The margin of error on the average user doing a liquid test has to be at least +- 10%. And maybe it is even as high as +- 25%. That is, if you read an ammonia reading as 2.0 ppm, +-10% is the reading is really between 1.8 ppm and 2.2 ppm. And it may even be as wide as 1.5 ppm and 2.5 ppm.
It may even be worse, since the color charts are typically exponential in nature. I.e. the ammonia color test goes 0.25 ppm, 0.5 ppm, 1.0 ppm, 2.0 ppm, 4.0 ppm, and 8.0 ppm. So when you say +-10% on an exponential curve the upper bound can be really high. I.e. +-10% on a reading of 2.0 ppm is probably between 1.8 ppm and 2.4 ppm -- where the higher bound is because there is two units between the color change on the upper side than the lower.
The errors are simply that only rarely does a color exactly match. So, it is a guess what the color is actually saying. And, then, a "drop" is not a terribly well calibrated unit of fluid. It depends on things like how tightly one squeezes the bottle, the temperature of the bottle and the air around it, and even a little bit on the humidity of the surrounding air.
So, as I said, I'd put error margins even as high as +-25% on any reading. In short, these aren't laboratory grade tests we are doing -- nor do we really want them to be. We really don't need very exact numbers. The presence of any ammonia and nitrite is a cause for concern. And, we don't need to know the pH much beyond +- 0.5 pH units for almost all purposes, nor degrees of hardness within a degree or two. To be successful, you really only need an idea of what is going on, only rarely are the exact details needed.
