Hold the definition steady
Identify the office, chamber, electoral level and unit in each observation. Distinguish election outcomes from later membership snapshots. A change in the size or structure of an institution may affect a comparison even when every recorded value is accurate.
Check whether the source's categories or collection methods changed. A new data series may cover participants or contests absent from an earlier record.
Preserve gaps and reference dates
Missing years should not be filled with invented values or repeated latest figures. A gap is information about the evidence available. A documented zero, by contrast, is a value that belongs in the analysis.
Keep the election date, publication date and reference date distinct where they answer different questions. A document published this year may describe an earlier election or institutional composition.
Explain the comparison
Percentages should remain separate by period; adding them does not create a meaningful historical total. Where counts can legitimately be combined, identify compatible cohorts and avoid counting national and regional records twice.
A trend can show a change in the stated measure without explaining its cause. Connect the pattern with legal and institutional context, and distinguish an observed movement from an interpretation about why it occurred. This makes historical evidence more useful and less vulnerable to overstatement.