This tutorial covers two kinds of unreliable samples. First, let's review what a representative sample is. A representative sample is one that's reliable. The relevant characteristics of the population are mimicked in the sample. They're both very similar. The kinds that we will look at in this tutorial are not reliable. They're not representative samples.
The two kinds of samples that we'll look at are convenience samples and voluntary response samples. Now, a convenience sample is one that is easy to select. It's convenient to choose, so, for example, if you wanted to do a survey and just walked around to interview the people near you. That would be very convenient. It's very easy for you. But it's not a very representative sample. It does not reflect the population that doesn't happen to be sitting closest to you.
Similarly, a voluntary response sample ends up being not representative but in a slightly different way. A voluntary response sample can also be called a self-selected sample. Individuals volunteer or choose to be part of it.
Now, when someone chooses to be part of a sample, they tend to have extreme opinions. One place that we see this a lot is on online customer satisfaction surveys. If you flew on an airline, you wouldn't really be likely to fill out their response if you had an OK time.
But if you had a fantastic flight attendant who took care of you every second of the way and gave you extra cookies in the middle of your flight, you'd want to make sure that the company knew about that so you would fill out the survey. And you'd give a very good survey response.
Now, on the other end, if you had a terrible time, if someone was kicking your chair, they lost your bag, everything under the sun happened to you, you'd also want to fill out that survey. So people who end up filling out the surveys have either extremely positive or negative opinions or somehow are on the extreme ends of whatever spectrum the survey is asking about.
In each of these four examples, we'll decide whether or not the scenario reflects a convenience sample, a voluntary response sample, or a representative sample. In our first example, a group of high school students wants to know about teacher attitudes. They interviewed teachers at their school.
This is very convenient for the high school students to do. They can just walk down the hallways, talk to the teachers that they run into. However, it's not representative of the population of teachers as a whole. They've only talked to one school. It could be very different at other schools. So this is an unrepresentative sample. And it's a convenience sample.
In our next scenario, a cell phone company representative wants to call to ask about customer satisfaction. Because you may choose to hang up the phone, or you might choose not to answer, this is a voluntary response sample.
In our next example, Jay is having a party. He puts a survey on Facebook to ask when it should be. Because Facebook does not force someone to respond to the survey, this is a voluntary response. Only the people who care really strongly either way-- maybe they're not available one weekend, maybe they want to have it on a Tuesday-- those are the people who are going to reply.
Our next example, Stephanie is deciding on a new office t-shirt and interviews five people from each department. This example here is representative. Stephanie makes sure to talk to people from each department. She talks to a handful of people. So this is a representative sample. This has been your tutorial on unreliable samples. Unreliable samples include both convenience samples and voluntary response samples, which are also known as self-selected samples.
