01Separate significant figures from decimal places

NIST's SI guide explains that significant digits are those needed to express the numerical value and notes the ambiguity of trailing zeroes in some integer forms. Scientific notation can make the intended precision clearer.NIST SP 811 — Expressing values of quantities (opens in a new tab)

Leading zeroes locate the decimal scale; they are not additional measured digits. A trailing zero after the decimal point can carry precision information. Count the meaningful digits rather than simply counting every character following the decimal point.

02Digits do not replace an uncertainty statement

A result written as 4.00 mg has a displayed last place of 0.01 mg. That does not prove its measurement uncertainty is exactly 0.01 mg, or that the true value lies within a particular interval. The uncertainty statement and method information are separate evidence.

For an illustrative comparison, two instruments might both display two decimal places while having different measurement limitations. The screen format alone cannot establish that their results have equivalent reliability. Read the reported uncertainty, validation or other relevant evidence rather than ranking by the number of digits.

03A unit conversion should preserve information

Suppose a source reports 0.0040 g. Expressing the same value as 4.0 mg preserves the two significant figures. Writing 4.000 mg adds digits; writing 4 mg may remove the indication that the second digit was significant. The unit change does not create a new measurement.

Exact prefix factors do not add measurement uncertainty of their own, but they do not reduce the uncertainty of the input either. A spreadsheet may display many digits after multiplication. Choose the reporting precision from the evidence and purpose rather than from the software's default format.

Keep a copy of the original value and unit if you standardise a table. That lets a reviewer see whether the conversion preserved the source's intended precision or whether rounding introduced an avoidable change.

04Specifications and results are different kinds of numbers

A limit is a decision criterion; a measured result is an observation produced by a procedure. Do not assume that equal numbers of decimal places make them directly comparable under every decision rule. Read how the laboratory applies the criterion and handles rounding.

An illustrative report might show a minimum criterion of 98.0% and a result of 98.04%. Preserve both as reported. Reducing the result to an integer before comparison can hide information, while appending zeroes to the criterion does not clarify the laboratory's actual rule.

If a value lies near a limit, ask for the decision basis rather than inventing one from the typography. The significant-figure question helps you notice a problem; it does not independently decide whether the material meets the requirement.

05Use an auditable reporting convention

• ✓Retain source digits when transcribing results.
• ✓Keep uncertainty and precision as separate fields.
• ✓Use scientific notation to clarify ambiguous trailing zeroes where appropriate.
• ✓Preserve information through exact prefix conversions.
• ✓Apply a documented rounding convention when producing a final summary.

For calculations, retain enough intermediate precision to avoid avoidable rounding error, then report the final result appropriately. A clear convention is more useful than adding decimals to make a specification appear technical.