1 Sampling
Why not just use all the data?
It is rare that a data set is complete, some people do not respond to surveys, sometimes there are just too many individuals or items about which to collect information. Sometimes the item is not usable once the information is collected, for instance, if you were collecting information about the life length of batteries. Therefore, it is not always possible to use the full data set; this is where sampling comes in.
Sample (verb)
To take a sample or samples of; to judge of the quality of (a thing) by a sample or specimen; to obtain a representative experience of. (OED, 2024)
Notice the word ‘representative’ in the definition. The sample taken should represent the rest of the population. All the data is known as the population and a sample (noun) is a subset of the population, containing one or more members. You can take more than one sample from the same population and doing so can strengthen your analysis of the population data.
The population need not be a population of people, it could be chocolate bars, light bulbs, batteries, fish, cities, etc.
Activity 1 Representative sample?
One pupil in a school could be considered a sample of the population of all pupils at the school. And the pupils in one class in the school could be considered a sample of all the pupils at that school.
Think about how representative of the whole school population these two samples would be and jot down your thoughts. Can you think of a way to choose a more representative sample of pupils in a school?
Comment
One item in a population is unlikely to be very representative of the whole population however the size of the population is important. If there were only 5 people, say, in the whole population then selecting one person might be sufficient to represent those 5 people. However, in a school situation where there are likely to be considerably more pupils, one pupil is not at all a representative sample.
Whilst the larger sample size here has the potential to make this sample more representative of the whole population, the fact that all the pupils are in the same class reduces the amount of variation in their characteristics. For instance, if the purpose of the investigation were to create a general classroom design to be used across the whole school based on the pupils in this class, the measurements taken would not necessarily be relevant across the whole pupil population.
One way to take a more representative sample would be to randomly select two pupils from each class.