I would be very careful about saying the test set was broken on your evidence. I have done a quick calculation, and taking into account a 2kt error in maintaining airspeed, a 2 kt variation in GPS ground speed, a 100 ft error in measuring/maintaining altitude, 2° error in temperature gives a measurment error of nearly 3 knots. These figures are based on real data, (although I have added a bit on to the altitude and airspeed taking into account you are not a test pilot. Sorry
) I have not attempted to look at the effect of heading, or the variation in wind conditions. The maths will take too long. As I said before, you cannot make a judgement on one test point. You need to at least cover the entire speed range to see what the trend is. Ideally, you need to repeat the tests several times.