To determine the NMR yield of fluorine containing product, I need to calculate %yield from 19F NMR spectra using trifluoromethylbenzene as reference standard.
If the Educt has also a 19F you do not need a reference, because the ratio of the 19F signals of educt and product is the yield. NMR signals directly represent molar ratios. For 19F you may have to take care for the excitation profile of the pulse you are using. I normally place the carrier in the middle between the educt and product signal (doing so, both have the same offset effect)...
In addition to what Alfred Ross said, Please also make sure that the relaxation delay is set sufficiently long to have a full relaxation for all fluorine atoms of interest. Otherwise the result may still be biased.