Photoshopping, fraud and circular logic in research