DEPARTMENT OF MATHEMATICS

QUESTION BANK (UNIT 3 AND UNIT 4)

CLASS: II M.SC
SUBJECT : MATHEMATICAL STATISTICS
--------------------------------------------------------------------------------------
ONE MARK:
1. If U is the most estimate of the parameter Q then almost
everywhere in the domain where g(u,Q) > 0 , U is the solution
𝜕𝑙𝑜𝑔𝐿
of equation = __________
𝜕𝑄

a)0
b)1
c)Q
d) None of the above
2. Express e = D2(U)/D2(U1) is taken if e = 1 the U1 is said to be
_______
a) Unbiased
b) Biased
c) Most efficient
d) Asymptotically most efficient
3. An unbiased estimate U of the parameter Q is the most efficient
estimate iff the estimate U is ________
a) Not sufficient
b) Sufficient
c) Necessary
d) Not neccessay
4. If f=gh then 𝜃 is ________
a)Unbiased
b)Efficient
c)Sufficient
d)Consistent
5.Rao-cramer inequality with respect to the variance of an estimator
gives____
a)Lower bound of variance
b)Upper bound of variance
c)Asymptotic variance
d)None
6.In case of ___ variables , Rao-cramer inequality if valid.
a)Continuous
b)Discrete
c)Both a and b
d)None
7.When the variance of an estimator obtains its Cramer Rao lower
bound for variance then this estimator will be
a)Unbiased
b)Efficient
c)Sufficient
d)All the above
8.A value of an estimator is called:
(a) Estimation
(b) Estimate
(c) Variable
(d) Constant
9.Estimation is of two types:
(a) One sided and two sided
(b) Type I and type II
(c) Point estimation and interval estimation
(d) Biased and unbiased
10.Estimation is possible only in case of a:
(a) Parameter
(b) Sample
(c) Random sample
(d) Population
11.Estimate and estimator are:
(a) Same
(b) Different
(c) Maximum
(d) Minimum
12.The process of using sample data to estimate the values of
unknown population parameter is called:
(a) Estimate
(b) Estimator
(c) Estimation
(d) Interval estimation
13.1The process of making estimates about the population
parameter from a sample is called:
(a) Statistical independence
(b) Statistical inference
(c) Statistical hypothesis
(d) Statistical decision
14.Statistical inference has two branches namely:
(a) Level of confidence and degrees of freedom
(b) Biased estimator and unbiased estimator
(c) Point estimator and unbiased estimator
(d) Estimation of parameter and testing of hypothesis
15.A specific value calculated from sample is called:
(a) Estimator
(b) Estimate
(c) Estimation
(d) Bias
16.Interval estimate is determined in terms of:
(a) Sampling error
(b) Error of estimation
(c) Confidence coefficient
(d) Degrees of freedom
17.The level of confidence is denoted by:
(a) α
(b) β
(c) 1 – α
(d) 1 – β
18.A single value used to estimate a population values is called:
(a) Interval estimate
(b) Point estimate
(c) Level of confidence
(d) Degrees of freedom
19.An interval calculated from the sample data and it is likely to
contain the value of parameter with some probability is called:
(a) Interval estimate
(b) Point estimate
(c) Confidence interval
(d) Level of confidence
UNIT 4
1.Hypothesis H0(Q=Q0), H1(Q=Q1), H2(Q=Q2),… are called-------
a)admissible hypothesis
b)null hypothesis
c)alternate hypothesis
d) statistical hypothesis
2. If P( ∈ 𝑤)/ Q=Q0)=𝛼, then w is called
a) level of significance
b) error of first kind
c) error of second kind
d) critical region
3. Type 1 error occurs when?
a) We reject H0 if it is True
b) We reject H0 if it is False
c) We accept H0 if it is True
d) We accept H0 if it is False
4. The probability of Type 1 error is referred as?
a) 1-α
b) β
c) α
d) 1-β
5. Rejecting H0 when it is true is called__________
a) error of first kind
b)error of second kind
c)both a and b
d) none of the above
6) Accepting H0 when it is false is called__________
a) error of first kind
b)error of second kind
c)both a and b
d) none of the above
7)_____________ expresses the probability of accepting the
hypothesis H0(Q= Q0)
When applying the test w as a function of Q
a)Power function
b)OC function
c)Statistical hypothesis
d)Alternative hypothesis
8)If 𝜔 contains all the Q≠Q0 such a test is called____________
a) most powerful test
b)Uniformly most powerful test
c)Both a and b
d)None of the above

&&&&&&&&&&&&

FIVE MARKS
Unit 3
1. The random variable X has the Poisson distribution 𝑃(𝑋 =
𝜆𝑘
𝑘) = 𝑒 −𝜆 , (k= 0,1,2,…) where the parameter  is unknown.
𝑘!
Considering n independent observations of X, estimate  by the
method of maximum likelihood.

2. Consider the statistic U = 𝑋̅ is a sufficient estimate of the

parameter ‘m’ of the normal distribution N(m; ) where  is
known show that U = 𝑋̅ estimate is also most efficient.
3. The characteristic X of the element of a population has the
normal distribution with the density 𝑓(𝑥, 𝜎 2 ) =
−1 𝑥−𝑚 2
1 ( 𝜎 )
𝑒 2 where m is known and  is unknown. Estimate
𝜎√2𝜋
2
𝜎 from a simple sample of n elements and choosing the
1
estimate of 𝜎 2 the statistic 𝑆02 = ∑𝑛𝑘=1(𝑋𝑘 − 𝑚)2 . Show that
𝑛
the estimate is most efficient
4. The Characteristic X of elements of a population has the
𝜆𝑘
Poisson distribution 𝑃(𝑋 = 𝑘) = 𝑒 −𝜆 , (k= 0,1,2,…) where
𝑘!
 the parameter  is unknown.Find the minimal variance of a
regular and unbiased estimate U of .
5. The characteristic X of elements of a population has the normal
distribution N(m; ), with unknown  and m known. Consider
1
the statistic 𝑈 = √0.5𝜋 ∑𝑛𝑘=1|𝑋𝑘 − 𝑚| , where the random
𝑛
variables Xk are independent and have the same distribution as
X. Prove that U is an unbiased estimate of  . Also find the
efficiency of U.
6. The characteristic X of elements of a population has the normal
distribution N(m; ) where the expected value with m is known
and standard deviation  is not known. As we know, the statistic
S02 unbiased estimate of the parameter 2. Show that S0 is not
an unbiased estimate of .
7. Consider the class D of distribution functions of random
variables whose second moment exist. The variance 2 of X is
1 ̅̅̅̅2 , where the random
unknown and let 𝑆 2 = ∑𝑛 𝑋 2 − 𝑋
𝑘=1 𝑘
𝑛
variables Xk are independent and have the same distribution as
X . Prove that S2 is not an unbiased estimate of the parameter 2
8. The characteristic X of elements of a population has the normal
distribution N(m; ) with m unknown. As an estimate of m,
take the median of simple samples of size n. Find the
asymptotic efficiency of this estimate.
9. Find the maximum likelihood estimate for the parameter 𝜆 of
the poisson distribution.
Unit 4
1. Explain one way classification.
2. The expected value m of a population, where the characteristic
x has the normal distribution N(m; 1) , is unknown. Test the
hypothesis H0 (m = m0) against the alternate hypothesis H1 (m =
m1) , where m1  m0 , by applying a most powerful test based
 on a simple sample for a given  and  . What value should n
have?
3. Find a most powerful test for testing the hypothesis H0 (m = m0)
against the alternate hypothesis H1 (m = m1) where m is the
unknown expected value of m of a normal population N(m,1)
using simple sample.

TEN MARKS
Unit 3
1. State and prove Rao-Cramer inequality.
2. State and Blackwell theorem.
3. Prove that an unbiased estimate U of the parameter 𝜃 is the
most efficient if and only if
(a) the estimate U is sufficient
(b) for g(u, 𝜃)  0 , the density g(u, 𝜃 ) almost everywhere
𝜕𝑙𝑜𝑔𝑔(𝑢,𝜃)
satisfies the relation = c(u-𝜃), where the number c is
𝜕𝜃
independent of u but may depend on 𝜃.
4. If U is the most efficient estimate of the parameter Q, then
almost everywhere in the domain f(u,Q)>0, u is the solution of
𝜕𝑙𝑜𝑔𝐿
equation = 0.
𝜕𝑄
5. Let the distribution function F(x) depend upon one parameter Q
that is m=1. If there exists a sufficient estimate U of the
𝜕𝑙𝑜𝑔𝐿
parameter Q then the solution of the equation = 0 is a
𝜕𝑄
function of U only.
Unit 4
1. State and prove Fisher’s lemma.
2. Explain uniformly most powerful test.
 3. State and prove Fundamental Lemma of Neyman and Pearson
(Particular Case).
4. Let the point  = (x1 ,x2 ,….,xn) in sample space be a random
variable of continuous type with density f( ,Q), where the
parameter Q is unknown and its admissible values belong to a
certain interval K (finite or infinite). Let H0 (Q = Q0) and H1 (Q =
Q1) be the null and alternative hypothesis, respectively and let
the point Q0 belong to the interior of the interval K. If the n-
dimensional density f( ,Q ) has at each interior point Q of the
interval K the partial derivative
𝜕𝑓(, Q)
𝑓1 (, Q) =
𝜕𝑄
Which satisfies the inequality |𝑓1 (, Q)| < 𝑔(), where the n-
∞ ∞
dimensional integral ∫−∞ … … . . ∫−∞ 𝑔(𝑥1 , 𝑥2 , … . 𝑥𝑛 ) 𝑑𝑥1 … . 𝑑𝑥𝑛
is finite then among the unbiased tests there exists a most powerful
test.
********************************************************