Page 306 - CyberSecurityEssentialsEnglish

bibliographic data. Do not treat these as one thing; always ask whether the tool is searching
a trusted index, summarizing files you attached, or generating a probabilistic answer.

The strategy begins by breaking the question into core concepts. For each concept we write
synonyms, abbreviations, spelling variants, older and newer terms, and Arabic and English

names where needed. We then use OR to gather alternatives within a concept, AND to link
concepts, and NOT with great caution because it may remove suitable studies. Quotation

marks, title or abstract field searching, and truncation are used according to the features of
each database.

Controlled vocabulary, such as subject headings or specialist term lists, helps find papers
even when authors use different words. But relying on it alone may miss recent topics whose

terminology has not settled, so it is usually combined with free-text terms. A tool can
propose synonyms, but they must be reviewed; it may merge two close concepts or offer a
translation that is not current in the field.

Testing the strategy is a necessary step. Identify a small number of studies known to be

relevant, then check whether the query finds them. If they do not appear, some concept,
synonym, or limit needs adjusting. Also inspect the first few dozen results: is the noise high?
Is one broad concept dominating? The goal is not the largest number of results, but a balance

between sensitivity in finding what is relevant and precision in reducing what is not.

Iterative searching does not mean changing the terms after seeing what agrees with your
view. You may refine a query after learning new terminology, but the versions, the dates,

and the reason for the change must be kept. In systematic reviews, the researcher records
the database name, the platform, the date of the search, the full query, and the number of
records. These details allow the search to be re-run and the review updated later.

Citation chaining adds two important routes: searching backwards through a paper’s
reference list, and forwards through the works that have cited it. You can also search the
output of key authors, similar studies, and reports that cited a finding. AI may help suggest

a network, but it is the citation services, the publisher, and the bibliographic data that
establish the links.

Searching in Arabic needs more than translating an English query. Terminology may differ
between countries, several spellings may exist, and official reports may use a term different

from academic articles. It helps to build a bilingual term table, test the words on university
and government sites and in Arabic journals, and record which term was used in each source.
We must be alert to a bias toward what is easy to find. Open content, or content that ranks

highly, is not necessarily the strongest evidence. Studies with negative results, or published
in other languages, or in unindexed reports, may be less visible. The researcher therefore
describes the limits of coverage rather than claiming to have found everything.

305