Agriculture Reference
In-Depth Information
Chapter 1
Essential Statistical Concepts, Definitions,
and Terminology
1.1
Introduction
Surveys are probably the most noticeable aspect of statistics. They are perhaps
universally criticized, and yet they continue to be widely used. If they are realized
and interpreted in an appropriate way, they are a valuable technique for gaining
information about a phenomenon.
Most surveys are developed using the same stylized sequence. Unless the survey
has been conducted before, its cycle starts by identifying information that is
required by one or more stakeholders. The delineation of the objectives is the
most important phase. Without this, the purposes of the survey will be unfocused,
and the process will be distorted from the beginning. The correct definition of the
purpose of the survey leads the ongoing development process for data collection.
When data are collected and processed, estimation and other analyses begin. At the
end of the process, the results are disseminated and evaluated by the stakeholders or
data users. Budget and costs are always relevant to the development of surveys.
National statistical offices, ministries, and private agencies must always be aware of
the trade-off between the costs and benefits of a survey.
There are two different types of surveys: a
census survey
and a
sample survey
.
They are executed in a different manner, and produce different results according to
sampling and non-sampling errors (see Chap.
5
for details). The choice of survey
often depends on how much time is available, as well as on other conditions.
Census surveys are usually conducted by public administrations. The required
information is collected at regional and country levels, for each unit (i.e., person,
household, field, factory, or shop) belonging to the target population (an aggregate
of all the units of a given type). This process is expensive and time-consuming,
because a large number of people are needed to collect and edit the information, and
appropriate policies must be formulated. However, if we are interested in the
information for each unit and the requested estimation domain is appropriate, a
complete enumeration survey is the only solution. Examples of these surveys