What can you say for sure about an individual after studying a sample of the population? For example, if 80% of my sample of people living in a certain town is bald, what are the chances that an individual from that town is bald?
Edit: This is all fictional, But let me be more precise. Let's say I have a sample (n = 350) of the population of a town (N = 5000). In this sample I find that 80% are bald. Can I say that John, who lives in this town has an 80% chance of being bald? 
 A: If 80% in your sample is bald, then that suggests that your sample consists overwhelmingly of men over 50 years old. I suspect that that is extremely far removed from your population. With such an extreme bias, I would say that pretty much nothing can be learned from that sample.
A: That's a great question. I don't think there is much you can say by drawing from one sample of the population and making an inference because we don't know the complete sample space. For example if we toss 3 coins, we know that there can be 8 outcomes and not all are equally likely, for example probability of a HHH is only 1/8. However, if I didn't know the sample space that they it came from I won't be able to compute the probability accurately. In that case I would probably bootstrap the event and try to derive the probability and I think that's what you need to do here to come to a conclusion. Draw repeated samples and calculate proportion of bald men. Plot it on a histogram, the centre of that histogram is the value that would be closer to the actual proportion. 80% could be a one off event or due to poor sampling method but assuming that random sampling has been applied it could just be an extreme event.
