For example, there are numerous cases in the past where it was felt that knowledge of the true distribution of tornadoes would hurt the image of a state, so tornado reports were systematically suppressed. In other cases, especially when people are killed by the weather produced in thunderstorms, pressure can be exerted to force the assessment of the event to conclude that it had to be a tornado, in spite of clear evidence that it was not; political pressure can force the issue, such that some non-tornadic events have been recorded as tornadoes.
 In addition to raising questions about their objectivity in this task, it ends up being an extra duty, above and beyond their normal full-time responsibilities, and certainly not supported with anything resembling adequate resources.
 It's like saying Tutankhamun had the most gold ever found in a Pharoah's tomb ... generally speaking, most of the Pharoahs' tombs had been robbed, so we have virtually no information about how much gold might have once been buried with the Pharoahs. Tutankhamun was a minor king, so it is unlikely that his tomb actually had the record amount of gold, but since the tomb had not been thoroughly looted, that gold which remained is, indeed, the most ever found.
 There are those who advocate another scale of windspeeds vs. damage, the so-called TORRO scale. If I am basically correct in assuming that the F-scale assignments are more precise than our ability to discrimate ... that is, the F-scale ratings are probably only accurate to within plus or minus one F-scale, so that we can only discern (reliably) three types of tornado damage: say, weak (F0-F1), strong (F2-F3), and violent (F5-F5), with any reliability ... then imagine the difficulties of a scale with twice as many divisions!
 NOTE: the wording here in my e-mail response is a bit ambiguous. The DOW observations certainly are above the surface, at about 100 meter heights or even higher, depending on the situation. The damage, however, is done at ground level ... hence, the F-scale values are meant to apply near ground level, not at a height on the order of 100 meters. As originally intended, the F-scale is connected to the Beaufort scale ... by international agreement, the Beaufort scale applies to a nominal height of 10 meters. Hence, the F-scale winds should be measured at the same height. I hope this clears up the ambiguity.