Population and samples
In a nutshell
When carrying out a statistical survey, different types and amounts of data can be collected. The population is the whole set of the items or individuals that are of interest. A census measures every member of the population. A sample measures only a subset of the population, and these measures are used to estimate measures for the population. There are advantages and disadvantages of using a census and using a sample.
Definitions
There are various definitions you need to know when working with data.
| NAME
| DEFINITION
|
| Population | The whole set of items or individuals of interest. |
| Census | A measurement or observation from every member of the population. |
| Sample | A selection of measurements or observations from a subset of the population. |
| Sampling units | The individual items or members of the population. |
| Sampling frame | A list of all the sampling units in the population. |
| Parameter | A population parameter is a value that can be calculated using the whole population. It summarises everything known about the population. |
| Statistic | A value that can be calculated from a sample. A statistic can be used to estimate a parameter. |
Census and sample
If data is collected from the entire sampling frame, i.e. the whole population, it is called a census. A census can be time consuming and expensive, however the information gathered would be representative of the population. It is usually quicker, easier and cheaper to take a sample, but it is important to ensure that the sampling method creates a sample which is representative of the population.
|
| ADVANTAGES
| DISADVANTAGES
|
| Census | - It gives a completely accurate result.
| - It is time consuming and expensive.
- Cannot be used when the testing process destroys the items in the population.
- It is difficult to process large quantities of data.
|
| Sample | - The process is less time consuming and cheaper.
- Less data to process.
| - It may not give a completely accurate result.
- The sample may not be large enough to represent the different subgroups of the population.
|
Example 1
A battery manufacturer wants to test the lifetime of the batteries they make by seeing how long they last. Should they use a census or a sample? Given a reason for your answer.
A census would require all the batteries to be used and they would consequently have no more batteries left to sell.
Therefore, the manufacturer should use a sample.
