Hi, this tutorial covers surveys. So let's take a look at three research questions. Number one, do residents in Duluth, Minnesota, support the president's re-election bid? Two, are students at a local high school satisfied with the hot lunch options? And three, do colleagues in your workplace prefer shopping online or in brick and mortar stores?
So all of the following questions could be addressed by taking a survey that addresses an appropriate variable of interest. So to defined variable of interest, a variable of interest is a variable that is researched in an observational study or an experiment. So really, all of these variables are really just-- you really just have two responses. So in the first one, the variable would be do they support the re-election bid or not support? This one, are they satisfied or not satisfied? This one, do they prefer shopping online or in brick and mortar stores? So the variables of interest are all pretty simple in this.
So now let's go ahead and take a look at the definition of what a survey is. So a survey is a collection of information from persons in order to make inferences about a general population. And a survey is a common tool for performing an observational study. So surveys generally go with observational studies and not really with experiments.
There are really two types of surveys, one called a census, and one called a sample survey. So a census is a survey of an entire population. So in order to do a census, you do need that entire population, whereas a sample survey is a survey of a sample of a population. So when you have a sample, that's going to be a smaller subset of your population. So generally, sample surveys are easier to do and less time consuming. If a sample surveys is used, a sample must be selected in a way that is representative of the population of interest. So let's go back to the three questions we started with and see, is it going to be more appropriate to do a census or a sample survey?
So if we look at the population of interest in this first example, your population of interest would be simply residents in Duluth, Minnesota. So Duluth is a pretty large city, so I think it would be impractical to do a census. So we would need to take a sample and then do a sample survey. So that one is pretty clear cut. A census would not be practical there.
Two, our population of interest here is students at a local high school. I guess it depends on the size of the school and how much time you have to do the study, but if the school is significantly large, I would say a sample survey would be more practical to do here. So take a representative sample of students and simply just ask them if they are satisfied. And the third one, do colleagues in your workplace prefer shopping online or in brick and mortar stores? I would say, certainly if your workplace is small enough, a census would be very practical. If you had a larger workplace, maybe you would have to do a sample survey, but I would say if your workplace is sufficiently small or depending on how you define your workplace, I would say a census is going to be possible in most cases there.
Now, after you decide whether to do a sample survey or a census, it's important to make sure that your survey is designed very well. So to define survey design, survey design is the process of addressing the following survey issues. So first of all, what statistics should be produced, for which population, when, and with what accuracy? Those are all important things to consider when designing a survey. So to obtain good information about a variable interest, the survey must be thoughtfully designed. So that has been your tutorial on surveys. Thanks for watching.