How to Round Numbers

A student rounds 4.837 to 4.83 instead of 4.84. A parent checks homework, sees the mistake, but cannot recall the rule. The three-step rounding method taught in grades 3 through 5 under Common Core 3.NBT.A.1, 4.NBT.A.3 and 5.NBT.A.4 walks through worked examples, explains why 5 rounds up (and when it does not), and points out common errors.

How to round numbers manually uses the standard round-half-up method that U.S. schools teach. Decimal-place rounding for a chemistry lab or significant-figure rules for a physics calculation are handled in separate resources. Here the focus is the basic rounding rule explained clearly with examples.

Rounding is the deliberate replacement of a number with a nearby, simpler number that has fewer digits. It makes numbers easier to work with, reduces the risk of errors when sharing numbers orally or in writing, and expresses measurements with the appropriate level of precision. For example, reporting $1,234.56789 as $1,235 avoids unnecessary precision. Rounding does not make numbers more accurate. It makes them more practical.

Rounding simplifies calculations, clarifies data presentation, and acknowledges the limits of measuring instruments. A ruler that measures only to tenths of a centimetre cannot justify a result of 12.345 cm. Rounding to 12.3 cm is honest. In finance, prices are rounded to the cent. In everyday life, time, distances and quantities are rounded when estimating. The process is fundamental to the Common Core standards and appears in every math curriculum from grade 3 onwards.

The standard rounding rule taught in schools is round half up. It follows three steps:

1. Identify the rounding digit. Find the digit in the place value you are rounding to. For example, if rounding to the tenth, the rounding digit is the digit in the tenths place.
2. Look at the digit to the right (the decision digit). This digit determines whether the rounding digit stays the same or goes up by one.
3. Apply the rule. If the decision digit is 5 or greater, increase the rounding digit by one. If the decision digit is 4 or less, leave the rounding digit unchanged. Then drop all digits to the right of the rounding digit.

This method is the default in most primary-school curricula and in many calculators. It is not the only rounding method. IEEE 754-2019, the standard for floating-point arithmetic, defines five rounding-direction attributes, and some software uses round half to even (bankers' rounding) to avoid cumulative bias. For homework and everyday manual rounding, round half up is the rule.

A place-value chart helps identify the rounding digit. Consider the number 5,827.346:

Place	Value	Digit
Thousands	1,000	5
Hundreds	100	8
Tens	10	2
Ones	1	7
Tenths	0.1	3
Hundredths	0.01	4
Thousandths	0.001	6

If rounding to the hundred, the rounding digit is 8 (hundreds place). The decision digit is the digit to its right, which is 2 (tens place). Since 2 is less than 5, the rounding digit stays 8 and all digits to the right become zero, giving 5,800.

Rounding a Decimal to Tenths

Round 3.74 to the tenth.

• Rounding digit: 7 (tenths place).
• Decision digit: 4 (hundredths place).
• 4 is less than 5, so the rounding digit stays 7.
• Drop all digits after tenths: 3.7.

Rounding a Whole Number to Hundreds

Round 2,863 to the hundred.

• Rounding digit: 8 (hundreds place).
• Decision digit: 6 (tens place).
• 6 is 5 or greater, so the rounding digit increases to 9.
• Drop all digits after hundreds: 2,900.

Rounding a Decimal to Hundredths

Round 9.2417 to the hundredth.

• Rounding digit: 4 (hundredths place).
• Decision digit: 1 (thousandths place).
• 1 is less than 5, so the rounding digit stays 4.
• Drop all digits after hundredths: 9.24.

When the rounding digit is 9 and the decision digit forces an increase, the 9 becomes 10 and the carry propagates left. Round 2.97 to one decimal place (tenths).

• Rounding digit: 9 (tenths place).
• Decision digit: 7 (hundredths place).
• 7 is 5 or greater, so the 9 becomes 10. The 10 sends a carry to the ones place: 2 becomes 3. The tenths digit becomes 0.
• Result: 3.0.

Another example: round 199.6 to the whole number.

• Rounding digit: 9 (ones place, the digit before the decimal).
• Decision digit: 6 (tenths place).
• 6 is 5 or greater, so the 9 becomes 10. The carry propagates: 199 becomes 200.
• Result: 200.

This carry is a common point of confusion, especially when the carry passes through multiple 9s. The result may have more digits than the original number (e.g., 9.999 rounded to two decimal places gives 10.00, which has four digits compared to the original's three).

The rule that 5 rounds up is a convention, not a mathematical truth. The number exactly halfway between two candidates (e.g., 4.5 is halfway between 4 and 5) has no unique integer. The round-half-up convention breaks the tie by always rounding the 5 upward. This avoids the confusion of leaving the tie unbroken, but it introduces a small upward bias in datasets that contain many ties.

In some contexts, scientific computing, statistics, some programming languages, round half to even (bankers' rounding) is used instead. Under that rule, 4.5 rounds to 4 (even), but 5.5 rounds to 6 (even). This reduces cumulative bias over many operations. IEEE 754-2019 defaults to roundTiesToEven for floating-point arithmetic, and Python's round() function uses it. For manual rounding in a grade-school or everyday context, the round-half-up rule is what is expected.

Double Rounding

Rounding a number more than once can introduce errors. For example, rounding 2.345 to two decimal places gives 2.35. If you then round that result to one decimal place, you might get 2.4, but the correct one-step rounding of 2.345 to one decimal place is 2.3 (because the decision digit is 4, which is less than 5). Always round only the final result.

Looking at the Wrong Digit

When rounding to tenths, the decision digit is the hundredths place, not the thousandths or any digit further right. Only the immediate digit to the right of the rounding digit matters. Looking at a digit further right can lead to an incorrect decision.

Confusing Decimal Places with Significant Figures

Rounding to two decimal places and rounding to two significant figures are different procedures. For example, 0.01499 rounded to one decimal place is 0.0, but rounded to two significant figures is 0.015. The first rule counts digits after the decimal point; the second counts digits from the first non-zero digit.

Incorrectly Handling Zeros

Zeros are placeholders. Rounding 0.00149 to two decimal places gives 0.00, which has zero significant figures. The zeros are not dropped; they indicate the precision of the rounded result.

Forgetting to Adjust Other Digits When Rounding Up

As shown in the carry example, rounding a 9 up can cause a cascade. Missing the carry leads to an incorrect result.

1. Round 4.837 to the hundredth. Answer: 4.84.
2. Round 2,863 to the hundred. Answer: 2,900.
3. Round 9.2417 to the tenth. Answer: 9.2.
4. Round 2.97 to one decimal place. Answer: 3.0.
5. Round 199.6 to the whole number. Answer: 200.
6. Round 0.01499 to two significant figures. Answer: 0.015.
7. Round 1.499 to one decimal place. Answer: 1.5.