Why am I writing about this 'silly' 'simple' thing? I think, time and again, we make an error and do not distinguish between 'sensible' rounding and estimation. They are different and the sample mean is a continuous variable, therefor, we are not allowed to round that 'sensibly'. That's how I look at it: expected value (or the mean value) is an abstract concept, 'it's the way we perceive in our mind'- reminds me, in a way. about histograms in statistics.
For clarification I will be using the following abstract from the book:
Fowler, Jim, Lou Cohen, and Phil Jarvis. Practical statistics for field biology. John Wiley & Sons, 2013.