Blog
Rounding and Significant Figures in SPM Mathematics
Rounding at the wrong moment quietly changes your final answer, and ignoring the question's instruction throws away an easy mark. Here are the rules that keep both safe.
Round once, at the end
The most common precision error is rounding a middle value and then carrying that rounded number forward, which drifts the final answer off. Keep every intermediate value at full precision, on the calculator, not on paper, and round only the final answer.
If you must write a middle value down, write more digits than you think you need.
Give exactly what the question asks
"Correct to two decimal places" and "correct to three significant figures" are different instructions, and answering in the wrong form can cost the accuracy mark. Read the required form, do the rounding at the end, and state the answer in that form.
When no instruction is given, three significant figures is a safe, conventional choice.
Know the difference between the two
Decimal places count digits after the point; significant figures count from the first non-zero digit. So 0.04052 is 0.04 to two decimal places but 0.0405 to three significant figures.
Mixing them up is a careless error that is entirely avoidable once you can see the distinction.
Zeros that count, and zeros that don't
Significant figures trip students up mostly because of zeros, since a zero sometimes counts and sometimes doesn't. Any zero between two other digits always counts, so 205 has three significant figures.
A zero at the very start of a decimal, before the first non-zero digit, never counts, so 0.0034 has only two significant figures, the 3 and the 4. Trailing zeros after a decimal point do count, so 3.40 has three significant figures, not two, because that final zero was written on purpose to show precision.
The tricky case is a whole number like 4500, without more context, it is genuinely ambiguous whether it has two, three or four significant figures, so SPM questions asking you to round to a given number of significant figures usually give you a decimal already, or expect scientific notation such as 4.5 × 10³ to remove the doubt. When in doubt, count from the first non-zero digit and stop counting once you reach the required number of digits, replacing anything after with a placeholder zero only if it is a whole number.
Rounding a number that is already in standard form
Standard form (also called scientific notation) writes a number as a × 10ⁿ, where a is between 1 and 10, and rounding one of these numbers means rounding only the coefficient a while leaving the power of ten alone, as long as a stays within that range. For example, 3.456 × 10⁴ rounded to 2 significant figures is 3.5 × 10⁴, not 35000 written out and then rounded separately, since converting back and forth needlessly risks losing track of the power.
The case that catches students out is when rounding the coefficient pushes it to 10 or beyond, because standard form no longer allows that. Take 9.96 × 10²: rounded to 2 significant figures, the coefficient rounds up to 10, which cannot stand as the coefficient, so the whole number has to be rewritten as 1.0 × 10³ to keep exactly one non-zero digit before the decimal point.
The power of ten increases by one purely to keep the form valid, even though the question only asked you to round, not to re-express the number in a different way.
The halfway rule: what happens to a trailing 5
One rounding case worth knowing cold is what happens when the digit immediately after the one you are keeping is exactly 5, since this is the situation that actually causes disagreement between students, calculators and even different pieces of software. The convention taught throughout Malaysian secondary school, and the one to default to in SPM Maths unless a question states otherwise, is round half up: if the first digit being cut off is 5 or more, the last digit you keep increases by one, and if it is 4 or less, it stays as it is.
So 2.35 rounded to 1 decimal place becomes 2.4, and 0.125 rounded to 2 decimal places becomes 0.13. In practice this exact-halfway case is rarer than it looks, because most values produced by a real calculation carry extra digits beyond that 5 which settle the rounding one way or the other before you even reach the decision.
The case worth watching for is a number that was already rounded once earlier in a multi-part question and now lands exactly on a 5 at the digit you need to cut, treating that as round half up is the safe default.
Frequently asked questions
What if the question doesn't say how to round the answer?
If no instruction is given, SPM Maths generally expects three significant figures unless the answer is already an exact value, such as a whole number from counting or a fraction that doesn't need rounding at all. This is a common convention rather than a rule printed on every paper, so if a specific instruction is given in the question, always follow that instead, since it overrides the general convention.
Does rounding differently at each step really add up to a noticeable error?
Yes, and the effect is often bigger than it looks, because each rounded value is reused in a formula that can multiply or raise the small error further. Two students starting from the same given values can land on final answers that differ enough to fall outside the accepted range purely from where they chose to round during the working, not from any actual mistake in method.
How many significant figures should I use if the question is silent about it throughout?
Carry your working to at least four significant figures on paper or keep it unrounded in your calculator, then round only your final answer to three significant figures as the safe default. This protects your working from small compounding errors while still giving a final answer in the form markers most commonly expect.