Champion Is Pretty Fair And Should Be In The Game - Change My Mind

You can't be 100% certain about wether or not the conclusion is false in the end when working with samples unless you know everything you want about the entire population, thing is. If you know everything you want from the entire population, there's no need to work with a sample.
Saying that because someone doesn't ask every single person involved there's no reason to believe their conclusion is correct is like saying every single survey ever taken was useless.

The difference between your and my analogy is that between inductive and deductive reasoning. With inductive reasoning, you randomly stumble upon a few observations, from which you draw a conclusion. With deductive reasoning, you start with a hypothesis, try things that would most likely disprove that hypothesis if it were incorrect, and if you fail to disprove it, you conclude that your hypothesis was most likely correct. Whilst still not completely fool-proof, the conclusions from deductive reasoning are definitely more trustworthy.