How do ESL learners react to training in oral peer corrective feedback? This quasi-experimental study was an exploration of that question. Over 13 weeks, a treatment group of francophone ESL learners was first observed, then trained in oral peer corrective feedback (i.e. the provision of corrective feedback to one learner by another) on the pronunciation of th - in English. All group activities were audio-recorded. A control group of francophone ESL learners was also observed and recorded. Both groups were pretested and posttested for accuracy of their pronunciation of th -. Additionally, two learners each from the treatment and control groups were followed throughout the study for changes in pronunciation of th -. Although oral peer corrective feedback on th - occurred in the treatment group, no significant differences between groups in accuracy of pronunciation of th - were found. Out of the four learners tracked, one learner from the treatment group and one learner from the control group demonstrated moderate improvement in the pronunciation of th - over the 13-week study. Possible reasons are discussed for the pattern and manner of oral peer corrective feedback provided by the treatment group, and the limitations and implications of the results are considered.