MCQ 1

Let X ∼ Binomial(n , p). What is Pr ⁡(X =0)?

(a) pn
(b) ¿
(c) np (1− p)
n!
(d) p¿
(n−1)! 1 !
Answer: (b) ¿
Steps / Explanation:

n k
 Binomial: probability of k successes in n trials is Pr ⁡(X =k )=( ) p ¿ .
k
n 0
 For k =0: Pr ⁡(X =0)=( ) p ¿.
0
 Intuition: all n trials independently fail.

MCQ 2

Suppose X ∼ Poisson(λ). Then Pr ⁡(X =1)is

(a) λ e− λ
(b) e− λ
(c) λ
(d) 1−e−λ

Answer: (a) λ e− λ

Steps / Explanation:
k −λ
λ e
 Poisson pmf: Pr ⁡(X =k )= .
k!
1 −λ
λ e −λ
 For k =1: Pr ⁡(X =1)= =λe .
1!

MCQ 3

Let X be exponential with rate λ . The memoryless property says:

Pr ⁡(X > s +t ∣ X> s )=¿?
(a) Pr ⁡(X > t)
(b) Pr ⁡(X > s)
(c) Pr ⁡(X > s) Pr ⁡(X > t)
(d) Pr ⁡(X > s)+ Pr ⁡( X >t)

Answer: (a) Pr ⁡(X > t)

Steps / Explanation:
  Exponential survival: Pr ⁡(X > u)=e− λu.

Pr ⁡( X > s+t) e− λ(s+t ) −λt

 Conditional: Pr ⁡(X > s +t ∣ X> s )= = − λs =e =Pr ⁡( X >t).
Pr ⁡(X > s) e
 This shows the distribution of remaining time doesn't depend on elapsed
time: memoryless.

MCQ 4

If X ∼ N ( μ , σ 2), approximate Pr ⁡(μ−σ ≤ X ≤ μ+ σ) (empirical rule).

(a) 50%
(b) 68%
(c) 95%
(d) 99.7%
Answer: (b) 68%
Steps / Explanation:
 Empirical (68–95–99.7) rule: ~68% within ±1σ, ~95% within ±2σ, ~99.7%
within ±3σ.
 So probability within one standard deviation ≈ 68%.

MCQ 5

X́−μ
Let X 1 , … , X ni.i.d. with mean μ, variance σ 2. By CLT, distribution of
σ /√ n
tends to:
(a) Student’s t
(b) Chi-square
(c) Standard normal N (0 ,1)
(d) Uniform

Answer: (c) Standard normal N (0 ,1)

Steps / Explanation:
 Central Limit Theorem: standardized sample mean converges in
distribution to N (0 ,1)as n → ∞, regardless of original distribution (under
mild conditions).

MCQ 6

2 1
For a sample of size n , the sample variance S = ∑ ¿. What is E [S2 ]?
n−1
(a) σ 2
n−1 2
(b) σ
n
 n 2
(c) σ
n−1
(d) σ

Answer: (a) σ 2

Steps / Explanation:
2 2
 S as defined is the unbiased estimator of population variance σ . Hence
2 2
E [S ]=σ .

MCQ 7

If X ∼U (a , b). What is Var ( X)?

(a) ¿ ¿
(b) ¿ ¿
b−a
(c)
12
b−a
(d)
2
Answer: (a) ¿ ¿
Steps / Explanation:

 For uniform on [a , b]: mean μ=(a+ b)/2, variance ¿ ¿. Derivation via ∫ ¿ .

MCQ 8
α
β α −1 −βx
Gamma(α , β ) with rate β (pdf x e ). What is E [X ]?
Γ (α )
α
(a)
β
β
(b)
α
(c) αβ
1
(d)
αβ
α
Answer: (a)
β
Steps / Explanation:

 Standard gamma mean (shape α, rate β) is α / β . (If parameterized with

scale θ, mean = αθ.)

MCQ 9
Given joint pdf f X , Y (x , y) . The marginal f X (x )= ∫ f X ,Y (x , y )dy . Which is true?
(a) The marginal integrates to 1 over x .
(b) The marginal integrates to 1 over y .
(c) ∬ f X , Y (x , y) dx dy =∫ f X (x ) dx .
(d) Both (a) and (c).
Answer: (d) Both (a) and (c)
Steps / Explanation:

 Marginal f X (x )integrates over x to 1: ∫ f X ( x)dx=∬ f X , Y (x , y)dy dx=1 .

 Option (c) is the equality used to show (a). Option (b) is wrong (marginal of
X not integrated over y ).

MCQ 10

For bivariate normal (X , Y )with correlation ρ , conditional variance

2 2
Var (Y ∣ X=x)=σ Y (1− ρ ). Which part shows it does not depend on x ?
(a) μY ∣ X
(b) σ 2Y (1−ρ2)
(c) The correlation term
(d) None of the above

Answer: (b) σ 2Y (1−ρ2)

Steps / Explanation:

 Conditional variance formula contains no x — it's constant regardless of the

observed x . This is what option (b) shows.

MCQ 11
2
(n−1) S
If X 1 , … , X n ∼ iidN (μ , σ 2 ), then 2
has distribution:
σ
(a) N (0 ,1)
(b) χ 2n−1
(c) t n−1
(d) F

Answer: (b) χ 2n−1

Steps / Explanation:

 For normal samples, (n−1) S2 /σ 2equals sum of squared standard normals;

thus chi-square with n−1degrees of freedom.

MCQ 12
In simple linear regression Y = β0 + β 1 X +ε with ε ∼ N ( 0 , σ 2) , which is true
about ^β 1?
(a) unbiased
2
σ
(b) Var ( β^ 1 )=
∑ ¿¿
(c) consistent
(d) all of the above
Answer: (d) all of the above)
Steps / Explanation:

 Under standard assumptions: ^β 1is unbiased, variance equals that formula,

and as n → ∞it is consistent.

MCQ 13
Which is NOT a Gauss-Markov assumption?
(a) Linearity in parameters
(b) No perfect multicollinearity
(c) E [ε ∣ X ]=0
(d) ε follows a uniform distribution

Answer: (d) ε follows a uniform distribution)

Steps / Explanation:
 Gauss-Markov assumptions require linearity, exogeneity, no perfect
multicollinearity, homoskedastic errors, etc. Error distribution need not be
uniform — often normality is assumed for inference, but uniform is not
required.

MCQ 14
In multiple regression, adding more regressors (even irrelevant) will:
(a) increase or keep R2same
(b) decrease R2
(c) keep R2same always
(d) may increase or decrease

Answer: (a) increase or keep R2same

Steps / Explanation:

2 SSR
 R =1− . Adding regressors cannot increase SSR (residual sum of
SST
squares) — it can only decrease or leave it same (OLS can always choose
coefficients that reduce SSR), so R2cannot decrease.
MCQ 15

AR(1): X t =ϕ X t −1+ ut with ∣ ϕ ∣< 1. Stationary variance Var ( X t )=¿?

2
σu
(a) 2
1−ϕ
2
σu
(b)
1−ϕ
(c) σ 2u (1−ϕ 2 )
(d) σ 2u
2
σu
Answer: (a) 2
1−ϕ
Steps / Explanation / Derivation:

 Let v=Var (X t ). From model and independence of ut :

2 2
v=Var (ϕ X t −1+u t )=ϕ v +σ u.
2
2 2 σu
 Solve: v (1−ϕ )=σ u→ v= 2
.
1−ϕ
 Valid when ∣ ϕ ∣< 1(stationarity).

MCQ 16
Type I error is:
(a) rejecting H 0when H 0is true
(b) accepting H 0when H 0is false
(c) rejecting H 1when H 1is true
(d) none of the above

Answer: (a) rejecting H 0when H 0is true

Steps / Explanation:
 Type I = false positive (reject true null). Type II = false negative (fail to
reject false null).

MCQ 17

Test statistic for population mean when σ unknown and n small:

X́−μ0
(a) Z=
σ /√n
(b) χ 2statistic
X́−μ 0
(c) t=
S /√n
(d) F-statistic
 X́−μ 0
Answer: (c) t=
S /√n
Steps / Explanation:

 When sample variance S2used in place of population σ 2, the correct

statistic follows Student’s t -distribution with n−1degrees of freedom (for
normal data).

MCQ 18

If X and Y independent with MGFs M X (t), M Y (t), MGF of X +Y is:

(a) M X (t)+ M Y (t )
(b) M X (t)M Y (t)
(c) M X (t−1) M Y (t +1)
(d) M X (t)/ M Y (t )

Answer: (b) M X (t)M Y (t)

Steps / Explanation:
t ( X +Y ) tX tY
 MGF: M X +Y (t )=E [e ]=E [e ]E [e ]=M X (t )M Y (t ) if independent.

MCQ 19

In simple random sampling without replacement from population size N ,

finite population correction (fpc) is

this factor. True or false?

√ N −n . You multiply the usual SE by
N −1

(a) True
(b) False
Answer: (a) True
Steps / Explanation:
 For population proportion or mean, standard error for SRSWOR uses fpc to

reduce variability: SEw/o rep =

√ N−n
N −1
⋅SEwith rep . So true.

MCQ 20

If X 1 , X 2 , X 3are i.i.d. Exponential( λ ), distribution of min ⁡(X 1 , X 2 , X 3)is:

(a) Exponential with rate λ
(b) Exponential with rate 3 λ
(c) Gamma with shape 3
(d) Gamma with shape 1/3

Answer: (b) Exponential with rate 3 λ

Steps / Explanation:
 For i.i.d. exponentials, minimum's survival function: \Pr(\min > t)=\
Pr(X_1>t,X_2>t,X_3>t)=e^{-\lambda t}^3=e^{-3\lambda t}.

 So min ⁡is exponential with rate 3 λ .

🧮 1. Descriptive Statistics

Concept Formula Notes

∑ Xi
Mean X́ = Arithmetic average
n
Use formula for grouped data:
Middle value (if ordered N
Median −F
data) 2
l+ ×c
f

( f 1−f 0) f 1=modal freq, f 0 , f 2=adjacent

Mode l+ ×c
2 f 1−f 0−f 2 freqs
2
Variance s =∑ ¿ ¿ Sample variance

Standard Deviation s= √ s2 Measure of spread

Coefficient of s
CV = × 100 Comparison of variability
Variation X́

Skewness (Karl X́−Mode

Measure of asymmetry
Pearson) s
Concept Formula Notes

μ4
Kurtosis (β₂) β 2= Flatness of distribution
¿¿

🎲 2. Probability Theory

Concept Formula Description

Addition Rule P( A ∪ B)=P( A)+ P( B)−P( A ∩ B) For any 2 events

Conditional B) = \frac{P(A \cap B)}
( P(A
Probability {P(B)} )

Multiplication Rule ( P(A \cap B) = P(A B) P(B) )

Bayes’ Theorem ( P(A_i B) = \frac{P(B

Total Probability ( P(B) = \sum P(B A_i) P(A_i) )

Independent
P( A ∩ B)=P( A) P( B) Independence condition
Events

Mutually Exclusive P( A ∩ B)=0 Can’t occur together

🎯 3. Random Variables and Expectations

Concept Formula Description

Expected Value E( X )=∑ x i P( x i )or ∫ xf (x)dx Mean of RV

2
Variance Var ( X)=E(X )−¿ Spread measure

Covariance Cov (X , Y )=E(XY )−E (X )E (YLinear

) relation

Cov (X , Y )
Correlation Coefficient r= −1 ≤r ≤ 1
σ X σY

Moment Generating tX k th moment:

M X (t)=E(e ) (k )
Function M X (0)

🔢 4. Discrete Distributions

Mea
Distribution PMF / Formula Variance
n
x
Bernoulli(p) p ¿ p p(1−p)

( n) p ¿
x
Binomial(n,p) np np (1− p)
x
 Mea
Distribution PMF / Formula Variance
n
x
−λ λ
Poisson(λ) e λ λ
x!
2
Geometric(p) ¿ 1/ p (1− p)/ p

Negative ( x−1 ) p ¿
r
r / p r (1−p)/ p2
Binomial(r,p) r−1

📈 5. Continuous Distributions

Distribution PDF Mean Variance

Uniform(a,b) 1/( b−a) (a+ b)/2 ¿

− λx 2
Exponential(λ) λe 1/ λ 1/ λ
1 −¿ ¿
Normal(μ,σ²) e μ σ
2
σ √2 π
α
β α −1 −βx
Gamma(α,β) x e α/β α/β
2
Γ (α )

1 α −1 α αβ
Beta(α,β) x ¿
B (α , β) α+β ¿¿

1 k/ 2−1 − x/2
Chi-square(k) f (x)= k /2
x e k 2k
2 Γ (k /2)

t- Mean = Var = v /(v−2), for

distribution(v) 0 v>2

F- v2 Used for variance

distribution(v₁,
v 2−2 ratio
v₂)

🧠 6. Sampling & Estimation

Concept Formula Description

Sample Mean
E( X́ )=μ
(unbiased)

2 Unbiased for population

Sample Variance s =∑ ¿ ¿ variance

Standard Error σ
SE=
(mean) √n
Concept Formula Description

Confidence Interval σ
X́ ± z α / 2 For known σ
(μ) √n
s
t-based CI X́ ± t α /2 , n−1 For unknown σ
√n

⚖️7. Hypothesis Testing

Concept Test Statistic Decision Rule

Z-test for mean (σ X́−μ0

Z= Compare with zα
known) σ /√n

t-test for mean (σ X́−μ 0 Compare with tα,n-

t=
unknown) s/√n 1

2
s1 Compare with
F-test for variance F=
s
2
2
Fα(v₁,v₂)

2
Chi-square test for (n−1)s Compare with χ²
χ 2=
variance σ 20 α,n-1

Probability of obtaining value ≥ test

p-value Reject if p < α
statistic

📉 8. Regression & Correlation

Concept Formula Description

Line of best
Simple Linear Regression Y =a+bX + ε
fit

Cov ( X ,Y )
Slope (b) b=
Var (X )

Intercept (a) a=Ý −b X́

Coefficient of 2 S Sres Goodness of

R =1−
Determination (R²) S S tot fit

Multiple Regression Y = β0 + β 1 X 1 + β 2 X 2 +...+ ε

2
σ
Variance of Estimate Var (b)=
∑¿¿

9. Time Series Analysis

Concept Formula Description

Trend (Least
Y =a+bt Linear trend
Squares)

Y t−1 +...+Y t −n
Moving Average MA n= Smoothing
n
Actual Seasonal
Seasonal Index SI= × 100
Trend variation

∑(Y t −Ý )(Y t−k −Ý ) Serial

Autocorrelation rk =
∑¿ ¿ correlation

🧮 10. Linear Algebra (for DSIM Quantitative Aptitude)

Concept Formula Notes

Matrix Addition A+ B=[aij +bij ] Element-wise

Matrix Multiplication ¿
Determinant (2×2) ( A

−1 1 d −b
Inverse (2×2) A = [ ]
ad−bc −c a
Eigenvalues ( A - λI

Variance-Covariance Used in multivariate

Σ=E ¿
Matrix analysis

📊 11. Econometrics / Advanced Statistical Concepts

Concept Formula Description

Ordinary least
OLS Estimator β=¿
squares

2 Covariance of
Var(β̂) σ ¿ estimator

Detects
Durbin–Watson Statistic DW =∑ ¿ ¿
autocorrelation

VIF (Variance Inflation 1 Detects

VIF= 2
Factor) 1−Ri multicollinearity

Ridge Regression β=¿ Regularization

📐 12. Sampling Techniques

Concept Formula Notes

2
S n Finite pop
SRS (mean) Var ( X́)= (1− )
n N correction
2
Stratified 2 Sh
Var ( X́ s )=∑ W h
Sampling nh
2
Cluster Sc
Var ( X́ c )=
Sampling nc

🧾 13. Nonparametric Tests (Conceptual)

Test Description

Sign Test Median test for paired data

Wilcoxon Rank Sum Compare two samples

ANOVA equivalent for non-normal

Kruskal–Wallis Test
data
2
Spearman Rank 6∑ d
1−
Correlation 2
n(n −1)
📏 14. Estimation Theory (Point & Interval Estimation)

Concept Formula Description

Difference between
Bias ^
Bias ( θ)=E ^
( θ)−θ estimator mean and true
value

Mean Square Error ^ ^

MSE ( θ)=Var ( θ)+¿ Combines variance and bias
(MSE)

Efficiency of Var ( θ^ 2 ) Lower variance ⇒ more

Eff ( θ^ 1 , θ^ 2 )=
Estimator Var ( θ^ 1 ) efficient

Cramér–Rao Lower ^ 1
Var ( θ)≥ Minimum possible variance
Bound (CRLB) I (θ)
2
−∂ ln ⁡L Amount of information in
Fisher Information I (θ)=E [ ]
∂θ
2 sample

MLE (Maximum
L(θ)=∏ f (x i ; θ); maximize Estimate that maximizes
Likelihood
w.r.t. θ likelihood
Estimator)

Method of Moments Equate sample moments Simpler estimation

Concept Formula Description

with population moments technique

📊 15. ANOVA (Analysis of Variance)

Concept Formula Description

Total SS SST =∑ ¿ Total variation

Between- Variation between group

SSB=∑ ni ¿
group SS means

Within-group
SSW =∑ ∑ ¿ Variation within groups
SS

MSB SSB /(k−1)

F-statistic F= = Used for ANOVA test
MSW SSW /(N −k )

Decision Rule Reject H 0if F> Fα , k−1, N −k Test for equal means

🧩 16. Nonparametric Tests (Formulas)

Test Formula / Concept Use

S = \text{#(positive
Sign Test Statistic For median testing
differences)}

Wilcoxon Signed
T =∑ ranks of positive differencesPaired samples
Rank

n 1(n1+ 1) Compare two independent

Mann–Whitney U U =n1 n2+ −R1
2 samples
2
12 Ri
Kruskal–Wallis H H= ∑ −3( N +1) Nonparametric ANOVA
N (N +1) n i
2
Spearman Rank 6∑d
ρ=1− Rank correlation measure
Correlation 2
n(n −1)

📚 17. Bayesian Statistics

Descriptio
Concept Formula
n

x) = \
Posterior Probability ( P(\theta
frac{P(x
 Descriptio
Concept Formula
n

Prior × Likelihood =
Posterior ∝ Likelihood ×Prior
Posterior

Bayesian Estimator ( \hat{\theta}_{Bayes} =

x) )
(Posterior Mean) E(\theta

x) = E[\
Posterior Variance ( Var(\theta
theta^2

📉 18. Correlation and Regression Extensions

Concept Formula Description

r 12−r 13 r 23 Correlation between 1 & 2

Partial Correlation r 12.3=
√(1−r 2
13 )(1−r )
2
23
controlling 3

√
2 2
Multiple Correlation (1−r Y 1)(1−r Y 2 ) Strength of combined
(Y on X₁,X₂)
RY .12= 1− relation
(1−r 212 )

Elasticity in X́
E=b × Responsiveness of Y to X
Regression Ý
Standard Error of
Estimate
SE= √ ∑ ¿ ¿ ¿ Prediction accuracy

🧮 19. Matrix Algebra (Advanced Formulas)

Concept Formula Description

Trace of Matrix tr (A )=∑a ii Sum of diagonal elements

Determinant
( AB
Property
'
Orthogonal Matrix A A=I Columns are orthonormal

2 Common in projection
Idempotent Matrix A =A matrices

Projection Matrix Projects Y onto column space

P= X ¿
(OLS) of X

Hat Matrix 2 '

Property P =P , P =P Symmetric and idempotent

Residual Maker
M =I −P Projects onto residual space
Matrix
🧩 20. Multivariate Statistics

Concept Formula Description

Covariance Matrix Σ=E ¿ Dispersion structure

2
Mahalanobis Distance D =¿ Multivariate distance

Principal Component Eigen decomposition of Σ : Dimensionality

'
Analysis (PCA) Σ=PD P reduction

Eigenvalue Proportion λi
Var explained= For each component
(PCA) ∑ λi

' Linear separation of

Discriminant Function D=a X groups

📈 21. Time Series & Forecasting (Advanced)

Concept Formula Description

Exponential
F t+ 1=α Y t +(1−α )F t Forecasting
Smoothing

Double Exponential
Lt =α Y t +(1−α )(Lt −1 +T t−1 ) Accounts for trend
(Holt’s)

Autoregressive (AR) X t =φ1 X t−1 +φ2 X t −2+...+ ε t Past value model

Moving Average (MA) X t =ε t +θ1 ε t −1+ ...+ θq ε t −q Error term model

ARMA(p,q) Combines AR(p) and MA(q) Stationary series

ARIMA(p,d,q) Differenced ARMA Non-stationary data

ACF (Autocorrelation Cov (X t , X t−k )

ρk = Lag correlation
Function) Var ( X t )

Correlation controlling for Used for AR order

PACF (Partial ACF)
intermediate lags selection

💡 22. Information Theory & Entropy

Concept Formula Description

Measure of
Entropy H ( X)=−∑ p i log ⁡p i
uncertainty

Combined
Joint Entropy H ( X ,Y )=−∑ p( x , y )log ⁡p( x , y )
uncertainty
Concept Formula Description

Conditional Entropy ( H(Y X) = H(X,Y) - H(X) )

Mutual Information I (X ; Y )=H (X )+ H (Y )−H (X ,YShared

) information

Kullback–Leibler
( D_{KL}(P
Divergence

🧪 23. Chi-square & Goodness of Fit

Concept Formula Description

Chi-square Test 2 Observed vs

Statistic χ =∑ ¿ ¿ expected

Degrees of k −1(for k
Freedom categories)
2
Contingency Table χ =∑ ¿ ¿ Independence test

Expected (Ro w i × Col j)

Eij = For 2-way tables
Frequency Total

🧩 24. Order Statistics

Concept Formula Description

PDF of rth Order n! Distribution of

f (r ) (x)= ¿
Statistic (r−1)! (n−r )! ordered samples

Expected
1
Minimum E( X (1) )= For U(0,1)
n+1
(Uniform)

Expected
n
Maximum E( X (n ))= For U(0,1)
n+1
(Uniform)

🧮 25. Reliability & Survival Analysis

Concept Formula Description

Reliability Function R(t )=P(T >t)=1−F(t) Survival probability

f (t ) Instantaneous failure
Hazard Rate h(t)=
R(t) rate
Concept Formula Description
∞
Mean Time to Failure
E(T )=∫ ❑ R(t) dt Expected life
(MTTF) 0

− λt Constant hazard
Exponential Reliability R(t )=e model