Despite the growing popularity of paired format speaking assessments, the effects of pre-task planning time on performance in these formats are not yet well understood. For example, some studies have revealed the benefits of planning but others have not. Using a multifaceted approach including analysis of the process of speaking performance, the aim of this paper is to investigate the effect of pre-task planning in a paired format. Data were collected from 32 students who carried out two decision-making tasks in pairs, under planned and unplanned conditions. The study used analyses of rating scores, discourse analytic measures, and conversation analysis (CA) of test-taker discourse to gain insight into co-constructing processes. A post-test questionnaire was also administered to understand the participants' perceptions toward planned and unplanned interactions. The results from rating scores and discourse analytic measures revealed that planning had limited effect on performance, and analysis of the questionnaires did not indicate clear differences between the two conditions. CA, however, identified the possibility of a contrastive mode of discourse under the two planning conditions, raising concerns that planning might actually deprive test-takers of the chance to demonstrate their abilities to interact collaboratively.