Source: IMAGES OF PEOPLE AND BIRDS CREATED BY JONATHAN OSTERS.
In this tutorial, you're going to learn about convenience samples and self selected samples. Both of these are a bad way to go. One of the things that we know about sampling is that it's important for samples to be representative of the population that they come from. What we mean by that is when we take our sample, which is some subset of a larger population, we want this sample to behave just like the population would if we sampled them all. Now, sampling everybody is not sample at all. N called a census.
But we want the sample to behave as similar to the population as we can, so that when we calculate statistics from our data, the statistics are as accurate about the population as they can be. We want to be able to generalize what we find in this box to the people outside the box. And so we want to be able to generalize, and we need a representative sample in order to do that.
The two ways that we're going to be learning about in this tutorial systematically don't give representative samples, and so they really should be avoided. They have major flaws-- these two designs that we're going to talk about. We would highly recommend that you not sample this way. Now this isn't to say that they're not done. We'll give you some examples of when they are done, but we would still recommend you not collect data this way.
One way is to take a convenience sample a sample that happens to be easy together. Suppose there's a crowd of people and this one guy with a clipboard, and he wants some data. He might take the nearest people to him, and say, hey, would you like to take my survey please. Now, these people might be representative of the population, but they might not. They all happen to be at the same place at the same time, which means that they might have some similarities that might make them not representative of the larger population. There's the risk in them not being representative is too high. It's not a good way to go.
Rule of thumb, convenience sampling is not valid, because people in these similar locations often feel the same way. We say, birds of a feather flock together. For instance, if you ask people about their spending habits, and they happen to be at the mall, well, that probably means they have similar ideas of how they should spend their money.
There's also self-selected samples. We also call them voluntary response samples. And those are samples where people can choose to participate. This is particularly common at the mall where you might have a lady with a clipboard say, "Would you like to take a survey, honey?" And you either walk right past and say, "No. Thanks." Or you say, "Sure, I'll stop and take your survey."
Now, this is problematic in a couple of ways. Number one is that you might get only participants who feel very strongly about the subject at hand. For instance, if your question is about politics, you might get only the very, very liberal people or the very, very conservative people. And then there's a lot of people who are ambivalent about politics. They don't really care much one way or the other. You'll get the most extreme viewpoints and not the viewpoints of the people in the middle.
The other reason that self-selected samples are probably not a wise idea is participants might be compensated for their time. They might be wanting to tell the interviewer exactly what they want to hear. If they get compensated in some way, you really just can't trust those results, because you might get people that are just saying whatever the interviewer wants them to say.
So to recap, representative sample is important if we want to generalize our findings to the population, which is exactly what we want to do. Convenience samples and voluntary response samples don't do that. So we really shouldn't do it that way. So we talked about representative samples, which are what we desire, and then bad ideas, convenience sample, and voluntary response samples. Good luck. We'll see you next time.