DOES FIELDWORK STILL MATTER: EVALUATING THE EMPIRICAL APPROACHES TO POLITICAL GOVERNANCE IN THE AI ERA

Povzetek

Rapid advances in artificial intelligence have transformed empirical approaches to political governance by enabling large-scale data processing and automated analysis. At the same time, traditional fieldwork remains central for capturing contextual, cultural, and institutional nuances that are often invisible in purely computational data. The objective is to evaluate the relative and complementary contributions of fieldwork and AI-based methods to the quality and reliability of empirical governance analysis in the contemporary AI era. The dataset consists of survey responses from 370 political governance professionals, including academics, policy analysts, and public administration practitioners. Data were collected using structured Likert-scale instruments. Fieldwork Engagement (FE) and AI-Based Method Usage (AIU) serve as independent variables. Contextual Understanding (CU) and Data Scale (DS) function as mediating variables. Quality of Governance Analysis (QGA) and Reliability of Findings (RF) represent outcome variables. Internal consistency was assessed using Cronbach’s Alpha. Multiple regression analysis was applied to examine direct relationships among variables. Partial Least Squares Structural Equation Modeling (PLS-SEM) was employed to evaluate the integrated model and mediating effects. FE demonstrated a substantial effect on CU (β = 0.58), which in turn significantly enhanced the QGA (β = 0.48). Similarly, AIU strongly influenced DS (β = 0.62), leading to improved RF (β = 0.51). AIU significantly improves DS, leading to higher RF. The integrated model demonstrates that fieldwork and AI methods contribute through distinct but complementary pathways. Empirical governance analysis benefits most from a hybrid approach in which contextual depth derived from fieldwork is combined with the scalability and consistency enabled by AI-based methods.