Speak “As Good As Men”? An empirical study of gender variation in Chinese courtroom audio

Do men and women perform differently within a male-dominated judicial system? We investigate this question by using a machine-learning algorithm to analyze oral speeches in criminal trials in China, measuring judges’ proficiency in Standard Mandarin as an external indicator of linguistic performance and professionalism. Results show little gender-based variation: male and female judges speak Mandarin with comparable fluency. Crime severity, a proxy for political salience and strategic case assignment, is positively associated with linguistic proficiency, suggesting that speech uniformity reflects both procedural norms and institutional symbolism. Rather than implying gender equality, this pattern illustrates how linguistic parity can serve the regime’s performative construction of legitimacy through hegemonic vernacular and gender representation. By combining computational analysis of audio data with theories of language and power, the study provides new empirical evidence on gender and institutional behavior under authoritarianism. It demonstrates the potential of audio data for studying judicial communication.