I noticed something during the 2016 election and polls - the number of people that actually take the polls can in no way represent the whole of the USA. A couple thousand people absolutely reflect roughly 160 million voters? Tell me how that can be statistically possible.