Assessing the validity and reliability of descriptor variables used in scenic highway analysis

The research presented here evaluated descriptor variables used by agencies to assess scenic quality along roads in their jurisdiction. The goal was to determine if these descriptors, when applied in an expert assessment, could produce results indicative of public preference for highway conditions. Several state and federal scenic highway programs were first reviewed to determine what descriptors were most widely used in their assessment efforts. From that review, four descriptors were selected: naturalness, vividness, variety, and unity. A photographic inventory was generated along a road corridor in California’s Central Coast region. A pool of landscape architects was shown test slides from that corridor, and was asked to rate each scene using the four descriptors. Untrained observers were then shown the identical slides, and were asked to rate the scenic beauty per-scene. Mean scenic beauty scores were compiled and later related to the expert judgments, using correlation, regression and factor analysis. Results indicate that vividness had a significant relation with preference, and was most correlated with the construct scenic beauty. Of the three other descriptors, only variety presented a significant relation to preference. The contribution of variety was; however, limited, as it did not supply additional information concerning preference beyond that provided by vividness in the regression equation. Of note is that naturalness did not significantly add to the analytical ability to predict preference. This runs contrary to discussions in the cited programs, which imply that natural conditions are the scenic ideal.