Lem and Visualappr. As you point out it becomes a question of basic instrument skills. I have had the opportunity to see an ASI fault inserted in a 737-300 simulator on several occasions with different crews. This is preceded by a full discussion on the pros and cons of an abort or continue - so its not just another bit of chicanery by an enthusiastic simulator instructor.
Despite the ideal situation of PNF spotting an ASI reading difference at the 80 knot call and advising the captain who then purses his lips and decides stopping or going it hasn't worked out that simple.
If the flight crew are not on the look-out for the fault (despite having been given the good gen at the briefing) there is several seconds of uncertainty and all the time the aircraft is accelerating fast.
To compound the problem I have also seen the PF commence rotation not on the bugs on his own ASI, but on the actual call "Rotate" by the PNF when in fact it was the PNF ASI that had the erroneous reading at the 80 knot call. The call "Rotate" when given in a firm manner is almost an instruction - not as it is meant to be, an airspeed reading like V1. "Rotate" used to be called "VR" in some airlines and still is in some.
High energy aborts are sometimes the cause of an accident and historically it is usually safer to continue a take off unless the aircraft will not fly for some reason. An ASI problem actioned well after the 80 knot call due to momentary confusion at which ASI is faulty has all the potential for a cock-up if an abort is decided. This is the advantage of simulator training - you can see the indecision that occurs when a mysterious erroneous ASI reading is discovered on the take off run. This is where good airmanship comes into play.