Statistics - sampling methods
certain cereal manufacturer wants to carry out a survey in the county Middlesex
to determine the most popular breakfast cereal among school kids. Asking each
and every child about their choice is next to impossible; the number of kids
runs into millions.
It is what we call in statistics, a vast population. When we
are faced with a situation of this kind, it makes sense to concentrate on a
manageable group of kids to get a true picture of their needs.
This small group
is called a sample and the
process is known as sampling.
be very careful about this small group of individuals – sample.
At the very
beginning it must fulfil the following requirements:
- The sample must be small enough to manage
- The sample must be large enough to represent the views of the whole population – in
this case all the kids in the county Middlesex
- When picking up individuals, bias should be avoided – skin colour, race, religion,
personal views of the interviewer towards the individuals to be selected. In
other words, we must ensure that each member of the population has an equal
chance of being selected.
With setting our sight on these points, the
next task is to number the whole population in a certain way; the letters of the
first names and surnames can be considered in this numbering process. Each and
every member of the population now has a number.
size of the sample is normally between 5%-10% of the population.
There are 600 students in Heathfield Grammar School. We want to carry out a survey to find
out how they travel to school every morning. - the sample size = 10 % of 600 = 60
population – in this case the kids in Heathfield Grammar School – are numbered
in the following way:
- Melani Yanks
- Nigel Zamir
- Owen Bright
- Raymond Hill
- Tim Sinclair
Now we have to choose 60 members from the above list. It is the process of
sampling. This can be done in three different ways:
- Random sampling
- Stratified sampling
- Selective sampling
Generate 60 random numbers by any of the
Throw the three spinners and take the number on display in order – 599, 400, 398 etc. Spin
all three 60 times. You can
get sixty random numbers by any one of these three methods. It is random
In this method, we divide the population into smaller units: we can divide them on the
basis of ethnicity of the members – just one way of grouping the population.
Then the members for the sample are chosen from these individuals, of course in
proportion to the size of each unit.
The population of Heathfield Grammar School is grouped in the following way:
The members for the sample in proportion to the size of the groups are chosen as follows:
|White Caucasians||10% of 150 = 15|
|Sikhs ||10% of 180 = 18|
|Muslims ||10% of 120 = 12|
|Afro-Caribbean||10% of 90 = 9|
|Other||10% of 60 = 6|
How do you choose these members from each group? You can use random sampling for that
In this method, a random number is generated by some means – calculator, spinner or a
simple computer programme.
If the number turns out to be 5, then every 5th member must be picked up for the sample.