We know that the issue is that if we were to take another random sample of emails, we probably wouldn’t get the same sample proportion that were sent by students. This means we can’t say that the proportion of all the lecturer’s work emails during 2022 that were from students is exactly 27.5%. The variation in estimates from different samples is due to sampling variation.
We use the term sampling variation to describe the way estimates made from samples vary (change) from sample to sample. We