Using the Wrong Sampling Frame

This lesson will teach you about about the problem of using the wrong sampling frame, which is an error that we can make when we sample. We’ll cover:

1. Sampling Frames
2. Using the Wrong Sample Frame

1. Sampling Frames

A nice way to think about sampling is with the analogy of a pot of soup. When you want to select a representative sample, that is like dipping your ladle into the pot and getting a little bit of every ingredient from the soup. However, when you taste what you’ve served out of the pot, it is possible for you to get the wrong idea of what's in the soup. That can affect your judgment.

So one problem might be choosing the wrong sampling frame.

Sampling Frame

The entire set of individuals from which we will sample. Ideally, this would be the entire population, but logistics often prevent us from obtaining every individual in the population.

So a sampling frame is the entire set from which we're going to sample. Now you might think to yourself, well isn't that the entire population? Don't want to generalize our findings to the population? And, yes, ideally that would be the case. But you don't always have the entire list of population at your fingertips.

If you are doing a sample for, say, a school, then you would have the entire population at our fingertips. All you’d have to do is check the attendance to obtain a list of every student in that particular population.

2. Using the Wrong Sampling Frame

When sampling, you don't always have the luxury of being able to access a list of the entire population.

Suppose that you are doing a sample of a city and you choose to do a telephone survey. The sampling frame, in that case, is all the individuals in the population of the city that have telephones.

Now you might wonder is that a big deal. Most everyone has a telephone. However, the people who don’t have phones cannot be included in your sample at all, even if they wanted to be.

Using this sampling frame, as opposed to the entire population, is only a problem if you feel that the excluded subset, that is the people without phones, feel substantially different about the topic that we're going to be asking them about than the people within the sampling frame.

For instance, maybe you are doing a phone survey about telephones. It might be that people without telephones feel differently about phones than people who have them, since they don't have or have chosen not to have telephones.

The only way that circumstance could result in bias is if you thought that there was a huge difference between the people that are part of the sampling frame and the people that aren't part of the sampling frame.

Now the question is, how does this affect the soup? How does this affect your ability to draw inferences about the population?

The analogy here is that, in this case, this sampling frame would be like setting the soup on the stove and cooking it up but leaving out some some ingredients that were called for in the recipe.

• When you take your sample spoonful of soup, you might get an accurate representation of what is in the pot.
• However, what is in the pot is not representative of the entirety of what should have been in there, based on the recipe.

Therefore, when you sample, you won't get an accurate reflection of what the soup should have tasted like had you included every ingredient.

It's important to choose the correct sampling frame and to use it if you can. If the entire population is available to be selected for the sample, then use the population as your sampling frame and sample from there. If the entire population not available to be selected, then the conclusions that you can draw might be a little bit limited. This might be the limits of using the wrong sampling frame. Thank you and good luck!