With the rapid development of autonomous driving technology, research on the factors influencing trust in autonomous driving has increased. However, due to differences in sample data and social contexts, there are inconsistencies in the direction and strength of these factors' effects on trust. Therefore, this study aims to identify the key factors that influence trust in autonomous driving, standardize their effects, and explore the moderating role of cultural context. This will lead to the creation of an analytical framework that balances theoretical insight with practical relevance.

Meta-analysis is a research method that statistically re-examines existing empirical findings. In this study, a literature search and screening were first conducted, including 134 studies in the meta-analysis. Next, data coding and extraction were performed, analyzing 265 effect sizes. The 20 influencing factors were then categorized into four dimensions: "individual characteristics, subjective perceptions, technical features, and traffic environment." Additionally, the study assessed publication bias using the Inverse Safety Coefficient, Egger's regression coefficient, and Begg's test. Heterogeneity was analyzed with the Q-test and I² test. Finally, main effect tests were conducted, and subgroup analysis was used to explore moderating effects.

The research results indicate that 16 factors are significantly associated with trust in autonomous driving, distributed across the four dimensions mentioned above. Among these, the traffic environment dimension exhibits the strongest correlation with trust in autonomous driving. Specifically, factors such as attitude toward autonomous driving, social influence, perceived benefits, perceived safety, system transparency, and traffic density show a strong positive correlation with trust. In contrast, factors including perceived risk, perceived ease of use, perceived usefulness, humanization, reliability, explainability, and automation level exhibit moderate correlations with trust. However, the correlations with driving experience, gender, and education level are relatively weak. Furthermore, cultural background plays a critical role in moderating these relationships, especially in the moderation of the relationship between gender, perceived risk, perceived ease of use, and trust in autonomous driving.

Through a comprehensive meta-analysis, we developed an integrated analytical framework for understanding trust in autonomous driving, which encompasses four critical dimensions: individual characteristics, subjective perceptions, technical features, and traffic environment. The framework establishes a systematic theoretical foundation for exploring the mechanisms underlying the formation of trust in autonomous driving. In addition, we identified and clarified the hierarchical relationships of effect strengths across these dimensions and key variables, a finding that provides guidance for prioritizing technological optimizations and formulating targeted promotion strategies. Furthermore, the examination of the moderating effect of cultural context elucidates the variability observed in prior research, offering a theoretical rationale for implementing differentiated promotion strategies tailored to diverse cultural contexts. Lastly, a thorough review of the existing literature highlights a predominant focus on mediating variables, such as personal self-perception, with numerous studies investigating their mediating roles within the "external stimulus---personal self-perception---trust in autonomous driving" logical framework.

This study provides valuable insights that can guide the research and development direction of autonomous vehicle manufacturers. It emphasizes the importance of enhancing the functional design of transparency and anthropomorphic features within autonomous driving systems. Furthermore, it advocates for the integration of design considerations that address the variability of traffic environments, as well as the diverse user priorities shaped by different cultural contexts, in order to foster greater user trust and facilitate the widespread adoption of the technology.