**Representative Sample Data**

The total group you want to find out about is called the **population**…but you might not to be able to get data for every person or thing in the population…so you take a **sample** instead.

The sample should be representative which means:

- It’s a random sample
- It’s large enough to represent the population

You might be asked whether a sample is biased or not. Think about:

- When, where and how the sample is taken
- How many members are in it

**Simple Random Sampling**

Simple random sampling is one way to to get a random sample. To do this you would have to:

- Assign a number to every member of the population
- Create a list of random numbers (using a computer or picking numbers out of a hat etc)
- Match the random numbers to the population and then take the sample from these members of the population

**Stratified Sampling**

“Strata” just means layer. A stratified sample is made up of different “layers” of the population. For example you might split up a group based on age.

In such cases each group’s share of the sample is calculated based on its share of the population.

To find the proportion of the population that the group represents:

Proportion Represented = Number In Group / Total Population

Then multiply this by the sample size to find the number from that group in the sample.

All this sounds confusing right? It’s not quite as bad as it sounds though – as this example shows:**Example:** In a school there are 104 children in year 9, 130 in year 10 and 95 in year 4. A sample of 40 children stratified by year is taken. How many are from year 10?

Proportion represented = 130 / (104 + 130 + 95) = 130/329

Number of year 10 in sample = 40 x (130/329) = 15.80547….

As you can’t have less than a whole person in the sample, round to the nearest whole number.

So the answer is 16.