Significant figures and errors

Significant figures (sig figs) express measurement precision. Rules: (1) All non-zero digits are significant. (2) Zeros between non-zeros are significant (e.g. 1002 has 4). (3) Leading zeros are NOT significant (0.0025 has 2). (4) Trailing zeros after a decimal point ARE significant (2.300 has 4). (5) Trailing zeros in a whole number without a decimal are ambiguous (1500 is treated as 2 sig figs unless written 1.500x10^3). Memory aid: 'Sandwiched zeros count, leading zeros don't, trailing-after-decimal do.' Scientific notation removes ambiguity: only the mantissa digits count. In JEE, exact constants (like the 2 in 2pi) have infinite sig figs and never limit precision. Always identify sig figs before applying rounding rules in numerical problems.

In a JEE problem, you can measure length to the nearest millimeter and time to the nearest hundredth of a second, but the moment you start combining these into a derived quantity like acceleration or density, those tiny measurement errors propagate. Knowing exactly how they propagate — and which formula to use when — is one of the cheapest topics to ace in the Physics and Measurement chapter.

Definition: An error in measurement is the difference between the measured value and the true value of a physical quantity.

Definition: Error propagation is the calculation of how errors in measured quantities (A, B, C ...) carry through to give the error in a derived quantity (Z) that depends on them.

The two big rules at a glance

Across all of error analysis, there are only two basic patterns to remember. Memorise these and you can derive everything else.

• For addition and subtraction (Z = A + B or Z = A - B): the absolute errors add.
• For multiplication and division (Z = AB or Z = A/B): the relative (fractional) errors add.

These two ideas drive every error MCQ in JEE Main.

Rule 1 — Addition and Subtraction (absolute errors add)

If Z = A + B or Z = A - B, then:

delta Z = delta A + delta B.

The plus sign is always used. Even for a subtraction, the maximum possible error is obtained by adding the magnitudes — this is the worst-case scenario.

Why we add and not subtract: if A is too high by delta A and B is too low by delta B, then A + B can be off by delta A + delta B in the same direction. If we wrote a minus sign, we would be hoping that the two errors cancel — that is a lucky case, not the maximum possible error.

Rule 2 — Multiplication and Division (relative errors add)

If Z = AB or Z = A / B, then:

delta Z / Z = delta A / A + delta B / B.

Again, the plus sign is always used, regardless of whether it is multiplication or division.

Definition: The relative (fractional) error in a quantity A is delta A / A — it is a dimensionless ratio. Multiplying by 100 gives the percentage error.

So a 2% error in A combined with a 3% error in B gives a 5% error in Z = AB or Z = A/B.

Rule 3 — Powers (errors multiply by exponents)

If Z = A^p B^q / C^r, then:

delta Z / Z = |p| (delta A / A) + |q| (delta B / B) + |r| (delta C / C).

The exponents come out as multipliers of the relative errors, and we use absolute values of the exponents because we are after the maximum possible error.

This is the rule that creates "shock" answers. A quantity raised to a high power dominates the total error. If your formula is Z = A^3 B, a 2% error in A becomes 6% in Z, while a 2% error in B stays at 2% — the error in A is three times more important than the error in B.

Worked example — density measurement

Question: The density of a metal cube is to be found from its mass M and side L. The relation is rho = M / L^3. The percentage errors are 1% in M and 2% in L. What is the maximum percentage error in rho?

Solution:
Step 1: Write rho in power form — rho = M^1 L^(-3).
Step 2: Apply the powers rule: delta rho / rho = 1 x (delta M / M) + |-3| x (delta L / L).
Step 3: Substitute: delta rho / rho = 1 x 1% + 3 x 2% = 1% + 6% = 7%.
Conclusion: The maximum percentage error in the density is 7%, dominated by the 6% contribution from L.

This single example captures the central JEE insight: in a derived quantity, the largest power-times-error term dominates, so reducing error in that variable is the highest-payoff design choice in an experiment.

Why we always use the worst case

A natural question is: why must we always add the magnitudes and not let them partly cancel? The answer is that the maximum permissible error is a guarantee — it tells the examiner (and the experimenter) what the worst possible deviation is, regardless of the signs of individual errors. Probabilistic cancellation is treated separately as statistical error, not here. JEE Main always asks for the maximum error unless explicitly stated otherwise.

Rounding off the final answer

After applying the rules, the final answer must be rounded sensibly:

• For addition and subtraction results: round to the least number of decimal places among the inputs.
• For multiplication and division results: round to the least number of significant figures among the inputs.

So if you add 12.3 m and 4.56 m, you write 16.9 m (one decimal place), not 16.86 m. If you multiply 1.234 cm by 2.0 cm, you write 2.5 cm^2 (two significant figures), not 2.468 cm^2.

Real-world example

Real-world example: In a JEE Main lab to find the value of g using a simple pendulum, the formula g = 4 pi^2 L / T^2 involves T raised to the power -2. If a student's stopwatch error is 0.1 s in a 20 s period (about 0.5%), it contributes 1% to the error in g. A 1 mm error in a 1 m length (0.1%) contributes only 0.1%. Conclusion: a more accurate stopwatch matters far more than a more accurate ruler in this experiment — straight out of error propagation.

Common misconception

Common misconception: "For subtraction, absolute errors should be subtracted to find the error." Wrong. Absolute errors are always added, even for subtraction. The worst case is when both errors push in the same direction.

Another wrong idea: "Higher powers reduce errors because the result becomes precise." Wrong — exactly the opposite. Higher powers multiply the relative error of that quantity, making them the dominant error contributors.

A third trap: "Percentage error = absolute error." Wrong. Percentage error = (delta Z / Z) x 100 — it is a relative quantity. Absolute error is in the same units as the quantity itself.

Worked example — combining rules

Question: A physical quantity P is given by P = (A^2 B^3) / (sqrt(C) D). The percentage errors are: A = 1%, B = 2%, C = 4%, D = 3%. Find the maximum percentage error in P.

Solution:
Step 1: Identify exponents — A: 2, B: 3, C: 1/2, D: 1 (all in absolute value).
Step 2: Apply the powers rule: delta P / P = 2(delta A / A) + 3(delta B / B) + (1/2)(delta C / C) + 1(delta D / D).
Step 3: Substitute: delta P / P = 2(1) + 3(2) + (1/2)(4) + 1(3) percent = 2 + 6 + 2 + 3 = 13%.
Conclusion: The maximum percentage error in P is 13%, with B (6%) the dominant contributor.

This is the structure of nearly every JEE Main error MCQ — assign each variable an exponent, multiply by its percentage error, then add.

Why it matters

Why it matters: Error propagation is the most predictable scoring topic in the Units and Measurement chapter — JEE Main asks one question on it almost every year, and it follows one of the patterns above. Practising 5-10 power-rule problems builds enough speed to solve any such MCQ under 60 seconds.

A small cube on a lab bench seems like an innocent object — until the question asks for the percentage error in its density. This is one of the cleanest places to see why JEE Main keeps recycling error analysis: a tiny relative error in the side length quietly becomes the biggest contributor to the final uncertainty, simply because volume depends on the cube of length.

Definition: Absolute error in a measurement x is the smallest amount by which the measured value can be wrong, written as Δx. Relative error is Δx/x, and percentage error is (Δx/x) × 100%.

Definition: Propagation of error is the rule that tells you how uncertainty in input quantities flows into a quantity computed from them. For a product or quotient like q = A^p · B^q · C^r, the relative errors add — and each is multiplied by the power.

Setting up the problem

We are given a cube with measured mass m = 10.0 g carrying an absolute error Δm = 0.1 g, and measured side L = 2.00 cm with ΔL = 0.01 cm. We need the density ρ = m/V and the percentage uncertainty in it.

Since the cube has side L, its volume is V = L³, so density becomes:

ρ = m / L³ = m · L^(-3)

This is a product-of-powers expression, exactly the form where the propagation rule applies cleanly.

Step-by-step computation

Question: Find ρ and its percentage error for the cube above.

Solution:

Step 1 — Compute the nominal density.
ρ = 10.0 / (2.00)³ = 10.0 / 8.00 = 1.25 g/cm³.

Step 2 — Apply the propagation rule. Because ρ = m · L^(-3), the magnitudes of the relative errors add, with each multiplied by the absolute value of its exponent:

Δρ/ρ = Δm/m + 3 · (ΔL/L)

Step 3 — Plug in the numbers.
Δm/m = 0.1 / 10.0 = 0.010
3 · ΔL/L = 3 × (0.01 / 2.00) = 3 × 0.005 = 0.015
Δρ/ρ = 0.010 + 0.015 = 0.025

Step 4 — Convert to percentage and find Δρ.
Percentage error = 2.5%.
Δρ = 0.025 × 1.25 g/cm³ ≈ 0.03 g/cm³.

Conclusion: ρ = 1.25 ± 0.03 g/cm³, i.e. a 2.5% uncertainty.

Why the length error dominates

Notice the punchline: ΔL/L (0.5%) was actually smaller than Δm/m (1%). Yet because the side enters as L³, the side's relative error gets a factor of 3, growing it to 1.5% — now larger than the mass contribution.

This is the heart of the power-multiplication rule: whenever a quantity appears raised to a high power, its relative error matters disproportionately. In a sphere, V = (4/3)πr³, so radius error is again tripled. In kinetic energy K = (1/2)mv², the velocity error is doubled. Question-setters love this asymmetry because it punishes students who add relative errors blindly without weighting by the power.

Why it matters: In every experiment in your lab manual — from simple pendulum's g to Young's modulus to refractive index — the final reported error is dominated by the quantity that is either raised to the highest power or measured with the coarsest instrument. Identifying that dominant term tells you which instrument to upgrade first if you want a better result.

Real-world example: A jeweller verifying that a small gold cube is solid (not plated) measures its mass on a digital balance and its side with vernier calipers. Pure gold density is about 19.3 g/cm³. To distinguish, say, 19.3 from 18.5 confidently, the jeweller needs the combined relative error well under 2%. A 0.1 mm error on a 5 mm side is already 6% in density — so for small samples, the calipers, not the balance, set the trust limit.

Common misconception: Many students think the absolute errors add: Δρ = Δm + 3ΔL, or that the side's error contributes the same way as mass's. Both are wrong. It is the relative (percentage) errors that add, and only when the formula is a product of powers. For sums and differences, the absolute errors add instead — never mix the two rules.

Reading the result correctly

Once we know Δρ ≈ 0.03 g/cm³, the answer must be reported with the right number of significant figures. The rule of thumb: round the absolute error to one significant figure, and round the value to the same decimal place. Here Δρ has one sig fig in the hundredths place, so ρ is reported to the hundredths place: 1.25 ± 0.03 g/cm³. Reporting "1.2500 ± 0.0312" is technically incorrect — the trailing digits pretend to a precision the data does not support.

If the question changes the numbers slightly (say L = 2.0 ± 0.1 cm), reapply the formula — do not memorise just the 2.5% answer. The whole point is the method.

Cover the answer, recall, then check. 12 cards on significant figures and error propagation for JEE Main.

Q1. How many significant figures are in 0.00420?
A1. Three (4, 2 and the trailing 0). Leading zeros are never significant; trailing zeros after a decimal are.

Q2. Rule for significant figures in multiplication/division.
A2. The result keeps as many significant figures as the factor with the fewest significant figures.

Q3. Rule for significant figures in addition/subtraction.
A3. The result is rounded to the least number of decimal places among the terms.

Q4. Define absolute, relative and percentage error.
A4. Absolute error Δa = |a_mean − aᵢ|; relative error = Δa_mean / a_mean; percentage error = (Δa_mean / a_mean) × 100%.

Q5. For Z = A × B or A / B, how do errors combine?
A5. Relative errors add: ΔZ/Z = ΔA/A + ΔB/B.

Q6. For Z = Aⁿ, what is the fractional error?
A6. ΔZ/Z = |n| · (ΔA/A). Powers multiply the fractional error.

Q7. For Z = A ± B, how does the absolute error combine?
A7. Absolute errors add: ΔZ = ΔA + ΔB (worst-case), regardless of the sign in the operation.

Q8. In Z = A²B³/√C, the fractional error is?
A8. ΔZ/Z = 2(ΔA/A) + 3(ΔB/B) + ½(ΔC/C).

Q9. How many significant figures in 6.320 × 10³?
A9. Four. Scientific notation shows exactly the significant figures; the power of ten is not counted.

Q10. Round 2.745 and 2.735 to 3 significant figures.
A10. 2.74 and 2.74 (round-half-to-even convention: preceding digit made even).

Q11. Distinguish systematic and random errors.
A11. Systematic errors have a consistent sign/cause (instrument zero error, calibration) and are correctable; random errors fluctuate unpredictably and are reduced by averaging many readings.

Q12. Which measurement contributes most to the error in a formula?
A12. The quantity raised to the highest power (its fractional error is multiplied by that power), so measure it most precisely.

Why it matters

Error analysis is a high-yield JEE Main topic — typically 1 dedicated question combining significant figures with error propagation. It is formula-light but rule-heavy, so a candidate who knows the rules scores instantly while others fumble.

Significant figure rules

• All non-zero digits are significant.
• Zeros between non-zeros are significant (2.03 → 3 s.f.).
• Leading zeros are NOT significant (0.0025 → 2 s.f.).
• Trailing zeros after a decimal ARE significant (2.500 → 4 s.f.).
• Scientific notation shows exact significant figures.

Error propagation formulae

Operation	Error combines as
Z = A + B or A − B	ΔZ = ΔA + ΔB
Z = A × B or A / B	ΔZ/Z = ΔA/A + ΔB/B
Z = Aⁿ	ΔZ/Z = n·(ΔA/A)
Z = Aᵖ Bᵍ / Cʳ	ΔZ/Z = p(ΔA/A) + q(ΔB/B) + r(ΔC/C)

Exam Tricks & Tips

• 🎯 Errors always ADD — never subtract, even for a division or subtraction. Assume worst case.
• 🎯 In addition/subtraction track decimal places; in multiplication/division track significant figures.
• 🎯 The variable with the largest power dominates the total error — questions test which measurement to refine.
• 🎯 % error of a square = 2 × % error of the base; of a square root = ½ × % error.
• 🎯 Averaging many readings reduces random error only; it cannot remove systematic (zero/calibration) error.
• ❌ Common mistake: counting leading zeros as significant — 0.00600 has only 3 significant figures (the two trailing zeros plus the 6).

Expected exam pattern

Usually one numerical: given percentage errors in measured quantities (e.g. in a formula like ρ = M/V or Y = FL/AΔL), find the percentage error in the derived quantity. Sometimes a pure significant-figures counting/rounding MCQ.

Quick recap

Non-zero digits and sandwiched/trailing-decimal zeros count; leading zeros don't. Add absolute errors for ±, add fractional errors (weighted by powers) for ×/÷/powers. Distinguish systematic vs random errors. Always take the worst-case (additive) combination.