ABSTRACT
We have seen in earlier chapters that data are usually collected in order to gain
information
about, and better understand, some particular population or process. In some circumstances it may be appropriate to attempt to collect data about every member of the population (a
census
), but this is often impractical because many of the populations we wish to study are impossibly large, or the sampling method is destructive. It is therefore more usual to collect information from only a sample of the population, as described in the previous chapter. In practice, samples of data may be collected in many ways, both random and nonrandom, but however the sample is taken, it provides data for only a portion of the entire population. So it is unlikely that a sample will give precise information about the population. Usually error will arise from using a sample to estimate population characteristics. With a random sample, this error is known as the
sampling error
. With a nonrandom sample, however, estimates may also be
biased,
in that a systematic error may arise because judgement has been used in the sampling process.
