This study examines English-Korean code-mixing in the everyday and digital discourse of Korean university students, investigating the pragmatic functions underlying this practice. A survey was developed by three experts in sociolinguistics and pragmatics and administered to 383 university students at a university in Gyeonggi Province, South Korea. Two rounds of exploratory factor analysis validated a five-factor, 16-item structure: communicative efficiency, playful function, translingual creativity, face-saving mitigation, and social identity. A repeated-measures ANOVA showed significant differences among the five functions, with communicative efficiency (M=3.40) rated highest, followed by playful function (M=3.13), translingual creativity (M=2.88), face-saving mitigation (M=2.65), and social identity (M=2.23). Code-mixing also varied significantly by interlocutor (friends > strangers/family > seniors > professors) and medium (text messaging > face-to-face > social media > email/presentations). Content analysis of two open-ended items revealed that face-saving mitigation and playful function, though lower on the Likert-scale items, emerged prominently in students‘ reported experiences, adding depth to the quantitative findings. These results show that English-Korean code-mixing, previously regarded as a marked choice signaling elite status, has shifted into an unmarked, agentive strategy oriented toward practicality and playful peer bonding, calibrated according to social context.