Beyond Resource Provision: Institutional Innovation Support, External Industry Connection, and Bottleneck-like Patterns in Faculty Innovation

Faculty innovation is central to the long-term innovative capacity of higher education institutions, yet institutional support is often studied either as a single overall condition or as a set of separate factors. Using a 2024 survey of 649 faculty members from 25 public higher education institutions in Zhejiang and Jiangsu, China, this study examines perceived institutional innovation support (PIIS) as a broad perceived support environment organized into three interrelated descriptive domains, namely internal resource and institutional environment, innovation service infrastructure, and external industry connection. Regression models assess associations with faculty innovation capability (FIC), measured as an observed index combining reported innovation behavior, activity, and performance, while Necessary Condition Analysis examines upper-end bottleneck-like patterns. Overall PIIS was positively associated with FIC across alternative inferential specifications. In the joint-domain model, innovation service infrastructure and external industry connection retained statistically distinguishable incremental associations, whereas internal resource and institutional environment did not. Across the association-based sensitivity analyses, innovation service infrastructure was the most consistently supported domain. The upper-end analysis revealed a different pattern, with external industry connection producing the largest observed ceiling effect under both ceiling techniques and remaining the clearest and most persistent bottleneck-like candidate across sensitivity analyses. Its empirical boundary became more restrictive mainly near the highest observed FIC levels. The findings suggest that conditions associated with variation across faculty need not be the same as those that become difficult to bypass near the upper end, while all conclusions remain associational and sample-grounded. PIIS is used here as a project-specific organizing framework rather than as a newly validated construct empirically distinct from adjacent support or climate constructs.