Stat2001 Practice Exam
Stat2001 Practice Exam
(a) You have two dice which are identical except that one of them is fair and the other one is
loaded, with probability 50% of coming up 6 when rolled (rather than 1/6). You now randomly
choose one of the two dice (both equally likely), roll it once, and observe that 6 comes up.
Find the probability that the die which you have rolled is loaded. [10 marks]
(b) You have three dice which are identical except that two of them are fair and the other one is
loaded, with unknown probability p of coming up 6 when rolled (rather than 1/6). In your mind
there is initially a 50% chance that p = 1/5 and a 50% chance that p = 1/4. You now randomly
choose one of the three dice (all equally likely), roll it once, and observe that 6 comes up.
Find the probability that the die is loaded and also the probability that p = 1/4. [20 marks]
Find the minimum number of values that you should sample from the standard uniform distribution
if you want the average of these values to lie between 0.44 and 0.56 with a probability that is:
Find the expected value and variance of the number of times that the
sequence HT will occur, both as a function of n and if n = 40. [15 marks]
Page 1
Problem 4 [60 marks total]
Suppose that:
(Y | X = x ) ~ U (0, 4 x )
X ~ Beta (3,1) .
(a) Find the joint pdf of X and Y and evaluate this pdf at x = y = 0.2. [10 marks]
(b) Find the marginal pdf of Y and evaluate this pdf at y = 0.2. [10 marks]
(d) Find and sketch the density of R = XY . Then calculate the value
of the density at r = 0.25 and illustrate this value in your sketch. [20 marks]
Balls are drawn from the bag, randomly, repeatedly, one-by-one and with replacement,
until a black ball has been drawn.
We observe y, the number of times that a ball is drawn (including the last draw).
(c) Find the mean and bias of the maximum likelihood estimator of k.
Clearly express each of these quantities as a function of k. [10 marks]
Page 2
Problem 6 [30 marks total]
(a) Let y1 ,..., yn be a random sample from a distribution with finite mean and variance.
Suppose that we observe y1 + ... + yn =4.25 and y12 + ... + yn2 = 27.34, where n = 250.
Find an approximate central 95% confidence interval for the mean of the distribution.
[10 marks]
(b) We observe a single value y from the uniform distribution between θ and 2θ , where θ > 0.
Find an exact central 80% confidence interval for θ , both as a function of y and if y = 5.4.
[10 marks]
(c) We observe x, the maximum of the n values in a random sample from the uniform distribution
between 0 and c, where c > 0.
Find an exact lower range 100(1 − α )% confidence interval for c, both as a function of x, n and α ,
and in the case where x = 2.4, n = 7 and α = 0.3. [10 marks]
A random sample of size n is taken from the uniform distribution between 0 and 2c,
where n is odd and c is an unknown positive constant.
Consider as two estimators of c the sample mean, Y , and the sample median, M.
(a) Find the variance of Y , both generally and in the case where c = 6 and n = 9. [10 marks]
(b) Find the efficiency of Y relative to M, both generally and in the limit as n → ∞ .
[20 marks]
Note: If the sample values are Y1 ,..., Yn , then the sample mean is Y = (Y1 + ... + Yn ) / n , and the
sample median is M = Y(( n +1)/2) , i.e. the (( n + 1) / 2)th order statistic, e.g. M = Y(7) if n = 13.
Page 3
Problem 8 [40 marks total]
(Y | θ ) ~ Beta (θ + 1, θ + 1)
θ ~ Bernoulli (1/ 2) .
(b) Find the posterior distribution and pmf of θ given the data y.
Then calculate the value of the posterior pmf at θ = 0 if y = 0.5. [10 marks]
(c) Find the posterior mean and mode of θ , each as a function of y. Draw these two functions in a
single sketch, using a solid line for the posterior mean and a dashed line for the posterior mode.
Then calculate and show in your sketch the values of the two functions at y = 0.1. [20 marks]
It has been alleged that p, the proportion of people with a particular disease, is 8%. We wish to test
this allegation against the counter-assertion that fewer than 8% of people have the disease. We have
decided to sample people from the population (of several hundred million people), randomly and
one-by-one, and test each person, until we have found a person with the disease. As a test statistic
we will use Y, the number of persons sampled (including the last person, who has the disease).
We want the test to have a significance level which is as close to 0.03 as possible but no larger.
Formulate a suitable hypothesis test. Clearly write down the two hypotheses involved, the exact
distribution of the test statistic, and the rejection region.
Next, find the probability of a Type II error, both as a function of p and if p = 0.04.
Then sketch this function, with a dot showing its value at p = 0.04.
Finally, find the p-value for the test, both generally as a function of y and if 37 people are sampled.
In the case of 37 people being sampled, conclude whether the null hypothesis should be rejected.
Page 4
Problem 10 [30 marks total] [to be attempted only by STAT6039 and STAT6013 students]
A random sample of size n = 50 from the beta distribution with parameters λ and 1
(and mean λ / (λ + 1) ) yields the following summary statistics:
n n n
∑ yi = 39.1251,
i =1
∑ yi2 = 31.9252,
i =1
∑ log y = −13.5177 .
i =1
i
(b) Use results in (a) and the large-sample properties of maximum likelihood estimators
to find an approximate central 95% confidence interval for λ . [20 marks]
END OF EXAMINATION
Page 5