Imagine an 8-segment spinner with segments labelled according to the data in the population dataset. Spinning this is equivalent to replacement after selection.
So the population (P) is infinite in size because of replacement. Now apply the population distribution to the population 100 (which we call R, the representative population) from which the sample of 16 will be taken.
⅜ of P are 1s, so in R there will be ⅜×100=37.5 1s, and so on for the other numbers: 12.5 2s, 4s and 6s, 25 5s. (Never mind that these are not all whole numbers. 37.5+12.5+12.5+25+12.5=100.)
The mean of P and R is 3.125. Their common variance (V) and standard deviations (SD) are about 3.859 and 1.984, the estimated V and SD for the sample are 3.859/16=0.241 and √V=0.491, or about 0.5. The mean of the sample is 3.125, the same as for P and R.