Ideally, if you want to learn about a population, you would study everyone in the population. Unfortunately, in reality, this is virtually impossible most of the time. 

In such cases, we have to rely on taking a sample of the population of interest and hope that whatever happens in the sample is representative of what’s happening in the population we drew the sample from. 

If we do it right, by taking a random sample in which everyone in the population we are interested in has an equal chance to be included, then we can use mathematics to infer that it may be true for the population of interest.