The Thing Nobody Gets Right About Rounding
Most students memorize the rule — the answer has as many significant figures as the least precise measurement — and then immediately mess it up by rounding too early or miscounting zeros. I watch this happen constantly, even from people who should know better. Here is what actually matters. When you multiply or divide, the result carries the same number of significant figures as the factor with the fewest of them. That is the entire rule. Everything else is just making sure you count correctly. Take 3.2 multiplied by 4.56. 3.2 has two sig figs. 4.56 has three. Your answer gets two. 3.2 times 4.56 is 14.592, which rounds to 15. Done. No drama.
Now the things that trip people up. Trailing zeros after a decimal point count as significant. So 2.500 has four sig figs, not two. Leading zeros never count — 0.0042 is two sig figs. Exact numbers like conversion factors or counted quantities have infinite sig figs and do not limit your answer at all. I ran into this with a student last year who was doing a density calculation. Mass was 12.50 grams and volume was 4.2 milliliters. She wrote the answer as 2.976 g/mL and lost points every time. The volume has two sig figs. The mass has four. The answer needed two. She kept writing four digits because she thought rounding to 3.0 would look wrong, like it had lost information. It hadn't. I told her to just accept that 3.0 g/mL is the correct answer and move on. She stopped second-guessing it. Another issue is carrying extra digits through intermediate steps. If you are doing a multi-step problem — say, converting units and then multiplying — keep at least one extra digit in your calculator until the final step. Round only at the end. Rounding after every operation introduces cumulative error that can change your final sig fig count, especially in longer calculations.
There is also a subtle edge case that comes up when you deal with numbers like 100 or 500. Without a decimal point, those trailing zeros are ambiguous. Are they significant or not? In practice, scientists write 1.00 × 10² when they mean three sig figs, or they put a bar over the significant zero. If your professor gives you 100 m and does not clarify, treat it as one sig fig unless the context makes it clear otherwise. That has caused more grading disputes than anything else I have seen. One counter-intuitive thing: sometimes the result of a division needs fewer sig figs than either input appears to warrant. Consider dividing 0.050 by 2.00. 0.050 has two sig figs. 2.00 has three. The answer is 0.025, which has two sig figs. The leading zeros in 0.025 are not significant. Students often look at 0.025 and think it has three sig figs because there are three non-zero digits visible. It does not. The leading zeros are placeholders, not measurements. For practice, work through problems in this order: identify each number's sig figs first, then compute, then round the final answer. Do not skip the identification step. I see too many people just multiply and hope for the best. That is how you end up with answers like 14.592 when the correct answer is 15.
Here are a few examples to work through: 6.22 × 3.0 = 19 (three sig figs times two sig figs gives two sig figs) 84.0 ÷ 2.00 = 42.0 (three sig figs divided by three sig figs gives three sig figs)
0.0034 × 5600 = 19 (two sig figs times two sig figs gives two sig figs — note that 5600 without a decimal is treated as two sig figs here) The main limitation of this system is that it is approximate. Significant figures tell you about precision, not accuracy. A measurement can have five sig figs and still be systematically wrong if your instrument is miscalibrated. The sig fig rules do not catch that. They only reflect the precision of the numbers you are given. If you need tighter error bounds, use uncertainty propagation instead, which gives you actual confidence intervals rather than a rough digit count. For homework sets, do the calculations by hand first to make sure you understand the counting, then verify with a calculator. Writing out the sig fig count for each number before you start prevents most mistakes. It adds maybe thirty seconds per problem but saves you from having to redo three questions because you rounded wrong.
Get the Full Details
