The construct validity of three entry level personality inventories used in the UK: a cautionary case study