OCR Strange Behaiviour

I am experiencing what, probably because of my lack of knowledge, is strange behaviour in importing PDF documents into DEVONthink Pro Office in that some are automatically converted to PDF+Text and soe are not. Any idea why this should be? Thanks.