To test a claim, we look at some appropriate data and see if it supports the claim or not.
Scenario 1: The claim is that this coin is "fair" - meaning that it is equally likely to come up heads or tails each time. You get some data by tossing the coin 20 times. When you did that, you got heads 2 times and tails 18 times. Using your intuition, do you think these data provide significant evidence against the claim that the coin is fair?
Scenario 2: The claim is that this coin is "fair" - meaning that it is equally likely to come up heads or tails each time. You get some data by tossing the coin 16 times. When you did that, you got heads 7 times and tails 9 times. Using your intuition, do you think these data provide significant evidence against the claim that the coin is fair?
In each of these scenarios, if the claim is true, and the coin is fair, then we expect approximately half the outcomes to be heads. We do not expect that exactly half the outcomes will be heads.
So the outcome in Scenario 2 is quite consistent with the claim - close to half the outcomes are heads. These data in Scenario 2 do not provide significant evidence against the claim of a fair coin.
The outcome in Scenario 1 is not consistent with the claim - the proportion of the outcomes that are heads is quite far from half. These data do provide significant evidence against the claim of a fair coin.
We can use data from a not-very-large sample (20 tosses of a coin), if it is taken by using a reasonable sampling method, to estimate a characteristic of a large population (all possible tosses of that coin.)
How much different from half and half must the data be for us to decide against the claim? Working through that idea is one of the main parts of our elementary statistics course. That's something to look forward to in our course.