The investigative interviewing of a vulnerable suspect is a complex and difficult task. Current best practice advocates for the use of open questions in order to elicit a free recall. However, those with mental health conditions have limited cognitive abilities that relate to free recall and episodic memory, and there is emerging evidence that suggests open questions may not always be most suitable for the vulnerable interviewee. As such, the present study examined the impact of two different interview models (best practice v modified interview) on the amount and accuracy of investigation relevant information obtained within an experimental vulnerable ‘suspect’ sample. Participants engaged in two tasks; a minor transgression and a matched non-transgression. Each participant was then subject to either a best practice (containing largely open questions) or a modified interview (containing largely closed questions). Vulnerable participants provided a significantly higher and more accurate amount of investigation relevant information during the modified interview rather than the best practice interview. In addition, participants that have mental health conditions sought more clarifications during the best practice interviews. Our findings challenge current best practice in that vulnerable participants performed worse in interviews containing more open questions than closed questions. These findings add to the emerging evidence base that vulnerable individuals may require an alternative method of questioning, including the use of closed questions as ‘scaffolding’ during an investigative interview.