For a very basic example, say you have 1 million people, 200 000 prefers burgers, and 800 000 prefers pizza, then say out you pick people out randomly from the group of 1 million people
How many do you need to pick out to have a 95% certainty that the ratio falls within 95% of the general distribution in the population? The answer is: 246. 246 is a big enough sample size for a 95% confidence that you are within 95% of the range of the general population distribution in this specific example
There’s a lot more to this, of course, but hopefully this is sufficient to showcase that you do not need large amounts of data to derive conclusive results
Usually in a scientific context you go more the route of calculating the confidence percentage that the data you got is random, also known as null-hypothesis testing, where the confidence percentage is the p-value. So the inverse of that is the confidence that it’s not random
But, again, there’s so much more to statistics than this, this is just the very basics.