Types of Sampling

Sample Estimates

Statistics revolves around finding trends in very large populations. For instance, you could be trying to find spending patterns for all males ages $18-35$ in America. This is a huge number of people that you would need to study, so it just isn't feasible to obtain all of their spending habits. We would need to break the population of $18-35$ year old males into smaller samples in order to estimate the true average in the population, hence the name sample estimate.

Not any sample will do. A sample taken from a population has to accurately reflect the larger population, but this doesn't just happen by chance. There are steps that can be taken to ensure a good sample.

Sampling Techniques

Here are the main types of sampling techniques that balance ease of giving the survey with the accuracy of the information received.

Simple random sampling is exactly what is sounds like. Every member of the population is given an equal chance of being selected to be surveyed. This is an easy way to sample if you can easily gather the entire population, but is hard with larger, more spread out groups. This can also lead to some inaccuracies due to location if you're using a group as large as the American young adult population. For instance, you could give every male an equal chance of being selected, but end up with most of your sample being from New York.

Systematic sampling is done by randomly selecting a starting point and moving an equal distance to each sample after the starting point. For instance, if you had a list of every member of your school, you could randomly select one person, then also select every $4$th person after that. This is a little bit more efficient than simple random sampling by being less time consuming.

Stratified sampling is performed by splitting the population into different groups or strata before randomly sampling. For instance, we could break each person in an office into their departments before we randomly select from those sections.

Snowball sampling is a type of non-probabilistic sampling technique used when it's difficult to obtain a list of the entire population. Initial participants are selected, and then they refer the researcher to other potential participants, who in turn refer more participants, creating a "snowball" effect.

Purposive sampling is more commonly known as judgmental or selective sampling, where researchers select individuals to participate in the sample based on specific criteria relevant to the research objectives. Participants need to meet certain characteristics or experiences that are relevant to the study.

Other important terms

• Representative sample: a sample that accurately reflects the characteristics of the population from which it's drawn. Representative samples help with generalizing research findings from the sample to the larger population.
• Target population: the entire group of individuals that the researcher is interested in studying and to which they intend to generalize the findings of their research
• Sample size: the number of individuals selected for inclusion in a sample from a larger population

Goal of a Sample

We use these sampling techniques to give ourselves the best possible sample to work with. If we get a good, unbiased sample, we can assume that the data we get accurately reflects the larger population. Statistical tests, which you'll learn about in a future section, are designed to test the accuracy of our sample estimates against a possible parameter of the population.

We will use these sampling in clinical research or in market research. Sampling will become important as learn about different types of research studies: observational studies, surveys, and experiments.

(Video) Sampling Methods

by Simple Learning Pro

In this video by Simple Learning Pro, you will get a basic introduction to methods of sampling.

As the video explains, there are various types of sampling. The three this video focuses on are: simple random sampling, stratified random sampling, and multistage sampling.

Simple random sampling is when everyone has an equal chance of being chosen. Therefore, this is an unbiased sampling method. The video provides the great example that this is analogous to picking names out of a hat.

Stratified random sampling is when we divide the population into different groups based on certain features, and then complete simple random sampling amongst each group. This type of sampling can help create a representative sample.

For multistage sampling, you can think of this like multiple rounds of simple random sampling. As the video explains well, you may complete a round of simple random sampling to chose a group and then from there conduct another round of simple random sampling to chose people amongst that group. This is another way to ensure a more representative sample, especially for large populations.

In the previous lesson, we learned what statistics was. In this lesson, we will move forward into methods of sampling.

Types of Sampling

We use statistics when we are trying to understand the distribution of a set of data or determine the probability of a certain occurrence. In order to figure those things out, we must first have a set of data that we want to understand or manipulate.

But to analyze a dataset, we must first get a dataset. We get this dataset through sampling!

Sampling is when we choose a subset of information or choose a sample of a population to collect data from for analysis. For example, if we want to understand how women do in bowling compared to men, we obviously cannot compare all women's scores and all men's scores. So, we would pick a group of women and we would pick a group of men to compare the samples. This is called sampling!

There are two many types of sampling: probability sampling and non-probability sampling.

Probability Sampling

Probability sampling is when each item in a sample has the same probability of being chosen and we can calculate that probability. An example of probability sampling is when we want to select a red marble from a bag that has a red marble, a blue marble, a green marble, and a yellow marble. There's an equal chance of selecting each of the marbles, and we can calculate that probability.

However, not all probability sampling is made the same! There are a few different types of probability sampling, but we are going to focus on simple random sampling. The previous example of selecting marbles is an example of simple random sampling.

A second type of probability sampling is called stratified sampling, which is when you divide a population into subpopulations, and then select a sample from the subpopulations. For example, say you want to understand math scores across an elementary school. You can't randomly select students from different grades because you may end up with more sixth graders than second graders, which would throw off your statistics. However, if you made each grade level a subpopulation, you could then select three people from each grade level. This is called stratified sampling!

Non-Probability Sampling

Non-probability sampling is when some items in a population have no chance of being selected, and we cannot make an accurate calculation of the probability of selecting each item in a population. There are a few types of non-probability sampling. We are going to talk about two: convenience sampling and quota sampling.

Convenience sampling is when we select items from a population based on convenience. For example, if you want to figure out the most popular type of Girl Scout Cookies, you could stand on the street and survey the first 100 people as they walk by. This is a convenience sample, since we cannot make a probability estimate as to the chance of someone being surveyed.

Quota sampling is similar to stratified sampling, except we cannot determine the probability of an item being chosen. Similar to stratified sampling, we divide the population into subpopulations. However, with quote sampling, we don't randomly select items for our sample. The sample can be chosen by convenience or volunteers.

In general, real-world sampling is not perfect. We may face biases or errors when creating a sample. Check out the next lesson on sampling biases.

Let's walk through some example problems of determining the best type of sampling.

Great! Let's look at another example.

Types of Sampling

Before we get deep into types of studies in statistics, we first need to understand the basics of the types of sampling. There are several ways to choose a smaller group from a larger population.

A simple random sample gives everyone an equal chance of being chosen. For example, you could put the names of $100$ students in a hat and randomly pick $20$ names.

A systematic sample follows a pattern, such as choosing every fifth student on a list. In general, we say every $nth$ item.

A stratified sample starts by dividing the population into groups, such as different grade levels. You then choose some people from each group to be part of the sample.

A cluster sample also divides the population into groups, but you choose entire groups instead of people from each group. For example, you could randomly choose two classrooms and survey every student in those classrooms.

An easy way to remember the difference is that stratified sampling chooses some people from every group, while cluster sampling chooses everyone from some groups. These different sampling methods help researchers choose a smaller group to learn more about a larger population.