Decimal places and the rounding digit
Digits to the right of the decimal point occupy tenths, hundredths, thousandths, and successively smaller places. In 12.3476, 3 is the tenths digit, 4 the hundredths digit, and 7 the thousandths digit. Rounding begins by identifying the requested place, not by counting digits from the end.
The digit immediately to the right is the deciding digit under the common round-to-nearest, ties-away-from-zero classroom rule described here. If it is 0 through 4, keep the rounding digit; if it is 5 through 9, increase the rounding digit by one. Remove later digits or replace integer-place digits with zeros as appropriate.
Place value extends left of the decimal as ones, tens, hundreds, and beyond. The same method rounds to whole numbers or powers of ten: identify the target place and inspect the next smaller place. Rounding 347 to the nearest ten gives 350 because the ones digit is 7.
Zeros between nonzero digits hold place value. In 5.0074, the tenths and hundredths digits are zero, the thousandths digit is 7, and the ten-thousandths digit is 4. Rounding to hundredths gives 5.01 because the 7 raises the hundredths zero; ignoring placeholder zeros would identify the wrong position.
A step-by-step rounding method
First locate and mark the requested place. Second inspect only the next digit. Third keep or increase the marked digit. Fourth remove digits to its right and retain trailing zeros when they communicate the requested decimal place. For example, 4.2 rounded to the nearest hundredth is written 4.20 when precision matters.
Rounding can carry across digits. Rounding 9.996 to the nearest hundredth inspects the thousandths digit 6, raises the hundredths 9, and carries to produce 10.00. Do not stop after writing an impossible digit such as 9.9(10).
When the deciding digit is below five, later digits do not change the decision: 2.349999 rounded to the nearest tenth is 2.3 because the hundredths digit is 4. When it is above five, later digits are likewise irrelevant. Only an exact tie may require a named tie-breaking rule.
Rounding to a decimal place differs from rounding to a number of significant figures. The value 0.004876 rounded to three decimal places is 0.005, while rounded to three significant figures it is 0.00488. Identify which instruction is requested before locating the rounding digit; this guide focuses on named decimal places.
Three complete rounding examples
Input: 18.746 rounded to the nearest tenth. Formula: inspect the hundredths digit. Calculation: the tenths digit is 7 and the next digit is 4, so keep 7. Result: 18.7. Interpretation: the original value lies closer to 18.7 than 18.8.
Input: 3.1458 rounded to the nearest hundredth. Formula: inspect the thousandths digit. Calculation: the hundredths digit is 4 and the next digit is 5, so the stated tie rule increases 4 to 5. Result: 3.15. Interpretation: the answer is reported to two decimal places.
Input: 0.9996 rounded to the nearest thousandth. Formula: inspect the ten-thousandths digit. Calculation: the thousandths digit is 9 and the next digit is 6; carrying produces 1.000. Result: 1.000. Interpretation: the trailing zeros show thousandth-place reporting.
Input: 27.995 rounded to the nearest hundredth. Formula: inspect the thousandths digit. Calculation: the hundredths digit is 9 and the next digit is 5; under ties away from zero, carrying moves through the tenths place. Result: 28.00. Interpretation: the zeros communicate a hundredth-place result after carrying.
Input: −6.384 rounded to the nearest tenth. Formula: inspect the hundredths digit. Calculation: the tenths digit is 3 and the next digit is 8, so the magnitude rounds upward to 6.4 and the negative sign remains. Result: −6.4. Interpretation: −6.4 is nearer to −6.384 than −6.3.
Negative numbers and tie rules
For negative numbers, apply the chosen rule to magnitude while preserving the sign. Under ties away from zero, −2.345 rounded to the nearest hundredth is −2.35. With a deciding digit below five, −2.344 becomes −2.34. Number-line thinking helps: choose the permitted value nearest the original.
Not every system uses the same rule for exact ties. Financial, statistical, programming, or regulatory contexts may require half-to-even, half-up, truncation, or another method. State the rule when ties matter and follow the governing software or policy rather than assuming one universal convention.
Half-to-even rounds an exact tie toward the candidate whose last retained digit is even. Under that rule, 2.345 to hundredths becomes 2.34, while 2.355 becomes 2.36. This can reduce aggregate directional bias, but it should not be substituted where a stated classroom, financial, or regulatory method requires another rule.
Rounding is not truncation
Truncation simply discards later digits. Rounding uses the next digit to choose a nearer permitted value. Truncating 6.789 to the nearest-hundredth format gives 6.78, while ordinary nearest rounding gives 6.79. The difference can matter when many results are totaled.
Do not round every intermediate step unless a procedure requires it. Early rounding can accumulate error or move a later result across a threshold. Keep sufficient internal precision, then round the final output. If a rule specifies staged rounding, document each stage.
Repeated rounding is unsafe. Rounding 2.449 to hundredths gives 2.45, and then rounding that display to tenths gives 2.5; rounding the original directly to tenths gives 2.4 because its hundredths digit is 4. Always round from the most precise available value to the final requested place.
Measurement, money, and reported precision
A converted or calculated number should not display more meaningful precision than its inputs support. Writing many decimal places can imply measurement certainty that does not exist. Trailing zeros may be useful when they deliberately show a requested place or measurement resolution.
Currency rules, tax calculations, invoices, laboratory reports, grades, and compliance thresholds can prescribe specialized rounding. This guide explains general decimal arithmetic and does not override those requirements. Use the authoritative rule for consequential decisions.
Threshold comparisons should normally use the unrounded value unless the governing rule says the rounded value controls. A measurement of 9.996 rounds to 10.00 at hundredths but remains below 10 before rounding. Confusing display rounding with the underlying comparison can change pass/fail or eligibility outcomes.
A stated tolerance and a rounded display are not automatically equivalent. Reporting 12.3 to the nearest tenth normally represents values near that display under the selected rule, but an engineering tolerance must be specified independently. Do not infer acceptance limits merely from the number of printed decimal places.
Using a calculator to verify rounding
A decimal calculator can verify arithmetic before rounding, while the rounding decision should be based on the unrounded result and specified place. Record both when reproducibility matters: for example, unrounded 12.3468; reported to hundredths, 12.35.
Check that the rounded value is one of the adjacent values at the requested increment and that it is no farther than half an increment from the original, subject to the tie rule. Retaining the original value prevents a rounded display from becoming false source precision.
Spreadsheet and programming functions may differ in tie behavior, binary floating-point representation, and treatment of negative values. A displayed decimal such as 2.675 may not be stored exactly in binary. For audited work, document the software, input representation, function, and required rule, and use decimal arithmetic where the governing process demands it.