Background: This study is a part of a larger research project where the effect of instructional guidance in terms of role definition and role switching was investigated on students' learning of Basic Life Support (BLS) during a 20-minute reciprocal learning episode with task cards. BLS is a lifesaving skill consisting of nine sub skills to be performed in a correct order. Students' BLS performance was individually assessed before (baseline), immediately after (intervention) and two weeks later (retention). BLS assessments were conducted on a Laerdal Resusci Anne manikin connected to a laptop to register the quality of chest compressions and rescue breathings. Data from the software together with qualitative data from trained BLS observers were entered into a scoring system to obtain individual BLS scores, ranging from 18 to 73 points. Results demonstrated that skill retention was enhanced in the structured group (i.e. the group where role switching and role definition was implemented; n?=?18) compared to the unstructured group (i.e. the group where students were free to organize themselves the activity; n?=?18). Instructing learners how to behave as doers and observers as well as prompting them to switch roles improves learning compared to when no guidance is offered. Purpose: In the present study, the author investigates this effect further by mapping and analysing the behaviours and verbal interactions of learners with and without instructional guidance on role definition and role switching. Methods: Behaviours and interactions of learners in both groups were analysed during the 20-minute learning intervention for acquiring BLS with task cards. All variables were analysed and reported on the level of frequency and duration. The latter was expressed in percentages of the 20-minute intervention time. Results and discussion: It was found that without guidance, learners spent significantly more time in reading the task cards in silence (27% vs. 6%). Although the frequency of motor time-on-task was significantly higher in the unstructured group (23 vs. 13), the percentage of time of motor time-on-task was significantly higher in the structured compared to the unstructured dyads (42% vs. 30%). Additionally, learners engaged more in the provision of instructions in the structured group (18% vs. 5%). Evaluating activities during learning correlated positively with students' performance of BLS at intervention. Asking for help irrelevant for learning the task and unsolicited feedback correlated negatively with student performance at intervention. Conclusion: Increased motor time-on-task and the provision of instructions together with could, therefore, be important factors in explaining short-term skill retention in structured peer learning settings. Future research could elaborate the effect of motor and cognitive evaluating activities in reciprocal peer learning.