Some harassment used coded or indirect
language that was hard to identify: In
Iraq, some forms of sexual harassment
and abuse use indirect expressions or
local references that are not easy to
detect through standard keyword
searches. Comment threads also
sometimes move away from the original
topic, which affects how comments
should be understood.
INSM developed a detailed guide for identifying harmful
language that included Iraqi-specific expressions and cultural
references. This guide was updated regularly, and the research
team took into account the full context of each comment
thread, not just individual comments.
The data came from different types of
content, so direct comparisons are not
always possible: The level and type of
harmful speech varied depending on the
content, for example, general posts
versus posts specifically about the
elections. This makes it difficult to
compare different parts of the dataset
directly.
INSM separated the data by content type and by election
phase (before, during, and after the vote). This allowed for fair
comparisons within each group. The variation between groups
is also treated as a meaningful finding in itself.
The total number of women candidates
was reported differently across sources:
Different official and unofficial sources
gave slightly different numbers, ranging
from 2,235 to 2,250 women candidates.
This made it difficult to rely on one exact
figure with full confidence.
INSM checked multiple sources, including official electoral
commission records and civil society data. The most
commonly reported figure (2,248 women candidates) was
used, and the report clearly notes that small differences exist
across sources.
79