Name the unit before reading the total
A candidate count may refer to people, ballot entries or participations in separate contests. A person can appear in more than one electoral context, so adding entries does not automatically produce a count of unique individuals. Verified identity links are needed before records can responsibly be treated as the same person.
The same care applies to institutional records. An elected outcome, a seat and a current member are not necessarily the same unit. A useful label tells readers what a number measures rather than relying on the general word representation. Without that distinction, a precise-looking figure can still support an imprecise conclusion.
A percentage needs a denominator
A share describes a numerator in relation to a defined total. The relevant denominator may be candidacies, elected outcomes, occupied seats or another clearly specified population. Those totals answer different questions. A women-candidate share should not be interpreted as an elected-seat share, even where both figures are drawn from the same election.
Unknown information also matters. If individual demographic classifications are not recorded, their absence should not be interpreted as evidence that the people concerned are men or belong to another category. A source-recorded sex field and a self-reported gender field should be labelled according to what the evidence actually establishes.
Keep periods and institutional levels distinct
An election result describes a particular contest or allocation process. A membership snapshot describes an institution at a reference date. Lawful changes after an election can mean that these records differ. A historical trend needs comparable observations rather than a mixture of dates and definitions presented as one continuous measure.
Aggregation requires particular discipline. National totals should not be counted again through regional components already included in them. Percentages should not be added across separate elections. Where compatible counts can be combined, the calculation and denominator need to be explained; otherwise, the source observations should remain separate.
Missing is not the same as zero
A documented zero is a substantive value. A missing value means that the relevant evidence is absent or unavailable within the selected scope. Treating both as zero can distort totals and comparisons, especially where reporting coverage differs between years, institutions or countries. An empty field cannot establish the absence of participation.
A complete assessment therefore looks at coverage as well as the figure itself. How many eligible records contribute to the measure? Which periods are represented? Does the source include every contest or only a defined collection? These questions help explain why a filtered result may change without implying that the underlying political reality changed in the same way.
Follow the evidence trail
A trustworthy record identifies its source, period, scope and calculation method. Readers should be able to distinguish original official information from subsequent analysis, and an independently verified figure from a value published with a stated limitation. Clear provenance supports scrutiny without implying that every record has the same level of completeness or precision.
GEWA's country information should be approached with these questions in mind. Read the context, examine source references and use filters to define the comparison. Evidence does not become more accurate merely because it is displayed in a chart; it becomes more understandable when the relationship between the number, its definition and its supporting record is visible.