Selective multimodal representation and visitor appraisal in educational heritage: evidence from Chinese academies

Introduction Research on heritage experience has examined material settings, cultural meanings, and visitor appraisal through psychological measures and digital visitor records. Their relationships within the same multimodal records remain less clear. Methods This study examines these relationships in educational heritage using 2,855 Ctrip reviews, 4,531 valid sentences, and 6,975 visitor-uploaded images from four Chinese academies, including 829 reviews with matched text and images. Sentence-level topic and valence classification, image classification, review-level association analysis, and coding of explicit evaluations of environmental qualities (EEQE) were combined. Results Buildings and cultural symbols accounted for 72.46% of images, while text covered a broader range of heritage, environmental, and practical visit content. Among reviews containing visible heritage imagery, 47.65% also contained heritage-related discourse. None of the 35 topic–image associations met the Benjamini–Hochberg FDR criterion after adjustment for academy, sentence count, and image count. In 617 identical reviews, the association between heritage-related discourse and textual valence weakened after heritage-related sentences were removed; the paired coefficient difference was 0.098 (95% CI 0.049–0.147). Heritage-related discourse showed no statistically reliable association with platform star ratings. EEQE coding identified judgments of atmosphere, tranquility, solemnity, artificiality, and related qualities that were only partly aligned with whole-review valence. Discussion The findings show how different media and evaluative indicators preserve complementary aspects of heritage experience. Connecting visual representation, heritage-related discourse, environmental-quality appraisal, textual valence, and ratings within individual reviews provides a naturalistic framework for examining visitor–environment relations while retaining the specific information conveyed by each form of expression.