Reading a Campaign's Own Evidence: How to Check the Numbers Candidates Cite
.jpg){kind=link}
Campaigns lean heavily on numbers. A candidate says crime is up, wages are flat, or a program saved a certain amount of money, and the figure lands with the weight of fact. Yet campaign materials are written to persuade, not to document, and the figures they cite are chosen for their rhetorical force. A civic reader does not need statistical training to test these claims. What helps is a consistent set of questions applied before a number shapes an opinion. This article sets out a working method for doing that, with attention to the gap between what a statistic measures and what a campaign says it proves.
Start With the Exact Wording of the Claim
Before evaluating a number, pin down what the claim says. Campaign speech compresses information. A figure like 'twice as many' or 'a record high' may refer to a specific measure, time span, or population that the listener never hears. Write the claim in full, including the time period, the geography, and the group it describes. If the speaker says unemployment fell, ask whether they mean the headline rate, the rate for a particular group, or the share of people working. Each of those can move in different directions over the same months.
This step often reveals that the claim is narrower than it sounds. A statistic about one city presented as a statewide trend, or a multi-year change presented as a recent one, is a common pattern. Rewriting the claim in plain terms makes the next steps easier, because you now know exactly what evidence would support it.
Identify the Original Source and Its Method
Next, find where the number came from. Strong campaigns cite a named source: a statistical agency, an academic study, an audit, or a published dataset. Weak claims cite 'experts,' 'studies show,' or an internal poll with no published methodology. When a source is named, go to it. Check whether the figure appears there, whether it is quoted accurately, and whether the surrounding text supports the use the campaign makes of it.
Methodology matters as much as the headline figure. Ask who collected the data, how, and from whom. A survey of likely voters from one party's list, a poll with a small sample, or a figure drawn from a convenience sample will not support the broad claims often built on them. Look for the sample size, the margin of error, and any notes on how the data were adjusted. If the methodology is hidden or unavailable, treat the number as unverified rather than false.
Check Whether the Comparison Is Fair
Many misleading statistics are accurate in themselves and misleading in their comparison. A campaign may compare its candidate's first year in office against an opponent's record in a different economic period, or set one program's cost against a total that includes unrelated spending. Ask what the number is being compared to, and whether the baseline is reasonable. Percent change can flatter a small starting figure, while absolute change can hide a large relative shift. Both framings are legitimate, but a campaign will usually choose whichever one helps it.
Also ask about causation. A claim that a policy caused an improvement needs more than a timeline showing the two happened together. Other factors may have changed at the same time, and a careful reader asks what else could explain the result. This is not a reason to dismiss every causal claim, only to recognize that the statistic alone rarely proves it.
A Practical Checklist for Campaign Numbers
Keep a short list you can run through quickly whenever a figure comes up. First, restate the claim precisely, including dates and population. Second, locate the named source and confirm the figure appears there. Third, read the methodology, noting sample size, collection method, and any caveats. Fourth, examine the comparison: is the baseline fair, and is the framing percent or absolute? Fifth, note whether the claim asserts cause, and what alternative explanations exist. Sixth, check for omitted context, such as a longer trend that tells a different story.
Some results will be clear: a number may be accurate, fairly framed, and well sourced, in which case it deserves weight. Others will be partly right, accurate in fact but stretched in meaning. A third group will fail the check entirely. Recording your conclusions in a notebook helps you notice patterns across a campaign, such as a candidate who repeatedly cites figures from a single partisan source. Over a season, that record tells you more about reliability than any single claim.