One of the main sources of inconsistencies in databases is the lack of homogeneous
criteria for recording data. In this document you will find advice on how to mitigate these
inconsistencies and give your database more coherence, uniformity and reliability.
Communicate in a clear and explicit manner
If you want your data to be reliable and trustworthy, you need to explain how you collected
and recorded them in your database. It is not about revealing or exposing your sources
(victims, witnesses, etc.), but about detailing how the study was made and how did you
build the indicators. This process, as we will see later on this document, is complex and
determines the results we will get. If we want to be reliable we need to reference the
standards or laws that we are using and explain where do our specific indicators come
from.
Avoid ambiguity
Polysemic, synonymic and ambiguous terms create inconsistencies when working with
data. There might be words or sentences with various meanings (polysemy) or, by contrary,
several words or phrases to name the same thing (synonymy).
For example, we often say indistinctly "right to asylum" and "right to refuge" to talk about
the protection and assistance that the state must guarantee for everyone forced to flee his
or her country because of persecution, war or violence. We use two different names to
refer to the same thing or very similar things, as in some contexts there might be
differences between them. If we store data related to this right in our database but we do
not select the preferred term, it is very likely that inconsistencies will appear. Similarly, if we
use polysemic words we should make clear which meaning we are referring to.
Also, there are some terms frequently used in our daily life or in media that do not have an
empirical correlation. Selecting the most adequate terminology in a database is a complex
process. For instance, the use of the term “political prisoner” in the media might have
different meanings that do not always match the specific way in which cases are registered
in the database.
Choosing the appropriate terminology for a database is a complex process. In the HR field,
the glossaries created by specialised organisations constitute a good reference that we can
take as a starting point.
2