Yes, blind tests are quite funny when talking about vision.

When doing statistics analysis, it is important that the people do not know what frequency it is currently using, or what other people have answered. The bad thing with our minds is that we are so very easy to fool even when we "know" that we are correct.

So the best tests are if each person will se a number of frequencies in random order, and say yes and no to them one-by-one. Then it is time for the next person - definitely in a new order if the second person where allowed to see/listen when the first person did the test. If the refresh rate is constantly increased, then it is quite likely that the poeple will continue to think that they see the flimmer until they start to notice the strange looks they get from any test staff who can't really believe them.

When I was very small, they did a quite stupid sound test. They said I should tell if I heard a sound (sound - not beep). I heard a lot of sounds, and they got a bit worried, and asked if I really heard a beep. Yes - I heard a beep too. Then they suddenly did send a beep and I got mighty irritated by the loud noise of that beep compared to all other sounds I heard.

Somehow, they could not understand that just having a CRT in a nearby room was enough to let me hear a constant beep and without knowing the reference levels, I had to assume that this was the beep I was supposed to hear :)

Different tests are differently hard to set up. Many scientific reports by the research elite gets dropped at late stages (or even gets accepted for many years) just because everyone involved has missed an important correlation factor that makes the test data or the conclusions corrupted.

A blink test will affect the current consumption of the processor. It may affect the light intensity of the diodes. If some part of the display scrolls, the scroll speed and the skew of the characters will vary. The power supply may produce intensity modulation from the 60Hz that gives similar interference as if running the display close to fluorescent lights. There are so many strange things that can vary even in a trivial test, that may somehow affect the tested persons. The last factor I mention - 60Hz modulation of the DC powering the diodes - is a very important factor since it not only explains why people can see close to 120 Hz refresh but should be applicable in real-life to a lot of fixed installations. In this case, you then found one of two possible solutions to the problem.