A census is a collection of information from all units in the population. The Census of Population and Housing is a statistical collection that aims to accurately measure the number of persons in Australia on Census night, their key characteristics and the dwellings in which they live.
An estimate is an inference for the target population using information obtained from a sample of the population.
A part of a population selected for the purpose of studying certain characteristics of an entire population of interest. A sample is used to represent the population. You can often get a response form a sample where it would not be possible to get a response form every member of the population.
The sample size is the number of units, including persons, households, businesses and schools etc, being surveyed. In general, the larger the sample size, the smaller the sampling error.
In a random sample, all units in the target population have an equal chance of selection.
Simple random sample
All members of the sample are chosen at random and have the same chance of being in the sample.
A Tattslotto draw is a good example of simple random sampling. A sample of six numbers is randomly generated from a population of 45 with each number having an equal chance of being selected.
Systematic random sample
The first member of the sample is chosen at random then the other members of the sample are taken at intervals.
Stratified random sample
Relevant subgroups form within the population are identified and random samples are selected from within each strata.
For example, a school has 24 Year 7 students. Eight of the students are 11 years old, twelve are 12 years old and four are 13 years old. The strata are the ages of students.
To take a stratified sample, select one quarter of the students in each age group – for instance, two students form the 11-year-olds, three students from the 12-year-olds and ones student who is 13 years old.
In this example, the strata are proportionally represented; however, this will not always be the case. The important thing to remember is to take a random sample from each strata.
In a non-random sample, the chance of a member of the population being in the sample is unknown. The accuracy of the sample in representing the population is unknown.
This is a type of stratified sampling in which selection within the strata is non-random. Quota sampling requires setting a number of participants to include in a survey – usually a proportion of the population.
Take the example of Year 7 students from the stratified random sample above who are in strata of age groups. Unlike stratified random sampling where participants are selected at random, participants in a quota sample are selected to fill the quotas.
For instance, the first 15 twelve-year-old Year 7 students to arrive at school on any given day may be selected. However, this sample may not be representative of all twelve-year-olds in Year 7.
In a convenience sample, participants are selected by how easy it is to reach them.
For example, the first ten students to walk through the front gates of the school is an easy sample to take. Convenience sampling does not produce a representative sample of the population because people or things that can be reached easily and conveniently are likely to be different to those that are harder to reach.
This is where participants volunteer to be part of the survey.
Phone-in sampling is a common method of volunteer sampling used by television and radio stations to measure public opinion. People are asked to telephone or SMS their vote on a particular issue by a certain time. There is no control over how many people vote.
There are two main problems with this type of sampling. Firstly, there is no limit to the number of times a person can vote, and secondly, those not interested in voting will not be included in the sample. People who don't call in may have different views to the people who are calling in.
Additionally, only those watching television or listening to the radio know that there is a survey taking place.
As such, volunteer sampling is unlikely to produce a sample that accurately represents the population.
Sampling error is the difference between an estimate derived from a sample survey and the true value that would result if a census of the whole population was taken.
Non-sampling errors are not caused by sampling methodology. They can be made by participants and interviewers when the questionnaire is being filled in. or they can happen when the questionnaire is being processed.