Round Blog Numbers Without Hiding the Difference

Calculate a comparison from the raw values, then choose a rounded display that preserves what the reader needs to know. Keep both in your working record. Two rounded numbers can look equal even when their original counts differ; subtracting rounded displays can also exaggerate that difference.

The Office for National Statistics rounding guide, checked on 8 October 2026, describes the tradeoff between readability and retained detail. Its house advice is to choose rounding for the intended use and keep detail that matters. That does not give every blog number the same required precision.

The examples below use original invented catalogue counts, in files. They are supplied teaching inputs, not observed data or ONS examples. The question is exact: how many more files does catalogue B contain than catalogue A?

Keep each raw count beside its display

Our declared teaching rule rounds a positive count to the nearest multiple of 10 or 100. None of the supplied inputs is halfway between two multiples, so these examples do not need a tie-breaking rule.

Catalogue A

Exact count: 1244 files. Rounded to the nearest 10, the display is 1240 files. Rounded to the nearest 100, it is 1200 files.

For the exact-count question, keep 1244 files as the display used in the answer. The shorter number is a separate display choice, not a replacement for the original count.

Catalogue B

Exact count: 1246 files. Rounded to the nearest 10, the display is 1250 files. Rounded to the nearest 100, it is 1200 files.

Keep 1246 files for the exact comparison. Both A and B display as 1200 at the coarser level, but their raw values are different.

Catalogue C

Exact count: 4 files. Rounded to the nearest 10 or the nearest 100, the display is 0 files.

The actual supplied count remains 4 files. For a reader deciding whether this catalogue contains any files, an unexplained zero would give the wrong answer. Keep 4 files for that task rather than let a coarse display erase the positive count.

Subtract the originals

The raw comparison is 1246 − 1244 = 2 files. A supported teaching sentence is: “Catalogue B contains 2 more files than catalogue A.” This compares supplied counts; it reports no growth over time or reason for the difference.

Subtracting the nearest-10 displays gives 1250 − 1240 = 10 files. That is a difference between displays, not the original catalogue-count difference.

Subtracting the nearest-100 displays gives 1200 − 1200 = 0 files. The equality of those displays does not make the original counts equal.

For a coarse overview, a different sentence can be appropriate: “Catalogues A and B each contain approximately 1200 files.” It declares approximation and supplies no exact difference. Do not use that sentence's displays as the inputs to an exact-count calculation.

Choose the intended question before shortening the number. The blogger's research guide helps define that question; here, the precise comparison requires information that a coarse overview omits.

Reuse the filled rounding record

Download the filled rounding record CSV. It contains all three completed catalogue records, including every raw count, both display levels, the unit and the exact-task display.

The columns separate different jobs:

• catalogue identifies A, B or C.

• exact_count retains the original supplied integer.

• nearest_10_display and nearest_100_display contain our declared display calculations.

• unit names files.

• exact_task_display retains the count used for the exact question.

• teaching_marker identifies these rows as invented teaching data.

To reuse the record, replace a row with a real sourced value and retain its source and meaning in your own notes. Decide the reader's question and rounding rule before creating a display. Calculate any comparison from the original inputs, then check the sentence against that calculation. The example CSV is completed teaching material, not a measurement log for your site.

Keep unknowns and uncertainty intact

A missing value stays unknown; it is not a zero to round. If a source supplies an estimate or uncertainty, retain that meaning instead of presenting extra digits as exact knowledge. Halfway or negative inputs need an explicitly chosen rule before extending this narrow example. The file-count exercise establishes no instrument precision or universal significant-figure policy.

When you write the article, keep the detail needed for its actual decision visible in the body. Use the article brief to name that decision. Rounding should serve the answer you intend to give, while the raw record preserves what the display leaves out.