Testing of fluid cracking catalysts in laboratory microactivity units is carried out in either fixed- or fluid-bed reactors. The latter are becoming increasingly popular because of the automation they provide. This paper focuses on understanding the effects of testing protocols on product selectivities in such reactors and explains the cause of artificial conversion and selectivity effects reported previously when a variable time-on-stream protocol was used. We show that coke deposited at the onset of a run lowers the coke selectivity of the subsequent feed to the reactor. This, in turn, affects all other product yields. The appropriate method to change conversion is to vary the mass of catalyst and thus change space time. This is best practiced without introducing other artifacts by using inerts to dilute the catalyst so that the total solid content in the reactor is constant. Finally, we have shown that a small fluid bed is significantly backmixed. Therefore, as is normal for a backmixed system, when compared to a plug-flow fixed bed at constant time on stream, the fluid bed showed lower conversion at the same space time, increased yields of final products, and lowered yields of intermediate products. However, yields from a backmixed fluid-bed reactor at longer times on stream approached yields from a plug-flow fixed-bed reactor operated at a shorter time on stream.