I often find myself pausing when a politician tosses a statistic into an interview or speech. Numbers give arguments weight — but they can also be misplaced, outdated, or taken out of context. Over the years I’ve developed a quick, practical routine for verifying quoted figures using only public sources, no freedom of information (FOI) request required. Here’s the five-source method I use when I want to check a claim fast and reliably.
My guiding principle: follow the chain of evidence
When someone quotes a number, my first thought is: where did that number come from? A trustworthy claim should point to an original or near-original source. Working backwards from the politician’s quote, I try to trace the statistic to a primary dataset, official report, or peer-reviewed study. If the trail ends at a press release or a partisan briefing, I treat the figure with caution.
The five public sources I always check
These are the five sources I use in this order. Together they cover most claims you’ll hear about employment, public spending, health, education, crime, migration and more.
1) Start with the official statistical office
In the UK, that’s the Office for National Statistics (ONS). In many countries there’s a national statistics institute (e.g., INSEE in France, BLS in the US). These portals are the most reliable place to find raw figures, time series, and metadata. I search the site for the indicator named by the politician, or use Google with site:ons.gov.uk plus keywords.
Practical search tips:
Red flags: a statistic that can’t be found on the national statistics portal is often a sign it was derived from a non-standard calculation or cherry-picked subset.
2) Check parliamentary records and official statements for context
Parliamentary transcripts (Hansard in the UK), ministerial statements, and official briefing notes are essential for context. Politicians sometimes compress a complex dataset into a neat soundbite — and the nuance gets lost.
What I look for:
Practical search tip: use site:parliament.uk plus a few words from the quote; Hansard is fully searchable and often indexed by Google.
3) Cross-check with international organisations or independent national agencies
If the claim involves comparisons (e.g., “we spend less on education than most EU countries”) or uses rates rather than absolute counts, international datasets are invaluable. The OECD, Eurostat, World Bank and UN databases standardize definitions across countries and can expose apples-to-oranges comparisons.
Why this matters: politicians sometimes compare figures that aren’t comparable — different age ranges, different accounting rules, or different years. International databases typically document those differences clearly.
Practical search examples:
4) Look for peer-reviewed studies or working papers that replicate or critique the number
Sometimes the statistic is new or derived from a novel analysis — in which case peer-reviewed papers, working papers (NBER, IZA), or preprints (SSRN, arXiv) can explain the methodology. I search Google Scholar for the statistic and the keywords used by the politician.
What I want to know from academic sources:
If the number depends on modelling choices, note that to readers: model-dependent figures are less definitive than direct counts.
5) Use reputable fact-checkers and archival news databases for secondary confirmation
Trusted fact-checking organisations (Full Fact in the UK, FactCheck.org, PolitiFact) often have already tested popular political claims. Major news organizations keep searchable archives and often publish explainers with sourced figures.
How I use these sources:
Putting the five-source method into practice (a quick walkthrough)
Imagine a minister says: “Immigration has driven a 15% increase in housing demand over the last five years.” Here’s how I would verify it.
After those steps I can say whether the 15% is verifiable, miscontextualised, or unsupported.
Practical tips and red flags to watch for
Here are practical heuristics I use every day:
How I report what I find
When I publish a verification, I include links to the primary datasets, the parliamentary transcript, and any papers I relied on. I spell out any assumptions I needed to make to reproduce the number (for example, excluding temporary residents). Transparency about method builds trust — and it helps readers and editors follow the same trail.
Verifying statistics doesn’t require specialised tools — just a methodical approach and a few reliable public sources. Stick to the five-source routine, document your steps, and you'll be able to tell readers not just whether a number is true, but how confident you are in that judgment and why.
