Mains PYQs
UPSC Mains Statistics paper 1 2018 questions
- We have a bag with 3 coins in it. One of them is a fair coin, but the others are biased trick coins. When flipped the three coins come up 'heads' with probability $0.5$, $0.6$, $0.1$ respectively. If one of these coins is picked at random and flipped three times,
(i) What is $P(HTT)$ ?
(ii) Assuming that $\{HTT\}$ occurred, what is the probability of having chosen a fair coin ?
Here $\{H, T\}$ denote Heads and Tails respectively.
- Let $X_1, X_2, \ldots, X_n$ be a sample from a distribution with density function
$$f(x\mid\theta) = \theta(\theta+1)x^{\theta-1}(1-x), \quad 0<x<1,\ \theta>0.$$
Obtain estimate of $\theta$ by method of moments.
- Consider two independent and identically distributed uniform $U(0,1)$ random variables. Derive the distribution of $U = \frac{X}{Y}$. Comment on $E(U)$.
- The following data represents lifetime (hours) of batteries for two different brands :
| Brand A | 40 | 30 | 40 | 45 | 55 | 30 |
|---|---|---|---|---|---|---|
| Brand B | 50 | 50 | 45 | 55 | 60 | 40 |
Use the Median test to conclude if the two samples come from the same population.
- Using central limit theorem, show that
$$\lim_{n\to\infty} e^{-n}\sum_{k=0}^{n}\frac{n^k}{k!} = \frac{1}{2}.$$
- The joint density of $X, Y$ is
$$f(x,y) = e^{-y}, \quad 0\le x\le y<\infty.$$
Find the marginal distributions of $X, Y$ and the conditional distribution of $Y$ given $X$.
- Let $X_1, X_2, \ldots, X_n$ be iid binomial $(1, p)$ random variables. Find the UMVUE for $p(1-p)$, based on the sample.
- Let $x_1, x_2, \ldots, x_n$ be a random sample from $N(0, \theta)$. The variance $\theta$ is unknown. Derive the most powerful test for $H_0 : \theta = \theta_0$ against $H_1 : \theta > \theta_0$, where $\theta_0$ is a fixed positive number. Is it a UMP test ?
- Shots aimed at a target reach the plane of the target at a distance $R$ from the centre, where $R$ has the probability density function
$$f_R(r;\sigma^2) = \frac{1}{\sigma^2}\, r\, e^{-r^2/2\sigma^2};\ r\ge 0.$$
A hit is scored if $R\le a$. If $n$ shots are fired and there are exactly $m$ hits, these being at distances $r_1, r_2, \ldots, r_m$ from the centre, prove that the maximum likelihood estimator of $\sigma^2$ is given by
$$\hat{\sigma}^2 = \frac{1}{2m}\sum_{i=1}^{m} r_i^2 + \frac{n-m}{2m}a^2.$$
- Let $\{X_n\}$ be a sequence of random variables and $Y_n = \frac{1}{n}\sum_{i=1}^{n} X_i$.
Prove that a necessary and sufficient condition for the sequence $\{X_n\}$ to satisfy the weak law of large numbers is that
$$E\left[\frac{Y_n^2}{1+Y_n^2}\right] \to 0 \text{ as } n\to\infty.$$
- (i) Let $S_n$ be equal to the total obtained in $n$ tosses of a symmetric die. Estimate an upper bound for
$$P\left[\left|\frac{S_n}{n} - 3.5\right| > \varepsilon\right], \text{ for } \varepsilon>0.$$
(ii) Let $X_n$ have the density function, for $n=1,2,\ldots$,
$P[X_n = 1] = \frac{1}{n}$ and $P[X_n = 0] = 1 - \frac{1}{n}$.
Check whether $X_n$ converges to zero in probability.
- If $X_n \xrightarrow{L} X$, $Y_n \xrightarrow{P} C$, then show that
(i) $X_n + Y_n \xrightarrow{L} X + C$
(ii) $X_n \cdot Y_n \xrightarrow{L} CX$
(iii) $X_n / Y_n \xrightarrow{L} X/C$, provided $C \ne 0$.
- Let $X \sim f_j$, $j = 0, 1$, where
| x | 1 | 2 | 3 | 4 | 5 |
|---|---|---|---|---|---|
| $f_0$ | $\frac15$ | $\frac15$ | $\frac15$ | $\frac15$ | $\frac15$ |
| $f_1$ | $\frac16$ | $\frac14$ | $\frac16$ | $\frac14$ | $\frac16$ |
Find the best critical region of size $\frac{2}{5}$. What is its power ?
- (i) For testing a simple hypothesis $H_0 : \theta = \theta_0$ against simple alternative $H_1 : \theta = \theta_1$, prove that the Sequential Probability Ratio Test (SPRT) terminates with probability one.
(ii) Let $X_1, X_2, \ldots$ be random observations from the Bernoulli distribution with $P(X=1) = p$ and $P(X=0) = 1-p$. Obtain a sequential probability ratio test for $H_0 : p = p_0$ against $H_1 : p = p_1$.
- The period of oscillation, $t$ of a pendulum is $2\pi\sqrt{l/g}$ where $l$ is the length of the pendulum and $g$ is the gravitational constant. The periods observed are $t_{ij}$ ($j = 1, 2, \ldots, n_i$) and lengths $l_i$ ($i = 1, \ldots, k$) of the pendulum, in an experiment. Assuming the errors of observations to be uncorrelated with zero means and variances $\sigma^2$, show that the best linear estimator of $\beta = 2\pi/\sqrt{g}$ is
$$\hat\beta = \frac{\sum_i (l_i)^{\frac12}\sum_j t_{ij}}{\sum_i n_i l_i} \text{ and } V(\hat\beta) = \frac{\sigma^2}{\sum_i n_i l_i}.$$
- Define principal components. Is the principal component affected by standardization ? Establish your claim by illustrating the same with the covariance matrix
$$\Sigma = \begin{bmatrix} 1 & 4 \\ 4 & 100 \end{bmatrix}.$$
- (i) Distinguish between total and partial confounding in a $2^k$ factorial experiment.
(ii) If a $2^3$ factorial experiment is laid in a block of size 4 with four replicates, write down the allocations of degrees of freedom for analyzing the results of such a design when the highest order interaction is totally confounded.
- A population has seven units consisting of 30, 10, 25, 5, 20, 6 and 4 fields. Select a PPS sample of units of size 3, using random numbers 72, 46 and 94.
- In a randomized block design with $t$ treatments and $r$ replicates, one observation was found missing. Explain the method of estimating the missing value and perform the analysis of variance.
- What do you understand by Balanced Design ? Suppose four treatments (1, 2, 3, 4) are laid out in five blocks (I, II, III, IV, V) as given below :
| Block | Treatments |
|---|---|
| I | 1 1 2 |
| II | 1 2 2 |
| III | 3 4 |
| IV | 3 4 |
| V | 3 4 |
Establish that the above design is unbalanced.
- If $X$ is a $p$-component vector of random variables distributed as $N_p(0, \Sigma)$, then show that $X'\Sigma^{-1}X$ is distributed as chi-square with $p$ degrees of freedom.
- Compare ratio and regression methods of estimation. Show that ratio method gives a biased estimation of population total. Derive an approximate expression of the bias.
- For a random sample of size $n$ from $N_4(\mu, \Sigma)$, we obtain $\hat\mu = (18, 15, 18, 14)$ and
$$\hat\Sigma = \begin{bmatrix} 9.5 & & & \\ 5.2 & 5.4 & & \\ 6.9 & 5.1 & 10.0 & \\ 4.6 & 3.5 & 5.6 & 4.5 \end{bmatrix}.$$
(i) Find the estimates of the parameters of the conditional distribution of $(X_3, X_4)$ given $(X_1, X_2)$.
(ii) Find the partial correlation $r_{34\cdot12}$.
- A $2^5$ experiment involving factors A, B, C, D and E is conducted in the blocks of size 8. The key blocks in the replicates 1 and 2 are respectively given by
1, bd, abde, cde, acd, bce, abc, ae and
1, ac, abd, ade, bcd, cde, be, abce.
Identify the confounded effects in each replication.
- Two dentists A and B make a survey of the state of the teeth of 200 children in a village. Dr. A selects a simple random sample of 20 children and counts the number of decayed teeth for each child with the following results :
| No. of decayed teeth | 0 | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 |
|---|---|---|---|---|---|---|---|---|---|---|---|
| No. of children | 8 | 4 | 2 | 2 | 1 | 1 | 0 | 0 | 0 | 1 | 1 |
Dr. B using the same dental technique, examines all 200 children, recording merely those who have no decayed teeth. He finds 60 children with no decayed teeth.
Estimate the total number of decayed teeth using
(i) Dr. A's results only
(ii) Both Dr. A's and Dr. B's results
Which estimate do you prefer and why ?
- Explain stratified random sampling and give reasons for its preference to simple random sampling.
With two strata, a sampler would like to have $n_1 = n_2$ for administrative convenience instead of using the values given by the Neyman allocation. If $V(\bar y_{st})$ and $V_{opt}(\bar y_{st})$ denote the variances given by the $n_1 = n_2$ and Neyman allocations respectively, show that the fractional increase in the variance is
$$\frac{V(\bar y_{st}) - V_{opt}(\bar y_{st})}{V_{opt}(\bar y_{st})} = \left(\frac{r-1}{r+1}\right)^2$$
where $r = n_1/n_2$ as given by Neyman allocation.
- Show that Hotelling's $T^2$ is a generalization of the Student's $t$. Prove the invariance property of $T^2$ under the non-singular linear transformations of the variables. Briefly mention different uses of $T^2$.
- For a model
$$Y_i = \beta_0 + \beta_1 X_i + \varepsilon_i, \quad i = 1, 2, 3$$
where $X_1 = -1$, $X_2 = 0$, $X_3 = 1$, find BLUE's of $\beta_1, \beta_0$. Suppose the model is not correct and the true model is
$$Y_i = \beta_0 + \beta_1 X_i + \beta_2 X_i^2 + \varepsilon_i.$$
Find the bias of the BLUE's obtained.