfolds along four paths: plant visual characteristics, animal activity cues, habitat structural complexity, and natural growth state. This is a multi-dimensional comprehensive evaluation based on directly perceivable environmental cues. This scale shifts the measurement of perceived biodiversity from specialized species estimation to public experience evaluation, providing standardized tools for urban green space ecological experience assessment, health benefit research, and public natural perception evaluation. At the same time, it can be used in conjunction with objective biodiversity data to identify deviations between ecological construction effectiveness and public perception experience. The limitations of the study lie in the sample source being adult park users in Chengdu and data collection being concentrated in a single season. The applicability of the scale in different regions, seasons, and populations still needs further verification.