Why Significant Figures Are a Simplification of Real Uncertainty

Rule Statement

Significant figures (sig figs) are a conventional method for expressing the precision of a measured or calculated value. The rule is straightforward: the number of significant digits in a value reflects the resolution of the measurement, with the last significant digit carrying uncertainty. For example, a measurement of 12.3 cm implies a resolution of ±0.1 cm, whereas 12.30 cm implies ±0.01 cm. However, this is a simplification of real uncertainty. In practice, uncertainty is a continuous quantity that depends on the measurement process, calibration, environmental factors, and statistical analysis. Sig figs collapse this complexity into a single digit position, often losing information about the actual distribution of error.

The core rule for determining sig figs: all non-zero digits are significant; zeros between significant digits are significant; leading zeros are not; trailing zeros are significant only if a decimal point is present. But these rules are arbitrary and can misrepresent uncertainty when the measurement error is not uniform or when the value is used in further calculations.

Significant figures are a pedagogical and practical compromise—they are not a rigorous uncertainty propagation method.

For a more rigorous approach, metrologists use the Guide to the Expression of Uncertainty in Measurement (GUM), which recommends expressing uncertainty as a standard deviation or expanded uncertainty with a coverage factor. Sig figs are a quick estimate, but they should never replace a proper uncertainty budget in critical applications.

Worked Examples

Let’s examine how sig figs approximate uncertainty in practice.

Example 1: Multiplication

Suppose you measure a rectangle: length = 12.34 cm (4 sig figs) and width = 5.6 cm (2 sig figs). The area is 12.34 × 5.6 = 69.104 cm². According to the rule for multiplication, the result should have the same number of sig figs as the factor with the fewest, which is 2. So we round to 69 cm². But what does this imply? The actual uncertainty in the area, if we treat each measurement as having an uncertainty of ±0.005 cm and ±0.05 cm respectively, propagates to about ±0.6 cm² (using partial derivatives). The sig figs result of 69 cm² implies an uncertainty of ±1 cm², which is a reasonable approximation but not exact. The true uncertainty might be ±0.6 cm², which would round to 69.0 cm² if we had one more sig fig. Thus, sig figs often overestimate uncertainty.

Example 2: Addition

Add 12.1 + 3.24 = 15.34. The rule for addition is to round to the least precise decimal place. The first number has one decimal place, so the result should be 15.3. If we assume each number has an uncertainty of ±0.05 and ±0.005 respectively, the combined uncertainty is √(0.05² + 0.005²) ≈ 0.050. The result 15.3 has a resolution of 0.1, which is about twice the actual uncertainty. So sig figs are conservative here.

These examples show that sig figs give a ballpark uncertainty, not an exact one.

Counter-Examples

There are cases where sig figs fail dramatically.

Counter-Example 1: Ambiguous Trailing Zeros

The number 1500 could have 2, 3, or 4 sig figs depending on whether the zeros are significant. Without scientific notation, it’s ambiguous. This ambiguity can lead to miscommunication in engineering and science. For instance, if a specification says “1500 N” and the tolerance is ±5 N, the value has 4 sig figs, but a reader might assume only 2. The solution is to use scientific notation (1.500 × 10³) or explicit uncertainty.

Counter-Example 2: Exact Numbers

Exact numbers, like conversion factors (1 inch = 2.54 cm exactly) or counted objects (e.g., 3 apples), have infinite sig figs. Applying sig fig rules to them can incorrectly limit the precision of a calculation. For example, if you measure 2.54 cm and multiply by the exact factor 1, the result should retain the original sig figs, not be rounded to 1 sig fig.

Counter-Example 3: Logarithmic Functions

Sig fig rules for logarithms are often misapplied. The rule is that the number of decimal places in the result should equal the number of sig figs in the argument. But this fails when the argument is close to 1. For example, log(1.00 × 10³) = 3.000, but the integer part is the characteristic and the mantissa carries precision. The rule works, but many students mistakenly round to the same number of sig figs as the argument, leading to errors.

Convention Comparison Table

r>

Each convention changes the last retained digit, which can affect the implied uncertainty. For example, rounding 2.35 to 2 sig figs with half-up gives 2.4, while half-even gives 2.4 (since 4 is even? Actually 2.35 -> 2.4 because 3 is odd? Wait half-even: 2.35, the digit before 5 is 3, which is odd, so round up to 2.4. If it were 2.45, half-even would give 2.4 because 4 is even. So the difference is subtle but can accumulate.)

Standards Citation

Several standards address significant figures and rounding. The most relevant are:

• ASTM E29 – “Standard Practice for Using Significant Digits in Test Data to Determine Conformance with Specifications” (Section 6) defines how to round to a specified number of significant digits for conformance testing.
• ISO 80000-1 – “Quantities and units – Part 1: General” (Section 7.3) provides rules for rounding and significant figures, emphasizing that the number of significant digits should be consistent with the measurement uncertainty.
• NIST TN 1297 – “Guidelines for Evaluating and Expressing the Uncertainty of NIST Measurement Results” (Section 7) recommends reporting uncertainty with two significant digits and aligning the measured value accordingly.
• GUM (JCGM 100:2008) – “Evaluation of measurement data – Guide to the expression of uncertainty in measurement” (Clause 7.2.6) states that the numerical value of the expanded uncertainty should be given to at most two significant figures, and the measurement result should be rounded to match.

These standards highlight that sig figs are a tool for communication, but the underlying uncertainty should be reported explicitly when possible.

Common Mistakes

1. Treating exact numbers as having limited sig figs. Exact conversion factors and counted values have infinite sig figs.
2. Rounding intermediate results. Always carry extra digits during calculations and round only the final answer.
3. Misinterpreting trailing zeros. Without scientific notation, 100 is ambiguous. Use 1.00 × 10² to clarify.
4. Applying addition rules to multiplication. The rules differ: addition uses decimal places, multiplication uses sig figs.
5. Ignoring the uncertainty of the last digit. Sig figs imply a resolution, but the actual uncertainty may be larger or smaller.
6. Using sig figs for tolerance intervals. For engineering tolerances, always use explicit ± values, not sig figs.

Practice Problems

Test your understanding.

1. Round 0.004567 to 2 sig figs.
2. Compute 2.5 × 3.14159 and round to the correct number of sig figs.
3. Add 12.11 + 0.2 + 1.003 and round appropriately.
4. Express 1500 with 3 sig figs using scientific notation.

Solutions:

1. 0.0046 (leading zeros are not significant; the first significant digit is 4, so 4.6 × 10⁻³).
2. 7.8 (2 sig figs because 2.5 has 2).
3. 13.3 (least decimal place is 0.1 from 0.2, so round to one decimal).
4. 1.50 × 10³.

Software Behavior Note

Different software tools handle sig figs and rounding inconsistently. For example:

• Excel uses half-up rounding by default, but its display formatting can hide the actual stored precision.
• Python (using round()) implements banker’s rounding (half-even) for floats, which can surprise users.
• MATLAB uses half-up for round but has separate functions for rounding to significant digits.
• TI-84 calculators typically use half-up and may display fewer digits than stored.

When using any tool, always check the rounding mode and the number of displayed digits. For critical work, rely on explicit uncertainty propagation rather than sig figs.

Quick Reference Table

Remember that these rules are heuristics. For rigorous work, always refer to the GUM and NIST guidelines. Our significant figures calculator implements these rules accurately, but we encourage you to understand the underlying principles.