Adding and Subtracting in Scientific Notation: Precision Rules and Rounding

Rule Statement

When adding or subtracting numbers expressed in scientific notation, the fundamental rule is that the result must be rounded to the least precise decimal place among the operands. This is a direct consequence of the error propagation principle for sums and differences: the absolute uncertainty of the result is the sum of the absolute uncertainties of the operands, and the limiting factor is the operand with the fewest decimal places (i.e., the largest absolute uncertainty).

To apply the rule correctly, follow these steps:

1. Convert all numbers to the same exponent (typically the largest exponent) by adjusting the mantissa (coefficient) accordingly.
2. Align the decimal points by writing each number with the same power of ten.
3. Perform the addition or subtraction on the mantissas.
4. Identify the operand that has the fewest decimal places in its mantissa after alignment. This determines the decimal place to which the result must be rounded.
5. Round the final mantissa to that decimal place, then rewrite in scientific notation (if necessary) while preserving the number of significant figures.

Key principle: In addition and subtraction, it is the position of the last significant digit (the decimal place) that matters, not the total number of significant figures. This differs from multiplication and division, where the number of significant figures is the limiting factor.

For example, consider 1.23 × 104 and 4.5 × 103. After aligning to the exponent 4, the second number becomes 0.45 × 104. The first mantissa has two decimal places (1.23), the second has two decimal places (0.45) – so the result should be rounded to two decimal places in the mantissa, i.e., to the nearest 0.01 × 10⁴ = 100. The sum is 1.68 × 10⁴.

This rule is universally applied in metrology, chemistry, physics, and engineering. It is codified in standards such as ASTM E29, ISO 80000-1, and the GUM (JCGM 100:2008).

Worked Examples

Example 1: Same exponent

Add 3.456 × 105 and 2.71 × 105.

1. Exponents are already equal (5).
2. Align decimal points: 3.456 and 2.71.
3. Add mantissas: 3.456 + 2.71 = 6.166.
4. Determine least precise decimal place: The first mantissa has 3 decimal places, the second has 2 decimal places. So the result must be rounded to 2 decimal places.
5. Round 6.166 to 6.17 (since the third decimal is 6, round up).
6. Final answer: 6.17 × 105.

Example 2: Different exponents

Subtract 8.90 × 103 from 1.234 × 104.

1. Convert to the larger exponent (4): 1.234 × 104 remains; 8.90 × 103 becomes 0.890 × 104.
2. Align: 1.234 and 0.890.
3. Subtract: 1.234 - 0.890 = 0.344.
4. Determine least precise decimal place: The first mantissa has 3 decimal places, the second has 3 decimal places (0.890). So the result is rounded to 3 decimal places.
5. 0.344 already has 3 decimal places, so no further rounding.
6. Final answer: 0.344 × 104 = 3.44 × 103 (or 3.44 × 103).

Example 3: Rounding changes exponent

Add 9.99 × 102 and 1.5 × 101.

1. Convert to exponent 2: 1.5 × 101 = 0.15 × 102.
2. Add: 9.99 + 0.15 = 10.14.
3. Least precise decimal place: first has 2 decimal places, second has 2 decimal places (0.15). So round to 2 decimal places: 10.14.
4. But 10.14 has an exponent 2, so we write as 1.014 × 103? Actually 10.14 × 10² = 1.014 × 10³. However, rounding to 2 decimal places in the mantissa of the original exponent (2) gives 10.14, which is 1.014 × 10³ – but the mantissa now has 3 decimal places. The rule is to round to the decimal place of the least precise operand, not to a fixed number of decimal places after rewriting. Since the least precise operand (0.15) has two decimal places, the result should be rounded to the same absolute precision: 0.01 × 10² = 1. So 10.14 rounds to 10.1 (one decimal place) because the third decimal is 4, so we round down. Actually 10.14 rounded to 1 decimal place is 10.1. Then 10.1 × 10² = 1.01 × 10³. So the final answer is 1.01 × 103. This is a common subtlety: the decimal place to which we round is determined before any normalization.

In practice, it is safest to perform the operation with the original exponent, round to the correct decimal place, and then convert to proper scientific notation.

Counter-Examples

The following examples illustrate common errors:

Error 1: Rounding to the fewest significant figures instead of the fewest decimal places.

Add 1.2 × 103 and 3.45 × 102. Incorrect: 1.2 has 2 sig figs, 3.45 has 3, so round to 2 sig figs. That would give 1.5 × 10³ (since 1.2+0.345=1.545, round to 2 sig figs = 1.5). But the correct rule: align to exponent 3: 1.2 and 0.345. The first has 1 decimal place, the second has 3 decimal places. So round to 1 decimal place: 1.2+0.345=1.545 → round to 1 decimal = 1.5. Actually both give 1.5 here, but consider 1.2 × 103 and 3.45 × 102 = 1.2 and 0.345 → sum 1.545 → round to 1 decimal = 1.5, same. But try 1.23 × 103 and 4.5 × 102 = 1.23 and 0.45 → sum 1.68 → round to 2 decimals (since 1.23 has 2) = 1.68, but if you round to 2 sig figs you get 1.7. So the correct answer is 1.68 × 10³, not 1.7 × 10³.

Error 2: Forgetting to align exponents before comparing decimal places.

Adding 2.5 × 104 and 3.2 × 102 – if you simply look at decimal places in the original mantissas (2.5 has 1, 3.2 has 1), you might think the result should have 1 decimal place. But after aligning to 10⁴, the second becomes 0.032 × 10⁴, which has 3 decimal places. The first has 1 decimal place, so the result should be rounded to 1 decimal place in the mantissa of 10⁴, i.e., to the nearest 0.1 × 10⁴ = 1000. The sum is 2.532 × 10⁴ → round to 2.5 × 10⁴. If you incorrectly used 1 decimal place from the original, you’d get 2.5 × 10⁴ as well, but consider a case where the second has more decimal places after alignment: 2.5 × 104 and 3.21 × 102 → aligned: 2.5 and 0.0321 → sum 2.5321 → round to 1 decimal = 2.5. That’s fine, but if the first had 2 decimal places and the second after alignment had 3, you’d round to 2. So the key is to align first.

Error 3: Rounding intermediate results.

When adding multiple numbers, do not round at each step. Only round the final result. For example, add 1.234 × 102, 5.6 × 101, and 7.89 × 100. Align to 10²: 1.234, 0.56, 0.0789 → sum = 1.8729. The least precise is 5.6 × 10¹ which becomes 0.56 (2 decimal places) – actually 0.56 has 2 decimal places, 1.234 has 3, 0.0789 has 4. So round to 2 decimal places: 1.87 × 10². If you rounded each to 2 decimal places first, you’d get 1.23 + 0.56 + 0.08 = 1.87, which is same here, but in other cases it may differ.

Common Mistakes

• Using significant figure count instead of decimal place. This is the most frequent error. Remember: for addition/subtraction, the absolute precision (decimal place) is the criterion.
• Forgetting to convert all numbers to the same exponent. Without alignment, you cannot compare decimal places correctly.
• Rounding too early. Always carry extra digits through intermediate steps and round only the final result.
• Misinterpreting trailing zeros. In scientific notation, trailing zeros in the mantissa are significant. For example, 1.50 × 103 has three significant figures and two decimal places.
• Not applying the rule when subtracting nearly equal numbers. When subtracting two close numbers, the result may have far fewer significant figures than the operands. The decimal-place rule automatically handles this, but be aware that the absolute uncertainty remains the same.

Quick Reference Table

Related Rules

Understanding addition and subtraction in scientific notation is part of the broader topic of significant figures. For multiplication and division, the rule is different: the result must have the same number of significant figures as the operand with the fewest significant figures. For mixed operations, apply the appropriate rule at each step, but avoid rounding until the final result. Also related is the concept of error propagation (GUM), which provides a more rigorous approach using uncertainties. For a deeper dive, see our articles on Multiplication and Division in Scientific Notation and Understanding Measurement Uncertainty.

Standards Citation

The rounding rules for addition and subtraction are specified in several international standards:

• ASTM E29-13 – Standard Practice for Using Significant Digits in Test Data to Determine Conformance with Specifications. Section 6.2 covers rounding of test results for addition and subtraction, stating that the result shall be rounded to the least precise decimal place.
• ISO 80000-1:2009 – Quantities and units – Part 1: General. Clause 7.3.4 discusses rounding of numerical values, emphasizing that for sums and differences, the number of decimal places is the determining factor.
• JCGM 100:2008 (GUM) – Evaluation of measurement data – Guide to the expression of uncertainty in measurement. Section 7.2.6 provides guidance on rounding of results, recommending that the numerical value of the expanded uncertainty be rounded to two significant figures and that the measurement result be rounded to match the uncertainty. This implies the decimal-place rule for addition/subtraction when combining uncertainties.
• NIST Technical Note 1297 – Guidelines for Evaluating and Expressing the Uncertainty of NIST Measurement Results. Appendix A discusses significant figures and rounding.

These standards are essential for laboratories, quality control, and any field where measurement results are reported with a specified precision.

Software Behavior Note

Different software tools handle rounding of scientific notation differently. For example:

• Python (using the decimal module) allows explicit control of rounding and precision. The default float type uses binary floating-point, which can introduce small errors; always use Decimal for critical calculations.
• Excel and Google Sheets typically round to the displayed number of decimal places, but the underlying value may retain more digits. When using ROUND functions, specify the desired decimal places explicitly.
• MATLAB uses double precision and displays results with a default format. Use format long or vpa for higher precision. The round function rounds to the nearest integer unless a number of digits is specified.
• TI-84 calculators have a Sci mode that displays results in scientific notation, but rounding is applied to the displayed mantissa based on the calculator’s internal precision (typically 14 digits). Always verify the least precise decimal place manually.

When using any software, it is best practice to perform the calculation with full precision and then apply the rounding rule manually or via a dedicated significant figures calculator, like the one on this site, to ensure compliance with standards.

Practice Problems

Test your understanding with the following problems. Answers are provided at the end.

1. Add 4.56 × 103 and 1.2 × 102.
2. Subtract 2.30 × 10-4 from 1.5 × 10-3.
3. Add 7.89 × 106, 1.23 × 105, and 4.5 × 104.
4. Subtract 9.99 × 102 from 1.00 × 103.

Answers: 1) 4.68 × 10³ (align to 10³: 4.56 + 0.12 = 4.68, round to 2 decimals) 2) 1.27 × 10^-3 (align to 10^-3: 1.5 – 0.23 = 1.27, round to 2 decimals) 3) 8.04 × 10⁶ (align: 7.89 + 0.123 + 0.045 = 8.058, round to 2 decimals) 4) 1.0 × 10¹? Actually 1.00×10³ – 0.999×10³ = 0.001×10³ = 1×10⁰. But rounding: both have 2 decimal places? 1.00 has 2, 9.99 has 2, so result should have 2 decimal places: 0.001 has 3 decimal places, but the least precise is 2, so round to 2 decimal places: 0.00 × 10³ = 0? That seems odd. Actually careful: 1.00×10³ and 9.99×10² = 0.999×10³. Subtract: 1.000 – 0.999 = 0.001. The least precise decimal place is 2 (from 1.00 and 0.999? 0.999 has 3 decimal places, 1.00 has 2, so round to 2 decimal places: 0.001 rounds to 0.00? That gives 0.00×10³ = 0. But that’s not correct because the absolute uncertainty is ±0.01×10³ = ±10, so the result could be 0 ± 10. So the answer is 0×10³ or 0 with an uncertainty of 10. In practice, we might write 0.0 × 10³? Actually 0.00 × 10³ = 0. So answer is 0. But that’s a special case. We’ll keep the answer as 0.0 × 10³ or simply 0. For simplicity, we’ll say 0.0 × 10³.