First 315 words of the document:
In this technique every member of the total population has an equal chance of being selected and the
selection of one member doesn't affect the probability of selection of another member. A sample
can be obtained using random numbers. This is the most accurate method as it has no bias. There are
3 alternative ways of using random numbers to sample areal distributions: random point, random line
and random area.
Point a grid is superimposed over the area of the map to be sampled. Points are then
identified using random number table and plotted on the map.
Line random numbers are used to obtain 2 end points which are then joined by a line, which
uses the same 8 random points.
Area areas of constant size are obtained using random numbers.
This is a sample in which values are selected in a regular way, e.g. choosing every 10th person on a
list. It can be operated using individual points, lines or areas. This is quicker and easier to use,
although some bias or selection is involved. It may either overstress or miss an underlying pattern.
Point this can show changes over distance.
Line may be used to choose a series of equally spaced transects across an area of land.
Area used for land-use sampling, to show change with distance or through time.
In this method the population is divided into categories and sampled within each. Once the groups
have been determined, they can be sampled either randomly or systematically.
Stratified systematic very useful, in many situations, e.g. political opinion polls.
Stratified random can be used to cover a wide range of data, both in interviewing and in
fieldwork and map work.