Explanations for the misinformation effect have focused on issues in long-term memory. This exploratory study re-examines data from previous research and suggests that a visual versus verbal distinction during encoding could account for seemingly contradictory results. The working memory model of Baddeley and Hitch (1974) incorporates such a distinction, and thus, using this framework, a dual-task study was conducted examining the roles of the phonological loop and the visuospatial sketchpad in processing misleading information, with a novel three forced-choice recognition procedure. Results indicate that the visuospatial interference task of manual tapping significantly increases subject’s performance to the same level as controls. Although non-significant, articulatory suppression led to an increase in reporting the misinformation item. Results favour an active role of working memory during the encoding phase, and further research is discussed, along with possible adaptations to the cognitive interview.