In this video I explain populations and samples, including random sampling, opportunity samples, stratified sampling, and issues related to representativeness.
Don’t forget to subscribe to the channel to see future videos! Have questions or topics you’d like to see covered in a future video? Let me know by commenting or sending me an email!
Viewers in China can find this video here: http://v.youku.com/v_show/id_XMTg0MTg3MjY2NA
Video transcript:
Hi, I’m Michael Corayer and this is Psych Exam Review. In this video I’m going to talk about populations and samples so we’ll start with the term population. So what does the population refer to? The population is everyone that we want to know about. So it’s the group of people that we want to learn about it, the group that we want to study. So if I want to know about college students then anyone who’s a college student would be part of that population.
Of course, that’s a very big population, there’s a lot of college students so I’m not going to be able to study all of them. Usually the populations we want to study are quite large. We want to know about all college students or all people in general and we can’t study everyone.
So we can’t study the population, instead we have to choose some people from that population that we actually study. And those people are called the sample. So the sample are the people that we actually study. So if I want to know about college students, any college student is part of the population but the college students that I actually bring in to the lab and study would be my sample.
So how do we go about choosing the sample from the population? Well, in ideal situations we don’t really choose, we make it random. We select a random sample. This would mean that anyone from the population has equal chance of being in this study. The odds are the same for all members of the population and we randomly choose a sample.
This is ideal but you can already sort of guess this is going to be very hard to do. If I want to know about all college students how am I going to random select from all college students all over the world? And then fly them to my lab for this study? That’s going to be impossible, I’m not going to be able to do that.
We almost never have random samples. The only time you can really have a random sample is if you have a very small population that you want to know about. Then you might be able to randomly select. Most of the time we want to know about big populations.
We want to know about all college students all people in general. So we can’t randomly select and have those be our participants. So what do we do instead? Instead we have what’s called an opportunity sample or you may also see it called a convenience sample.
This is just saying we studied the people we had the opportunity to study or it was convenient for us to study. So if I want to know about college students, that my population. I might choose a sample that comes from a local college.
That’s not random but it’s the people who were available to me. I had access to these college students and I wanted to know about college students said that’s how they ended up being in my study.
I might also choose to have a stratified sample. What a stratified sample refers to is that I think there are certain features of the population that are important certain demographics and I want to make sure that my sample has those same demographics. I might say OK-maybe gender is important for the thing I’m studying, it might be relevant, so if in the population there’s about half females and half males then I want to make sure my sample also has that. Now that’s not random, that wouldn’t be guaranteed to happen by random chance but I might think it’s important, so I choose to stratify my sample that way.
Similarly I might say ok I want to know about U.S. citizens and I see certain ethnic groups, maybe I think it’s relevant for my study so I might choose to say OK, I want this percentage of this ethnic group in my study and I want this percentage of some other ethnic group in my study and so I can sort of go through and stratify my same to ensure that the sample kind of looks like the population and the reason for doing this is that we want to have representativeness.
All this term refers to is the idea that the sample represents the population. So representativeness is the idea that the group of people I’m studying are able to stand in for, or represent the entire group that I want to know about.
Now this is a very common criticism in psychology. It’s often the case that we have a sample that’s said to be nonrepresentative. It doesn’t really represent the population.
Why is this such a common criticism? Well, the reason is fairly simple. Most research happens at universities and who are the people that they’re able to study in these universities? Mostly it’s going to be undergraduates. It’s going to be freshmen and sophomores, often psychology majors, who we’re able to get to come to the psych lab and do some experiments. And those people probably don’t represent all people.
College students are different from the rest of the population in certain ways and that might matter. So they tend to younger than average they tend to be better educated than average they tend to be wealthier than average, and these traits might be influencing the results that we’re getting.
So often we have nonrepresentative samples but this is not the end of the world, we don’t give up, we don’t say “ok I can’t get a representative sample for all people, therefore what’s the point.”
So there’s three reasons why this is ok. The first reason is fairly simple, we say that OK you’ve got limited data if you have a nonrepresentative sample but limited data is better than no data.
We’d rather have some information about a nonrepresentative sample, we might find an interesting effect here that’s better than not having any information at all. Now we have to be careful because when we have limited data, we have to remember our conclusions must also be limited.
If I only study people from this particular sample then I need to make conclusions only about this particular sample. I can’t necessarily make conclusions about the whole population.
The second reason why nonrepresentative samples can be ok is that we can compare them. If researchers share their information with one another then we can say, alright I did my study with a nonrepresentative sample and you did your study with a different nonrepresentative sample, and somebody else studied the same thing with another nonrepresentative sample, we can sit down and look at these three sets of data and see how they compare to one another.
If we find a similar effect among all three that would imply that maybe this effect is general, maybe this does apply to the whole population. So, by comparing studies with groups of nonrepresentative samples, we’re able to get a better picture, and maybe we can make some general conclusions
The third reason why nonrepresentativeness is not the end of the world is that sometimes we assume generality. Sometimes we look at the results from a nonrepresentative sample and we simply assume that they apply to everyone else.
Now you might say “Mike that sounds like terrible science. You’re not supposed to do that” you shouldn’t be looking at a nonrepresentative sample and drawing large conclusions. And that’s true but sometimes it’s OK.
So when is it OK? Well let’s say that I have a new drug treatment and I give it to a bunch of people and they all get worse. They all have side effects and their symptoms worsen. I don’t need to then say “well this was a nonrepresentative sample I’d better give this drug to a lot more people.” That is probably not a good idea because in this case we’re going to assume generality. We’re going to say OK even though these people were a different ethnic group or even though they had a different educational background or even they were a different gender than you, we’re going to assume that these bad effects are going to happen to you too.
We don’t really know, but we’re going to assume it to be on the safe side. We actually do this in the extreme in some cases, so if I give a bunch of rats a drug and they all die, you probably don’t wanna try this drug. Now you would certainly be correct in saying “well rats are not a representative sample for the population of humans” but I’m willing to assume generality in this case.
So that’s populations and samples, I hope you found this helpful, if so please like the video and subscribe to the channel for more. Thanks for watching!
