Error analysis — types of errors, propagation rules, significant figures

Question

What are the different types of errors in measurement? How do errors propagate through calculations, and what are the rules for significant figures?

Solution — Step by Step

Systematic errors: Consistent bias in one direction. Caused by faulty instruments, wrong calibration, or personal bias. Can be minimised by improving experimental technique.

Random errors: Unpredictable variations in readings. Caused by fluctuations in conditions, observer limitations. Can be reduced by taking multiple readings and averaging.

Gross errors: Outright mistakes — misreading a scale, recording wrong data. Eliminated by careful experimentation and rejecting outliers.

If measured quantities have absolute errors $\Delta a$ and $\Delta b$ :

Addition/Subtraction ( $z = a \pm b$ ): Absolute errors add.

\Delta z = \Delta a + \Delta b

Multiplication/Division ( $z = a \times b$ or $z = a/b$ ): Relative (percentage) errors add.

\frac{\Delta z}{z} = \frac{\Delta a}{a} + \frac{\Delta b}{b}

Power ( $z = a^n$ ): Relative error multiplies by the power.

\frac{\Delta z}{z} = n \cdot \frac{\Delta a}{a}

For a formula like $z = \frac{a^p \cdot b^q}{c^r}$ :

\frac{\Delta z}{z} = p\frac{\Delta a}{a} + q\frac{\Delta b}{b} + r\frac{\Delta c}{c}

All non-zero digits are significant: 345 has 3 significant figures
Zeros between non-zero digits are significant: 3045 has 4
Leading zeros are NOT significant: 0.0045 has 2
Trailing zeros after decimal point ARE significant: 3.40 has 3
Trailing zeros without decimal point are ambiguous: 2300 could be 2, 3, or 4

In calculations: The result should have the same number of significant figures as the least precise input (for multiplication/division) or the same number of decimal places as the least precise input (for addition/subtraction).

graph TD
    A[Error Types] --> B[Systematic: consistent bias]
    A --> C[Random: statistical fluctuations]
    A --> D[Gross: mistakes]
    E[Error Propagation] --> F{Operation?}
    F -->|Addition/Subtraction| G[Add absolute errors]
    F -->|Multiplication/Division| H[Add relative errors]
    F -->|Power: a to n| I["Multiply relative error by n"]

Why This Works

Error propagation follows from calculus — specifically, the total differential of a function. For $z = f(a, b)$ :

\Delta z = \left|\frac{\partial f}{\partial a}\right|\Delta a + \left|\frac{\partial f}{\partial b}\right|\Delta b

The formulas above are special cases of this general rule. For multiplication ( $z = ab$ ), $\partial z/\partial a = b$ , so $\Delta z/z = \Delta a/a + \Delta b/b$ .

The key exam insight: the quantity with the highest power in a formula contributes the most to the error. For example, in $T = 2\pi\sqrt{L/g}$ , if we measure $T$ to find $g$ :

g = \frac{4\pi^2 L}{T^2} \implies \frac{\Delta g}{g} = \frac{\Delta L}{L} + 2\frac{\Delta T}{T}

The error in $T$ is doubled because $T$ appears squared. So measuring time accurately matters more.

Alternative Method

For JEE problems asking “which measurement should be most accurate to minimise error?”:

Look at the power of each variable in the formula. The variable with the highest power needs the most accurate measurement.

Example: Resistivity $\rho = \frac{R \cdot \pi d^2}{4L}$ . Here $d$ has power 2. If percentage errors in $R$ , $d$ , and $L$ are each 1%, the error in $\rho$ is: $1 + 2(1) + 1 = 4\%$ . The diameter measurement contributes the most.

Common Mistake

Students confuse absolute error and relative error in propagation rules. For addition/subtraction, we add absolute errors (not percentage). For multiplication/division, we add relative (percentage) errors (not absolute). Mixing these up gives completely wrong results. JEE Main 2024 tested this by giving a formula with both addition and multiplication — students had to apply different rules at different stages.

Question

Solution — Step by Step

Why This Works

Alternative Method

Common Mistake

Try These Next