This study outlines the steps taken to create an academic multi-word unit list derived from corpus data. It gives details on the procedure used and the rationale behind why certain approaches were utilised. It also compares existing resources and makes some suggestions for practical use of the resulting resource.