This study explores the differences between face-to-face (FTF) interaction and text-synchronous computer-mediated communication (SCMC) regarding corrective feedback (CF). Eight native speakers (NSs) and eight second language learners were paired up and divided into two groups: instructed and uninstructed. Native speakers in the instructed group were asked to provide corrective feedback to help improve their partner's English skills and given a short training on providing corrective feedback. Adapting a counterbalanced design, the NS-learner dyads completed a jigsaw task in each communication mode. We analyzed the corrective feedback episodes in each mode to see which mode provided more opportunity for corrective feedback negotiation. Furthermore, participants' attitudes toward using each mode for corrective feedback negotiations was examined qualitatively. The findings revealed that the instructed group produced more CF episodes than the uninstructed group. Additionally, participants focused more on negotiation of meaning during FTF interactions, but they paid more attention to form during text-SCMC interactions. NSs' and learners' attitudes toward corrective feedback varied depending on the mode they used for interaction. The findings indicate that both FTF and text-SCMC interactions are beneficial for learners' second language development aligned with each communication mode affordances. [Note: The volume number (9) shown on the PDF is incorrect. The correct volume number is 10.]