Errors in Measurement

Every physical measurement you take in the NEET physics lab — be it a length with a vernier, a time with a stopwatch, or a voltage with a multimeter — carries an unavoidable uncertainty. Understanding how these uncertainties behave and combine is what separates a careful experimenter from a sloppy one, and NEET tests it almost every year.

Definition: An error is the difference between the measured value of a quantity and its true (or accepted) value. Errors are not mistakes — mistakes can be avoided by repeating the experiment; errors are inherent to the measurement process.

Two Big Families of Errors

Errors fall into two broad classes, and the very first MCQ trick is to separate them correctly.

Systematic errors are reproducible biases — they push the reading in one consistent direction every time. Their sources include:

• Instrumental errors: a vernier caliper with a non-zero error, a meter scale whose end has worn off, a thermometer that always reads 0.5 degC high.
• Imperfect experimental technique: measuring the temperature of a hot body in a draughty room so it cools while you read it.
• Personal errors: an observer who always presses the stopwatch a fraction late, or who consistently reads the meniscus from above.

The defining feature of a systematic error is that it is one-directional — and because it is one-directional, it is correctable. Once you discover it, you can subtract the zero error or apply a calibration correction.

Random errors are the irregular wobbles around the true value caused by fluctuating conditions you cannot control — small temperature changes, air currents, observer reaction-time variation, mains voltage hum. They are equally likely to be positive or negative, so repeating the experiment many times and taking the average reduces them.

Quantifying the Error

For a quantity measured several times, NEET expects you to compute four numbers in order.

Definition: Absolute error of one reading = |true value - measured value|. When the true value is unknown, we use the mean of the readings as our best estimate of the true value.

Definition: Mean absolute error = the arithmetic average of the absolute errors of all individual readings. It is written as Δa_mean = (|Δa_1| + |Δa_2| + ... + |Δa_n|) / n.

Definition: Relative error = Mean absolute error / Mean value, i.e. Δa_mean / a_mean. It is a pure number — a fraction — with no units.

Definition: Percentage error = Relative error × 100, expressed as a percentage. This is the form most commonly asked in MCQs.

The point of moving from absolute to relative is that a 1 mm error means very little in a 1-metre rod but is catastrophic in a 1-cm wire. Relative error puts the uncertainty in proportion.

The Three Combination Rules

When derived quantities are computed from measured ones, errors propagate. NEET tests three rules; learn them as a single block.

Rule 1 — Sum or Difference. If Z = A + B or Z = A − B, the absolute errors add:

ΔZ = ΔA + ΔB

Notice that for both addition and subtraction the maximum absolute uncertainty is the sum, never the difference, because the worst case is when both errors push the same way.

Rule 2 — Product or Quotient. If Z = A × B or Z = A / B, the relative errors add:

ΔZ/Z = ΔA/A + ΔB/B

Again, even for division, fractional errors add — the same worst-case logic.

Rule 3 — Power. If Z = A^n, then

ΔZ/Z = n × (ΔA/A)

If the formula has several powers, e.g. Z = A^p × B^q / C^r, the relative errors combine as

ΔZ/Z = p(ΔA/A) + q(ΔB/B) + r(ΔC/C)

— powers come out as multipliers, and the sign of the power does NOT matter for error propagation.

Why it matters

Every time you set up an experiment for NEET-style numerical reasoning — measuring g with a simple pendulum, finding Young's modulus from a Searle's apparatus, calculating resistance through a meter bridge — you are computing a derived quantity from several measured ones. The combination rules tell you which measurement is the weakest link and therefore where you should put the best instrument or the most careful technique.

Worked example

Question: In an experiment, the period of a pendulum is T = 2π√(L/g). The length L = (100 ± 1) cm is measured with a metre scale and the time period T = (2.00 ± 0.01) s is measured with a stopwatch. Find the percentage error in the calculated value of g.

Solution:
Step 1: Solve the formula for g: g = 4π²L / T².
Step 2: Apply Rule 2 and Rule 3 — the relative error in g is the sum of relative errors of L (power 1) and T (power 2):
Δg/g = ΔL/L + 2(ΔT/T)
Step 3: Plug in numbers: ΔL/L = 1/100 = 0.01 and ΔT/T = 0.01/2.00 = 0.005.
Δg/g = 0.01 + 2(0.005) = 0.01 + 0.01 = 0.02
Step 4: Convert to percentage: 0.02 × 100 = 2%.
Conclusion: The percentage error in g is 2%. Note how the time measurement, despite appearing more accurate, contributes equally to the error because it enters with power 2.

Real-world example

Pharmacists in Indian hospitals titrate intravenous drips by measuring drops per minute. A 5% error in drop volume can become a much bigger error in dose because dose = (drops/min) × (volume/drop) × (concentration) — a product of three measured quantities, all carrying their own errors. The error-combination logic of NEET physics is literally the same logic that ensures patient safety.

Common misconception

Many students think that in subtraction, errors should subtract. They write ΔZ = ΔA − ΔB. This is wrong. Errors are uncertainties — you do not know whether A is too high or too low — so in the worst case, both push in the same direction. For both sum and difference, absolute errors ADD. Similarly, fractional errors add for both products and quotients.

Significant figures convey precision. Rules: (1) all non-zero digits are significant; (2) zeros between non-zero digits are significant (1002 has 4); (3) leading zeros are NOT significant (0.005 has 1); (4) trailing zeros after a decimal ARE significant (2.300 has 4); (5) trailing zeros in a number without a decimal are ambiguous (use scientific notation). In addition/subtraction, the result keeps the least number of DECIMAL PLACES. In multiplication/division, the result keeps the least number of SIGNIFICANT FIGURES. Rounding: if the digit to drop is 5 with nothing after, round to make the preceding digit even.

Error combination is one of those NEET-Physics topics where one tiny rule decides the answer — and one tiny slip wipes out the mark. Master the power rule for maximum percentage error and the whole chapter shrinks to a one-line check.

Definition: When a derived quantity is expressed as a product or quotient of measured quantities, each raised to some power, the maximum fractional (or percentage) error in the result is obtained by adding the fractional errors of the measured quantities, each multiplied by the magnitude of its power.

The general power rule

If P is a quantity defined by

P = a^p · b^q · c^r · d^s

(where p, q, r, s are real numbers — positive, negative or fractional), then the maximum fractional error in P is

ΔP / P = |p| (Δa / a) + |q| (Δb / b) + |r| (Δc / c) + |s| (Δd / d)

Multiply both sides by 100 to convert to percentage error:

(ΔP / P) × 100 = |p| (Δa/a × 100) + |q| (Δb/b × 100) + |r| (Δc/c × 100) + |s| (Δd/d × 100)

Two things to internalise:

1. The signs of the powers do not matter for maximum error. A term in the denominator contributes just like a term in the numerator — both add to the worst case.
2. Roots are fractional powers. √c means c^(1/2), so its contribution is (1/2) × (Δc / c × 100).

Why it matters: NEET, JEE, AIIMS and every State CET Physics paper asks at least one error-combination question. Most are direct one-line applications of this rule.

The worked NEET example

Question: A physical quantity P is defined by

P = (a^3 · b^2) / (√c · d)

The percentage errors in a, b, c and d are 1%, 3%, 4% and 2% respectively. Find the maximum percentage error in P.

Solution:

Step 1: Identify the powers of each measured quantity.

• a has power +3 (in the numerator).
• b has power +2 (in the numerator).
• c has power −1/2 (since √c = c^(1/2) is in the denominator).
• d has power −1 (in the denominator).

Step 2: Take the magnitudes of these powers — for maximum error, signs are ignored:

• |power of a| = 3
• |power of b| = 2
• |power of c| = 1/2
• |power of d| = 1

Step 3: Multiply each fractional error by the magnitude of its power:

• Contribution of a = 3 × 1% = 3%
• Contribution of b = 2 × 3% = 6%
• Contribution of c = (1/2) × 4% = 2%
• Contribution of d = 1 × 2% = 2%

Step 4: Add these contributions (always add for maximum error — never subtract, never average):

ΔP / P (%) = 3 + 6 + 2 + 2 = 13%

Conclusion: The maximum percentage error in P is 13%.

Why we always add, never subtract

Some students wonder — since c and d are in the denominator, shouldn't their contributions reduce the error? No. The +x and −x errors in different measurements are independent random events; the worst case is when they all pile up in the same direction. Mathematically, you take the differential of ln P,

ln P = 3 ln a + 2 ln b − (1/2) ln c − ln d

dP/P = 3 (da/a) + 2 (db/b) − (1/2) (dc/c) − (dd/d)

In the worst case, the sign of each da/a (etc.) can be chosen to make the right-hand side as large as possible — so we take absolute values and add. That is why the denominator's terms still add, never subtract.

Why it matters: This is exactly the trap NEET examiners set every other year. The wrong answer "3 + 6 − 2 − 2 = 5%" is always on the option list.

A second worked example

Question: The kinetic energy K = (1/2) m v². If the percentage errors in m and v are 2% and 3% respectively, find the maximum percentage error in K.

Solution:

Step 1: K depends on m^1 and v^2. (The constant 1/2 has no error contribution — constants are exact.)

Step 2: Contribution of m = 1 × 2% = 2%. Contribution of v = 2 × 3% = 6%.

Step 3: ΔK / K (%) = 2 + 6 = 8%.

Conclusion: The maximum percentage error in K is 8%.

This little example explains why velocity measurements need to be very precise — a small error in v doubles in the kinetic-energy expression because of the square. Doubling occurs because the exponent multiplies the fractional error.

A third example to lock the rule

Question: Density ρ = m / V. m is measured to 1%, V to 2%. Find max % error in ρ.

Solution: ρ depends on m^1 and V^(−1). Both contribute their full fractional errors. ΔP / P (%) = 1 + 2 = 3%.

If V is measured by V = π r² h with errors 1% in r and 2% in h, then by the rule (ΔV / V)% = 2 × 1 + 1 × 2 = 4%, and then ρ would carry 1 + 4 = 5% error. Compound the rule whenever a derived quantity itself depends on other derived quantities — apply the rule at every layer.

Common misconceptions

Common misconception: "Errors in the denominator subtract because they are 'opposite'." Wrong. Errors in the denominator add in the worst case, exactly like the numerator.

Common misconception: "Roots have no error contribution." Wrong. √c is c^(1/2); the (1/2) is its power. It scales the fractional error by 1/2 but never makes it zero.

Common misconception: "Multiply the percentage error of a constant (like 1/2 or π) by something." Wrong. Constants are exact; they do not enter the error sum.

Real-world example: In a school-lab pendulum experiment, g = 4π² L / T² with L measured to 0.5% and T measured to 1%. The maximum percentage error in g is then 1 × 0.5 + 2 × 1 = 2.5%. The doubling on T is exactly why every Class XI textbook insists on multiple readings of the period.

Speed rules for the exam

A 30-second checklist for any error-combination MCQ:

1. Write the formula and identify the exponent of every measured quantity, with sign.
2. Take the magnitude of each exponent.
3. Multiply each by the given percentage error of that quantity.
4. Add the four (or so) products. The sum is the answer.
5. Constants and pure numbers contribute nothing. Roots contribute with exponent 1/2.

A guard-rail against arithmetic slip

Quick sanity check: the answer can never be smaller than the largest single contribution. In our worked example, the largest contribution is 6% (from b^2 × 3%); the final 13% is larger than 6%, which is consistent. If you get an answer of 4% with one of the terms contributing 6%, you have made an arithmetic mistake — usually subtracting instead of adding.

Every measured value carries uncertainty, and NEET regularly asks you to propagate that error through a formula — e.g. find the % error in density from mass and radius errors. This is a scoring topic because the rules are mechanical once memorised.

Types of error

• Systematic (one-directional): instrumental (zero error), imperfect technique, personal bias. Reducible.
• Random: scatter from unpredictable causes; reduced by repeating and averaging.
• Gross: human blunders.

Quantifying error

For readings a₁, a₂, … with mean a_mean:

• Absolute error of a reading: |Δaᵢ| = |a_mean − aᵢ|
• Mean absolute error: Δa_mean = average of |Δaᵢ|
• Relative error = Δa_mean / a_mean
• Percentage error = relative error × 100%

Combination (propagation) of errors

Operation	Result Z	Error rule
Sum / Difference	Z = A ± B	ΔZ = ΔA + ΔB (absolute add)
Product / Quotient	Z = AB or A/B	ΔZ/Z = ΔA/A + ΔB/B
Power	Z = Aᵖ Bq / Cʳ	ΔZ/Z = p(ΔA/A) + q(ΔB/B) + r(ΔC/C)

Errors always add, never subtract — you take the worst case.

Exam Tricks & Tips

• 🎯 For sums and differences, the absolute errors add; for products, quotients and powers, the relative (%) errors add.
• 🎯 A power multiplies the % error: radius r appearing as r³ in volume contributes 3× its % error.
• 🎯 A difference of two nearly equal numbers gives a huge relative error — avoid such measurements.
• 🎯 Least count = smallest division; instrument error is usually taken as ± half the least count (or ±1 LC if stated).
• 🎯 Significant figures in a result follow the least-precise factor for ×/÷, and the least decimal places for +/−.
• ❌ Never let errors cancel: if Z = A/B you still ADD ΔA/A and ΔB/B, you don't subtract.

Expected exam pattern

"% error in a derived quantity" — density = m/V, g from a pendulum T = 2π√(L/g), resistance from V and I — given individual % errors; or rounding to the correct number of significant figures.

Quick recap

Sum/difference → add absolute errors. Product/quotient/power → add relative errors weighted by the power. Report the answer to the correct significant figures. Errors accumulate, never cancel.

Cover the answer, recall, then check. 12 cards on error analysis.

Q1. Name the three broad types of error.
A1. Systematic, random and gross errors.

Q2. Define absolute error of a single reading.
A2. |Δaᵢ| = |a_mean − aᵢ|, the magnitude of its deviation from the mean.

Q3. Relative error and percentage error?
A3. Relative error = Δa_mean / a_mean; percentage error = relative error × 100%.

Q4. For Z = A + B, how do errors combine?
A4. Absolute errors add: ΔZ = ΔA + ΔB.

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 = A²B³/C, the fractional error is?
A6. ΔZ/Z = 2(ΔA/A) + 3(ΔB/B) + (ΔC/C).

Q7. Density ρ = M/V with V ∝ r³. If M has 2% error and r has 1% error, % error in ρ?
A7. 2% + 3×1% = 5%.

Q8. Why is the difference of two nearly equal quantities risky?
A8. The absolute error stays the same but the small result makes the relative error very large.

Q9. How do you reduce random errors?
A9. Take many readings and average; random errors partly cancel out.

Q10. How is a systematic error like zero error corrected?
A10. Identify the fixed offset (e.g. instrument zero error) and add/subtract it from every reading.

Q11. Significant figures in a product/quotient result?
A11. Keep as many significant figures as the factor with the fewest.

Q12. Instrument (least-count) error is usually taken as?
A12. ± half the least count (or ±1 least count if the question states so).