Sampling (statistics)

Sometimes we cannot check every single thing.

It might take too much time or money. Instead, we pick a small group to study. This small group is called a sample.

A good sample should look like the whole group. This helps us learn about everyone or everything. We can use a sample to check things like weight or color. It can even help us learn about stars in space. Sampling is a smart way to find out facts quickly.

Sometimes, we cannot check every single thing in a group.

It might take too much time or money. We might even find it is impossible. For example, we cannot measure every star in the sky. Instead, we pick a small group to study. This small group is called a sample.

A good sample should reflect the whole group. This is called being representative.

There are many ways to pick a sample. In a simple random sample, every part has an equal chance to be picked.

You can also use stratified sampling. This means you split the group into parts first. This helps make sure the sample is fair.

If a sample is not fair, it can lead to mistakes. In 1936, a big study in the US failed. It used magazine lists to pick people. But those lists mostly had Republicans on them. This made the sample biased, or not fair. Because of this, the study gave the wrong answer. Today, we use math to help make samples better. This helps us learn about the world quickly and accurately.

Imagine you want to know the average height of every person in a huge city. It would be nearly impossible to find every single person and measure them. It would cost too much money and take too much time. Instead, you can pick a smaller group to study. This smaller group is called a sample.

If you pick the group carefully, it can represent the whole city. This process is called sampling. It helps us understand big groups without checking every single member.

Scientists and businesses use this to learn about things like stars or medicine.

To make a sample work, you need a way to find people. This is called a sampling frame. A good frame is a list where you can find every person. For example, a phone book or an electoral register can be a frame.

In a probability sample, every person has a known chance of being picked. This makes the results much more reliable. You can even use weights to adjust the data. This helps make sure the sample truly reflects the whole population. If a person has a lower chance of being picked, you can count them more to stay fair.

People have used sampling ideas for a very long time. The Bible mentions using lots to pick items. In 1786, Pierre Simon Laplace used a sample to estimate the population of France. He even used math to figure out how much error might be in his guess. Later, in the 1870s, Alexander Ivanovich Chuprov brought sample surveys to Imperial Russia. These early thinkers helped turn sampling into a real science. They showed how small groups could tell us big truths.

Their work laid the foundation for how we study the world today.

However, sampling can go wrong if the group is not fair. This is called bias. In 1936, a study in the US tried to predict an election. They asked over two million people from magazine lists and phone books. They thought a huge sample would be perfect. But the lists were biased toward Republicans. Because the sample did not represent everyone, the prediction was wrong.

This shows that a large sample is not always a good sample. You must be careful about who is included in your list.

Today, sampling is used in many important places. In Singapore, officials use sample counts during elections. This helps stop rumors and misinformation. It gives a quick idea of the results with a small margin of error.

Businesses also use it to check if their products are good enough to sell. Researchers might even study rats to learn more about human health. Whether it is checking the quality of copper or studying penguins, sampling helps us see the big picture. It turns a hard job into a smart one.

Sampling is a vital tool used in statistics, quality assurance, and survey methodology. It involves selecting a subset of individuals from a larger group, known as a statistical population. This subset is called a sample. The primary goal is for the sample to reflect the characteristics of the entire population. Measuring every single member of a population is often impossible or impractical. For example, it is impossible to measure the size of every star in the universe. Sampling provides faster data collection and lower costs. It allows researchers to gain insights when studying a whole population is infeasible.

To begin sampling, a researcher must first define the population. A population includes all people or items with the specific characteristics being studied. Sometimes this definition is simple, such as a batch of manufactured material. In these cases, the batch is the population, and a manufacturer uses acceptance sampling to see if the lot meets specifications. Other times, populations are defined by time or space. A study on endangered penguins might look at hunting grounds over a specific period. A researcher might even study a "superpopulation." This is a larger group that does not yet exist, such as everyone in a country who might one day use a new medical program.

A successful sample requires a sampling frame. A frame is a way to identify every element in a population so they can be selected. The most straightforward frame is a list, such as an electoral register or a telephone directory. Without a proper frame, it is difficult to ensure every member has a chance to be included. For instance, there is no way to identify every rat in existence. However, a list of names provides a clear way to pick individuals for a survey. A good frame helps researchers avoid leaving certain groups out entirely.

There are two main types of sampling: probability and nonprobability sampling. In probability sampling, every unit in the population has a known chance of being selected. This allows statisticians to use mathematical theory to produce unbiased estimates. One method is simple random sampling, where every subset of the frame has an equal chance of selection.

Other methods include systematic sampling, stratified sampling, and cluster sampling.

In stratified sampling, the population is divided into groups to ensure representation. Researchers can also apply weights to the data to adjust for the specific design used.

Nonprobability sampling occurs when some elements have no chance of being selected. In these cases, the probability of selection cannot be accurately determined. This often happens through convenience sampling, where researchers pick whoever is easiest to reach. For example, interviewing the first person who answers a door is nonprobability sampling. This is because certain people, like those who stay home more often, are more likely to be picked. This creates exclusion bias, which limits how much information the sample can provide. Because the selection is not random, researchers cannot easily estimate sampling errors.

The history of sampling includes both great discoveries and famous mistakes. In 1786, Pierre Simon Laplace used a sample to estimate the population of France. He used a ratio estimator and Bayes' theorem to compute probabilistic estimates of error. In the 1870s, Alexander Ivanovich Chuprov introduced sample surveys to Imperial Russia. However, sampling can fail if the frame is biased. In 1936, the Literary Digest tried to predict a US presidential election. They used a massive sample of over two million people from magazine lists and phone books. Because these lists were biased toward Republicans, the prediction was deeply flawed.

Today, sampling remains essential across many different fields. In Singapore, officials use sample counts during elections to reduce misinformation. These counts provide an indicative result with a 4% margin of error at a 95% confidence interval. In business, companies use sampling to check the electrical conductivity of materials like copper. Even in casinos, people like Joseph Jagger have used sampling to study the behavior of roulette wheels. Whether studying human health through rats or checking product quality, sampling turns the impossible task of measuring everything into a manageable science.