Statistics

Measurement of deflection

Measurement of deflection in statistical and analytical contexts investigates the deviation, dispersion, or shift of a continuous or discrete distribution from a baseline or expected model.

For: UP PGT Mathematics UP TGT Mathematics DSSSB PGT / TGT KVS / NVS PGT LT Grade

📖 Topic progress

○ Mark as complete

Definition and statement

Theorem of Mean Deflection and Quadratic Deviation Bounds

Let $X$ be a random variable with probability density function $f(x)$ on $[a, b]$, or a continuous function representing error profiles. The integral deflection measure $D_2 = \int_{a}^{b} (x - \mu)^2 f(x)\,dx$ attains its minimum when $\mu$ is the expected mean, bounded precisely by the domain span.

📐 Geometric Interpretation

Geometrically, deflection measures the root-mean-square or absolute average displacement of curve ordinates or data points from a central axis of symmetry or mean regression line.

⚙️ Algebraic & Structural Meaning

Algebraically, it establishes minimization properties of quadratic forms and links variance directly to second-order differential operators and Taylor polynomial remainders.

Conditions and hypotheses

Check each condition and what changes when it is omitted:

Hypothesis / Condition	Domain / Interval	Necessity	What Fails If Condition Dropped
Continuity of Distribution / Function	`[a, b] (Closed interval)`	Strictly Mandatory	Discontinuities lead to improper Riemann integrals and undefined expectation moments.
Square Integrability ($L^2$ condition)	`(a, b) (Open interval)`	Strictly Mandatory	Infinite variance or undefined deflection bounds if the second moment diverges.
Normalisation of Weight / Probability	`Entire Support`	Essential for relative deflection	Total probability not summing to unity ruins scaling factors.

Useful results and shortcuts

Use these identities and results when their conditions apply:

Variance / Deflection Shift Invariance

$$ \text{Var}(aX + b) = a^2 \text{Var}(X) $$

Rule: Adding a constant shifts the mean but leaves deflection (variance/spread) completely unaltered. Scaling multiplies deflection by the square of the scalar.

Application: Directly tested in UP PGT to find new variance after linear transformation of data.

Raw Moments to Central Moments Relation

$$ \mu_2 = \mu_2' - (\mu_1')^2 $$

Rule: Use second raw moment minus square of first raw moment to compute deflection instantly without shifting individual data points.

Application: Saves 2 minutes of calculation in statistics numerical problems.

Uniform Distribution Deflection Formula

$$ \sigma^2 = \frac{(b - a)^2}{12} $$

Rule: For any continuous uniform distribution over $[a, b]$, the variance (deflection measure) is immediately given by the length squared divided by 12.

Root Mean Square Deflection for Discrete Sets

$$ \text{RMS} = \sqrt{\frac{1}{n} \sum_{i=1}^{n} x_i^2} $$

Rule: When measuring deflection from the origin, bypass mean calculation if data is already centered.

Application: Frequently featured in DSSSB PGT mathematics sections.

Chebyshev's Deflection Bound

$$ P(|X - \mu| \ge k\sigma) \le \frac{1}{k^2} $$

Rule: Instantly bound the probability of extreme deflection using reciprocal of squared standard deviations.

Rapid Option Elimination via Dimensional Analysis

$$ [\text{Deflection}] = [\text{Data Unit}]^2 $$

Rule: Variance and mean square deviations must have squared units. Eliminate options with linear units immediately.

Application: Used to eliminate 2 incorrect options in quantitative statistics items.

Common errors and counterexamples

These examples show how a conclusion can fail when a condition is missing:

⚠️ COMMON ERROR #01

Confusing Mean Deviation with Standard Deviation

Common Pitfall: Using absolute linear deviations when quadratic root-mean-square deflection is explicitly demanded by the problem statement.

Counterexample: Data set: $\{-2, 0, 2\}$

Why it fails: Mean absolute deviation about mean is $\frac{4}{3}$, whereas standard deviation is $\sqrt{\frac{8}{3}}$. Mixing formulas yields wrong option selection.

🛡️ Exam Defense: Always check whether the question specifies 'Mean Deviation' (absolute) or 'Standard Deviation/Variance' (quadratic).

⚠️ COMMON ERROR #02

Origin Shift Neglect in Scaling

Common Pitfall: Applying linear scaling factors directly to standard deviation without taking the absolute value ($|a|$) or squaring for variance.

Counterexample: If $Y = -3X$, then $\text{SD}(Y) = |-3| \text{SD}(X) = 3 \text{SD}(X)$.

Why it fails: Dropping the absolute value leads to negative standard deviation errors.

🛡️ Exam Defense: Remember that scale factor for standard deviation is always positive: $|a|$.

Comparison

Concept / Theorem	Hypotheses / Assumptions	Key Formula / Outcome	Geometric Meaning	Exam Identification Clue
Range Deflection	Bounded domain $[a, b]$	$R = x_{\max} - x_{\min}$	Total span of spread	Keywords: 'extreme values', 'span'
Mean Deviation	Continuous or discrete data	$\frac{1}{n}\sum \|x_i - A\|$	Average absolute distance from center	Keywords: 'mean deviation about mean/median'
Standard Deviation	Finite second moment	$\sqrt{\frac{1}{n}\sum (x_i - \bar{x})^2}$	Root-mean-square quadratic deflection	Keywords: 'standard deviation', 'variance'
Root Mean Square (RMS)	Any real data set	$\sqrt{\frac{1}{n}\sum x_i^2}$	Deflection distance from origin	Keywords: 'RMS value', 'root mean square'

Solved examples

Each problem includes a full solution and a shorter method:

CANONICAL PROBLEM 01 Dual-Method Breakdown

If the variance of a set of observations $1, 2, 3, \dots, n$ is $10$, then the value of $n$ is:

(A) $10$(B) $11$(C) $12$(D) $19$

📚 Method A: Formal Textbook Derivation

Step 1: Use standard formula for variance of first $n$ natural numbers: $\sigma^2 = \frac{n^2 - 1}{12}$. Step 2: Equate to given variance: $\frac{n^2 - 1}{12} = 10$. Step 3: Solve $n^2 - 1 = 120 \implies n^2 = 121 \implies n = 11$. Wait, let us check: for $n=11$, variance is $(121-1)/12 = 10$. Wait, checking options: if $n=19$, $(361-1)/12 = 30$. Let's re-verify formula: variance of $1..n$ is $(n^2-1)/12$. If variance is 10, $(n^2-1)=120 \implies n^2=121 \implies n=11$. Option (B) is 11.

⚡ Method B: 25-Second Exam Shortcut

Memorize standard variance of first $n$ naturals as $\frac{n^2-1}{12}$. Set equal to 10, solve $n^2 = 121 \implies n = 11$.

💡 Tip: Standard UP PGT direct formula application. Do not waste time deriving.

CANONICAL PROBLEM 02 Dual-Method Breakdown

The variance of 20 observations is 5. If each observation is multiplied by 2, the new variance of the resulting observations is:

(A) $5$(B) $10$(C) $20$(D) $40$

📚 Method A: Formal Textbook Derivation

Step 1: Let original observations be $x_i$ with $\text{Var}(X) = 5$. Step 2: New observations $Y = 2X$. Step 3: $\text{Var}(Y) = \text{Var}(2X) = 2^2 \text{Var}(X) = 4 \times 5 = 20$.

⚡ Method B: 25-Second Exam Shortcut

Multiply old variance by square of the scaling factor: $5 \times (2)^2 = 20$. Done in 5 seconds.

💡 Tip: Watch out whether the question asks for Standard Deviation (multiply by $|a|$ ) or Variance (multiply by $a^2$ ).

Practice questions

Choose an answer to check it and see the solution:

MCQ 1: If standard deviation of $x_1, x_2, \dots, x_n$ is $\sigma$, then standard deviation of $ax_1, ax_2, \dots, ax_n$ ($a > 0$) is:

A $a\sigma$

B $a^2\sigma$

C $\sigma/a$

D $\sigma$

Derivation & Analysis:

Standard deviation scales linearly with positive factor $a$: $\text{SD}(aX) = a\text{SD}(X) = a\sigma$.

MCQ 2: The variance of the first $n$ even natural numbers is:

A $\frac{n^2-1}{4}$

B $\frac{n^2-1}{12}$

C $\frac{n^2-1}{3}$

D $\frac{n^2}{12}$

Derivation & Analysis:

Even natural numbers are $2, 4, \dots, 2n$, which is $2(1, 2, \dots, n)$. Variance scales by $2^2 = 4$, so $4 \times \frac{n^2-1}{12} = \frac{n^2-1}{3}$.

MCQ 3: If the mean deviation of a normal distribution about its mean is 12, its standard deviation approximately equals:

A $10$

B $15$

C $12$

D $18$

Derivation & Analysis:

For normal distribution, $\text{Mean Deviation} \approx 0.8 \times \text{SD}$. Thus $12 \approx \frac{4}{5}\sigma \implies \sigma = 15$.

MCQ 4: Sum of squares of deviations of 10 observations measured from 4 is 50, and sum of deviations is -20. The variance is:

A $1$

B $2$

C $5$

D $10$

Derivation & Analysis:

Use shifting property: $\sum (x_i - 4)^2 = 50$ and $\sum(x_i - 4) = -20$. Variance is independent of origin shift: $\sigma^2 = \frac{50}{10} - \left(\frac{-20}{10}\right)^2 = 5 - 4 = 1$.

MCQ 5: Which of the following measures of deflection is independent of the unit of measurement?

A Variance

B Standard Deviation

C Coefficient of Variation

D Mean Absolute Deviation

Derivation & Analysis:

Coefficient of Variation ($\text{CV} = \frac{\sigma}{\bar{x}} \times 100$) is dimensionless, hence independent of units.

MCQ 6: If $\sum x_i = 20$, $\sum x_i^2 = 200$, and $n = 10$, the variance is:

A $16$

B $\sqrt{16}$

C $20$

D $4$

Derivation & Analysis:

Variance $\sigma^2 = \frac{\sum x_i^2}{n} - \left(\frac{\sum x_i}{n}\right)^2 = \frac{200}{10} - (2)^2 = 20 - 4 = 16$.

MCQ 7: The root mean square deviation is minimum when measured from:

A Mean

B Median

C Mode

D Origin

Derivation & Analysis:

By least squares property, root mean square deviation (standard deviation) is minimum when measured about the arithmetic mean.

MCQ 8: Assertion (A): Standard deviation is unaffected by change of origin. Reason (R): Adding a constant to all observations shifts the mean but does not alter dispersion.

A Both (A) and (R) are true and (R) is correct explanation

B Both (A) and (R) are true but (R) is not explanation

C (A) is true, (R) is false

D (A) is false, (R) is true

Derivation & Analysis:

Standard deviation measures spread, which remains invariant under translation (change of origin).

MCQ 9: If $X$ and $Y$ are independent random variables, then $\text{Var}(X - Y)$ equals:

A $\text{Var}(X) - \text{Var}(Y)$

B $\text{Var}(X) + \text{Var}(Y)$

C $\text{Var}(X) + \text{Var}(Y) - 2\text{Cov}(X,Y)$

D Zero

Derivation & Analysis:

For independent variables, $\text{Var}(X \pm Y) = \text{Var}(X) + \text{Var}(Y)$.

Q10

MCQ 10: The variance of first $n$ natural numbers is:

A $\frac{n^2-1}{12}$

B $\frac{n(n+1)}{2}$

C $\frac{n^2}{2}$

D $\frac{n^2-1}{6}$

Derivation & Analysis:

Standard textbook formula for variance of $1, 2, \dots, n$ is $\frac{n^2 - 1}{12}$.

Common questions

Variance squares the deviations to prevent positive and negative deviations from cancelling out, making it a quadratic deflection measure. Standard deviation takes the square root to restore the original dimensional units.

Confusing whether a scaling factor affects standard deviation linearly ($|a|$) or variance quadratically ($a^2$), or mixing up change of origin rules with change of scale rules.

Memorize standard results like variance of first $n$ natural numbers ($\frac{n^2-1}{12}$), uniform distribution variance ($\frac{(b-a)^2}{12}$), and property shifts to bypass lengthy algebraic summation.

← Previous Topic

Index

Next Topic →

Pictorial representation of statistical data