Significant Figures and Rounding in Unit Conversions

Introduction

You measure a piece of wood and find it is 7 inches long. You convert to centimeters using the exact factor: 7 × 2.54 = 17.78 cm. But should you report the answer as 17.78 cm? Or 17.8 cm? Or 18 cm? The question is not about the math — the math is exact. The question is about what the original measurement actually tells you, and how many digits you are justified in reporting. This is the domain of significant figures.

What Is Precision vs Accuracy?

Before getting to significant figures, it helps to separate two terms that are often confused:

Accuracy is how close a measurement is to the true value. A bathroom scale that consistently reads 1 kg too high is inaccurate but precise.

Precision is how reproducible or fine-grained a measurement is. A ruler marked in millimeters gives more precise readings than one marked only in centimeters, regardless of whether either is calibrated correctly.

Significant figures deal with precision. They express how many digits in a number carry meaningful information based on how the measurement was taken.

The Rules for Significant Figures

Which Digits Are Significant?

1. All non-zero digits are significant. In 4.73, all three digits are significant.
2. Zeros between non-zero digits are significant. In 4.073, all four digits are significant.
3. Leading zeros are not significant. In 0.0047, only the 4 and 7 are significant (2 sig figs).
4. Trailing zeros after a decimal point are significant. In 4.700, all four digits are significant — the trailing zeros indicate the measurement was made to the nearest thousandth.
5. Trailing zeros before a decimal point are ambiguous. In 4700, it is unclear whether 2, 3, or 4 digits are significant. Scientific notation resolves this: 4.7 × 10³ (2 sig figs), 4.70 × 10³ (3 sig figs), 4.700 × 10³ (4 sig figs).

Counting Significant Figures: Examples

Sig Figs in Unit Conversions

Multiplication: Match the Least Precise Operand

When you multiply or divide (which is what most unit conversions require), the result should have no more significant figures than the least precise number in the calculation.

Example: Convert 7 inches to centimeters.

7 in × 2.54 cm/in = 17.78 cm

The number 7 has one significant figure. The conversion factor 2.54 is exact (defined by international agreement, not measured), so it does not limit sig figs. Therefore the result should be reported to one significant figure: 18 cm.

Example: Convert 7.00 inches to centimeters.

7.00 in × 2.54 cm/in = 17.778 cm → 17.8 cm

Now 7.00 has three significant figures, so the result is reported to three sig figs: 17.8 cm.

The same distance, measured to different precisions, justifies different numbers of decimal places in the converted result. This is not a rounding convention — it is a statement about what the original measurement actually tells you.

Addition and Subtraction: Match the Least Precise Decimal Place

For addition and subtraction, the rule is about decimal places, not sig figs:

1.234 m + 0.56 m = 1.794 m → 1.79 m

The second number is only known to the hundredths place, so the result rounds to the hundredths place. This comes up in temperature conversions:

°C = (98.6°F − 32) × 5/9
°C = 66.6 × 5/9
°C = 37.0°C

Here 98.6°F has one decimal place; 32 is exact; the subtraction result is 66.6 (one decimal place); the multiplication by 5/9 (exact ratio) gives 37.0°C. Reporting this as 37.000°C would misrepresent the precision of the original measurement.

Rounding Methods

Once you know how many figures to keep, you need to decide how to round. There are three common methods:

Half-Up (Standard Rounding)

The most familiar method: if the digit being dropped is less than 5, round down; if 5 or greater, round up.

• 17.45 → 17.5 (rounds up because the dropped digit is 5)
• 17.44 → 17.4 (rounds down)
• 17.35 → 17.4 (rounds up)

This is the rounding most people learn in school, and it introduces a small systematic upward bias because every "5" case rounds up.

Banker's Rounding (Round Half to Even)

Also called "round half to even" or "statistical rounding." When the dropped digit is exactly 5 (with no further digits), round to the nearest even number.

• 17.45 → 17.4 (4 is even, so round toward 4)
• 17.55 → 17.6 (6 is even, so round toward 6)
• 17.25 → 17.2 (2 is even)
• 17.35 → 17.4 (4 is even)

Over many calculations, half the "5" cases round up and half round down, eliminating the systematic bias. This method is used in IEEE 754 floating-point arithmetic, financial calculations, and statistical analysis. Python's built-in round() function uses banker's rounding.

Truncation

Simply drop the unwanted digits without adjusting. 17.78 truncated to one decimal is 17.7. This is rarely appropriate in unit conversions because it always introduces a downward bias, but it appears in some computing contexts.

Which Method Should You Use?

For everyday unit conversions: standard half-up rounding is fine. The systematic bias is negligible for one-off calculations.

For repeated calculations or financial contexts: banker's rounding reduces accumulated error.

For engineering calculations that feed into further calculations: carry extra digits throughout and round only at the final step.

How Many Decimal Places to Report?

A practical guide by context:

Scientific and Technical Work

Match the significant figures of your least precise measurement. If you measured 12.3 cm (3 sig figs), your converted result should have 3 sig figs: 12.3 cm = 4.84 inches (not 4.8425197 inches).

Everyday Conversions

Report enough decimal places to be useful, but not so many as to imply false precision: - Body weight: nearest 0.1 kg or 0.5 lb is sufficient - Cooking: nearest gram or nearest teaspoon - Room dimensions: nearest centimeter or nearest quarter-inch - Running distances: 5.0 km is appropriately precise; 4.97589 km is not

Currency Conversions

Always report to exactly two decimal places (cents), regardless of the input's precision. This is a convention, not a sig-fig rule.

GPS Coordinates

Latitude and longitude to six decimal places (0.000001°) resolves to about 11 cm — more than sufficient for navigation. Many apps report more, but beyond six, the digits are meaningless noise.

Common Mistakes

Reporting All Calculator Digits

A calculator returns 17.78 cm from 7 × 2.54. Reporting "17.78 cm" when the original measurement was "7 inches" implies you measured to the nearest hundredth of a centimeter, which is not what a basic ruler measurement supports. The number of digits in the result should match the information content of the input.

Treating Exact Conversion Factors as Limiting

The conversion factor 1 inch = 2.54 cm is exact — it is not a measured quantity. It does not limit significant figures. Only measured values with finite precision limit sig figs.

Similarly, the constant π = 3.14159... does not limit significant figures. If you use π in a calculation, carry enough decimal places of π to match the precision of your measurement.

Rounding Intermediate Results

In a multi-step calculation, rounding at each intermediate step accumulates error. The correct approach: carry at least one or two extra significant figures through intermediate steps, and round only the final answer.

Example: Convert 1 mile to meters to centimeters.

Correct: 1 mile × 1609.344 m/mile × 100 cm/m = 160,934.4 cm

Wrong: 1 mile → 1609 m (rounded) → 160,900 cm (14 cm of accumulated error)

Precision in Context: A Practical Perspective

The rules for significant figures can feel pedantic in daily life. No one needs to know whether a 7-inch measurement converts to "18 cm" or "17.78 cm" when deciding whether a shelf will fit. In that context, extra digits are harmless.

Sig figs matter most when: - Measurements will be used in further calculations - Results will be used to make decisions with real consequences (medical dosing, structural engineering, chemical synthesis) - You need to communicate the reliability of your measurement to someone else

The unit converter at UnitFYI returns results to high precision by default, which is appropriate for a reference tool. When you use those results in practice, apply the sig-fig rules to determine what precision is actually justified by your original measurement.

Summary

• Significant figures express the precision of a measurement, not just its magnitude
• In multiplication/division conversions, the result has as many sig figs as the least precise input
• Exact conversion factors (like 1 in = 2.54 cm) do not limit significant figures
• Standard rounding (half-up) is appropriate for everyday use; banker's rounding reduces bias in repeated calculations
• Round only at the final step in multi-step conversions
• Report as many digits as are justified — no more, no fewer