number, then certain other values are
determined based on the characteristics
of the group. The most important part of
statistical theory is sampling. This is
because in most applications, the statisti-
cian is not only interested in the charac-
teristic of the sample, but also the charac-
teristics of some much larger population.
(For information about samples and pop-
ulations, see below.)Why are populationsimportant to
statistics?
A population is the entire collection of
items—from people, animals, and plants
to street numbers and various other
things of any size—from which the statis-
tician collects data. These data are of par-
ticular concern because in most cases the
statistician is interested in describing or
drawing conclusions about the population
(also called the target population). For example, take a population of 10 cats. None of
them are identical, but certain common features between the cats can be measured,
such as color, fur length, and weight. The data collected about one of the common fea-
tures, such as the fur length of the 10 cats, would be defined as the population.What is a samplein statistics?
A sample is a generalization about a population and is represented by a group of units
selected from the population (also called a subset of the population). The sample is
meant to be representative of the population; thus, in many studies, there are many
possible samples. There are also types of samples, such as a matched sample, in which
two of the members are paired; an example would be the IQ of twins. There is often a
good reason for taking samples of a population: Most of the time, a population is too
large to study as a whole.
For example, take the “smaller” example of the above population of 10 cats. Again,
none of them are identical, but certain common features between the cats can be mea-
sured, including color, fur length, and weight. If data is collected about the fur length
of the 10 cats (the population), then if we chose only to take the cats with long fur,
that would be a sampling. Another example is the population for a study of physical
condition of all children born in the United States in the 1970’s; the sample could be
256 all children born on July 5 in any of those years.
A matched sample is a type of sampling method used
in statistics. For example, in studying population
IQs, identical twins could be paired up to measure
and compare intelligence. Photographer’s Choice/
Getty Images.