Survey Data Collection and Processing - Sampling Spatial Units for Agricultural Surveys - page 229

9.5 Quality Assurance
Validation is the process that evaluates if the collected data can be considered
consistent with the purposes for which they were gathered. Therefore, validation
activities can be defined as the set of operations through which we compare the
planned quality targets and the achieved results.
From this definition, it follows that the quality targets should be set at the design
stage, and must be expressed in measurable terms. We must develop procedures for
measuring the quality of the collected data to assess if the targets have been
achieved.
Validation has two goals: to assess whether the quality of the data is sufficient to
allow the dissemination of statistical information to users, and to identify the
sources of error. A consequence of the second goal is that validation can determine
relevant changes to the production process that reduce the effects of errors on
subsequent surveys.
The nature and intensity of the analyses performed in these two cases are
different. In fact, in the first case, validation should be conducted within an
appropriate amount of time so that we can exclude data. In the second case,
however, we can take more time. These validation procedures can be more ambi-
tious and assess the impact that certain sources of error may have on the accuracy of
the collected data.
These considerations lead to a quite general list of validation measures.
The first is to facilitate user judgments, adequately documenting the quality
targets, definitions, and processes. This action is based on the importance of a data
user understanding the validation procedures, because they must assess the useful-
ness of certain statistical information. In fact, the user must assess whether the data
can be considered valid for its own purposes. Additionally, many qualitative
evaluations can be made using knowledge of the characteristics of the production
process, regarding possible interpretations of the available information. To this end,
the checklist is a useful tool for evaluating the definition and operational aspects of
a survey. The checklist facilitates comparisons by making the survey documenta-
tion uniform. Data users need to know the accuracy and reliability of a land cover/
land use survey so that they can assess whether the data quality agrees with their
specific needs.
The LUCAS 2009 survey (Eurostat 2009a ) can be used as a practical example.
The technical document available for data users contains a list of the main checks
used in the data production process. It includes information regarding the data
import and export, the surveyor, the point (how it is fixed and how its accuracy is
automatically checked using the GPS coordinates and the observation distance
entered by the surveyor), the start and end times, the type of observation, the
latitude/longitude, the distance and direction, the elevation, the description of the
path to the point, the land cover and land use, the area, the height of mature trees,
the width of a feature, the land management, the water management, the source, the
type of irrigation and delivery system, and other characteristics. The document also