Sampling Flashcards
Parent Population/ Population
Parent Population/ Population (same term) – The overarching set of values. (This can be finite, such as all professional golf players, or infinite, for example the points where a dart can land on a dart board.)
parameters
A parent population is usually described in terms of its parameters, such as its mean (μ). By convention, Greek letters are used to denote these parameters and Roman letters are used to denote the equivalent sample values. I DO NOT UNDERSTAND WHAT TF THIS MEANS- ASK MR GRANT. (any of it really).
Sampling Frame
A list (or other representation) of the items available to be sampled.
In many situations no sampling frame exists, nor is it possible to devise one, for example for the cod in the North Atlantic
Sampling Fraction
The proportion of the population that is actually sampled.
Census
a 100%-of-population sample
Sampling Error
An estimate of a parameter derived from sample data will in general differ from its true value. The difference is called the sampling error.
To reduce the sampling error, you want your sample to be as representative of the parent population as you can make it.
Bias
A systematic error.
Simple random sampling
In a simple random sampling procedure, every possible sample of a given size is equally likely to be selected.
Simple random sampling is fine when you can do it, but you must have a sampling frame. The selection of items within the frame is often done using random numbers; these can be generated using a calculator or computer, or you can use tables of random numbers.
Strata
Strata- Sub-group of a population.
EG. Different types of subgroups which you might expect to have different voting patterns: low, medium, and high income groups; urban, suburban, and rural dwellers; young, middle-aged, and elderly voters; men and women; and so on.
Proportional Stratified Sampling
the procedure when the numbers sampled in the various strata are proportional to the size of their populations.
The selection of the items to be sampled within each stratum is usually done by simple random sampling.
Cluster sampling
Cluster sampling also starts with sub-groups, or strata, of the population, but in this case the items are chosen from one or several of the sub-groups.
Clusters
the subgroups chosen to represent the population in cluster sampling.
The sub-groups are now called clusters. It is important that each cluster should be reasonably representative of the entire population.
If, for example, you were asked to investigate the incidence of a particular parasite in the puffin population of Northern Europe, it would be impossible to use simple random sampling. Rather you would select a number of sites and then catch some puffins at each place. This is cluster sampling. Instead of selection from the whole population you are choosing from a limited number of clusters.
Systematic sampling
Systematic sampling is a method of choosing individuals from a sampling frame. If you were surveying telephone subscribers, you might select a number at random, say 66, and then sample the 66th name on every page of the directory.
Systematic sampling is particularly useful if you want to sample from items listed in a spreadsheet. (I think it just means that it’s easier because they are numbered).
When using systematic sampling you have to beware of any cyclic patterns within the frame.
For example, a school list which is made up class by class, each of exactly 25 children, in order of merit, means that numbers 1, 26, 51, 76, 101, . . ., in the frame are those at the top of their classes. If you sample every 50th child starting with number 26, you will conclude that the children in the school are very bright.
Quota sampling
Quota sampling is a method often used by companies employing people to carry out opinion polls. An interviewer’s quota is usually specified in stratified terms: for example, how many males and how many females, etc. The choice of who is chosen to be sampled is then left up to the interviewer and so is definitely nonrandom.
Opportunity sampling
As its name suggests, opportunity sampling is used when circumstances make a sample readily available.
As an example, the delegates at a conference of hospital doctors are used as a sample to investigate the opinions of hospital doctors in general on a particular issue.
This can obviously bias the results and opportunity sampling is often viewed as the weakest form of sampling. However, it can be useful for social scientists who want to study behaviours of particular groups of people, such as criminals, where research will lead to individual case studies rather than results which are applied to the whole population. Opportunity sampling can also be useful for an initial pilot study before a wider investigation is carried out.