Integrating patent analytics and employee sentiment: a multi-dimensional methodology for assessing organizational culture and R&D performance by company size in the Korean construction industry

In this study, we explore the relationship between organizational culture and R&D performance in the Korean construction industry using an integrated approach combining patent analytics and employee sentiment analysis. We analyzed 2,699 employee reviews and 672 patents from 2015-2024 across public institutions, large enterprises, and SMEs, employing the BM25 text-weighting algorithm to derive eight innovation indicators such as Concentration Score (CS), Diversity Score (DS), and Consistency Score (ConS). Organizational culture was evaluated using Psychological Safety Metrics (PSM) and Network Density Indicators (NDI). The findings reveal distinct correlational patterns: in public institutions, a strong positive correlation was observed between psychological safety and innovation consistency (ρ = 0.771), while network density was positively correlated with patent intensity (ρ = 0.51). In large enterprises, network density was positively correlated with innovation diversity (ρ = 0.420). Conversely, SMEs showed a negative correlation between network density and patent intensity (ρ = −0.430). Notably, peer support was negatively correlated with concentration scores (ρ = −0.35), indicating that highly collaborative environments may not always foster breakthrough innovations in certain contexts. These insights are valuable for customizing innovation strategies to organizational traits in the Korean construction sector and enhance understanding of factors affecting information-driven R&D performance. While focused on a specific context, this methodological framework offers a scalable model for assessing the culture-innovation link in other industries and geographies.