Probability exercises XIV

1. A bottling machine can be regulated so that it discharges an average of µ ounces per bottle.
It has been observed that the amount of fill dispensed by the bottling machine is normal
with σ = 1 ounce. If n = 16 bottles are randomly selected from the machine and the sample
mean Ȳ is computed, then what is P (|Y − µ|) ≤ 0.3.
2. An environment protection agency is concerned with the problem of setting criterion for the
amounts of certain toxic chemicals to be allowed in freshwater lakes and rivers. A common
measure of toxicity of any pollutant is the concentration of the pollutant that will kill half
of the test species in a given amount of time (usually 96 hours for fish species). This
measure is called LC 50 (lethal concentration killing 50% of the test species). In many
studies the values contained in the natural logrithm of LC50 measurements are normally
distributed and hence the analysis is based on ln(LC50) data.
Studies of copper on a certain species of fish (say, species A), show the variance of ln(LC50)
to be around 0.4 with concentration measurements in milligrams per litre. If n = 10 studies
on LC50 of copper are to be completed find the probability that the sample mean of
ln(LC50) will differ from the true population mean by no more than 0.5.

3. Suppose X1 , X2 , . . . , Xm and Y1 , Y2 , . . . , Yn are independent normal samples with the

variables Xi normally distributed with mean µ1 and variance σ12 and the variables Yi
normally distributed with mean µ2 and variance σ22 . Find the E(X − Y ), Var(X − Y ).
Suppose that σ12 = 2, and σ22 = 2.5. Find the sample sizes so that (X − Y ) will be within
one unit of µ1 − µ2 .
4. Suppose Z1 , Z2 , . . . , Z6 denote a random
! sample from a standard normal distribution. Find
X6
a number b such that P Zi2 ≤ b = 0.95.
i=1

5. In problem 2 suppose that n = 20 observations are taken on ln(LC50) measurements and

that σ 2 = 1.4. Let S 2 denote the sample variance of 20 measurements.
(i) Find a number b such that P (S 2 ≤ b) = 0.975.
(ii) Find a number a such that P (S 2 ≥ a) = 0.975.
(iii) If a and b are as in previous parts what is P (a ≤ S 2 ≤ b).
6. If Y is a random variable that has F distribution with ν1 numerator and ν2 denominator
degrees of freedom, show that U = 1/Y has F distribution with ν2 numerator and ν1
denominator degrees of freedom.

7. Suppose Y1 , Y2 , . . . , Y5 be a random sample of size 5 from a normal population with mean 0

5
1X
and variance 1, and Y = Yi . Let Y6 be another independent observation from the
5 i=1
same population.
5
X
(a) What is the distribution of W = Yi2 . Why?
i=1
5
X
2
(b) What is the distribution of U = Yi − Y . Why?
i=1
5
X
2
(c) What is the distribution of V = Yi − Y + Y62 . Why?
i=1

8. Suppose Y1 , Y2 , . . . , Yn , Y , W, U are as defined in problem 5.

√
5Y6
(a) What is the distribution of √ .
W
2Y6
(b) What is the distribution of √ .
U
2
2(5Y + Y62 )
(c) What is the distribution of .
U
 Probability exercises XV
1. Workers employed in a large service industry have an average wage of 7.00 per hour with a
standard deviation of 0.50. The industry has 64 workers of a certain ethinic group. These
workers have an average wage of 6.90 per hour. Is it reasonable to assume that the wage
rate of the ethinic group is equivalent to that of a random sample of workers from those
employed in the service industry.
2. The length of time required for the periodic maintainance of an automobile or other
machine usually has a mound shaped probability distribution. Because some occasional
long service times will occur, the distribution tends to be skewed to the right. Suppose the
length of time required to run a 5000-mile check and to service a new automobile has mean
1.4 hours and standard deviation 0.7 hours. Suppose also that the service department plans
to service 50 automobiles per 8-hour day and that in order to do so it can spend a
maximum average service time of only 1.6 hours per automobile. On what proportion of
the workdays will the service department have to work overtime.
3. Suppose X1 , X2 , . . . , Xn and Y1 , Y2 , . . . , Yn are independent random samples from
populations with means µ1 and µ2 and variances σ12 and σ22 respectively. Show that the
random variable
(X − Y ) − (µ1 − µ2 )
Un = q
2 2
σ1 +σ2
n

satisfies the conditions of Central Limit Theorem and thus the distribution function of Un
converges to the standard normal distribution function as n → ∞.
4. The result in problem 3 holds even if sample sizes differ. That is if X1 , X2 , . . . , Xn and
Y1 , Y2 , . . . , Yn are independent random samples from populations with means µ1 and µ2 and
variances σ12 and σ22 respectively, then (X − Y ) will be approximately normally distributed
σ2 σ2
with means µ1 and µ2 and variance 1 + 2 . The flow of water through soil depends on,
n1 n2
among other things the porosity (volume proportion of voids) of the soil. To compare two
types of sandy soil, n1 = 50 measurements are taken on the porosity of soil A and n2 = 100
measurements are taken on the porosity of soil B. Assume that σ12 = 0.01 and σ22 = 0.02.
Find the probability that the difference between sample means will be within 0.05 unit of
the difference between population means (µ1 − µ2 ).
 Probability exercises XVI
1. Using the identity Θ̂ − Θ = [Θ̂ − E(Θ̂)] + [E(Θ̂) − Θ] = [Θ̂ − E(Θ̂)] + B(Θ̂), prove that
M SE(Θ̂) = E[Θ̂ − E(Θ̂)]2 − [Θ − E(Θ̂)]2 = V ar(Θ̂) − [B(Θ̂)]2
2. Suppose that E(Θ̂1 ) = E(Θ̂2 ) = Θ, V ar(Θ̂1 ) = σ12 , V ar(Θ̂2 ) = σ22 .
Consider the estimator Θ̂3 = aΘ̂1 + (1 − a)Θ̂2 .
(i) Show that Θ̂3 is an unbiased estimator of Θ.
(ii) If Θ̂1 and Θ̂2 are independent, how should the constant a be chosen in order to
minimize the variance of Θ̂3 .
3. Suppose Y1 , Y2 , Y3 denote a random sample from an exponential distribution with
probability density function given by
 
 1 e− yθ for y > 0
f (y) = θ
0, otherwise


Y1 +Y2 Y1 +2Y2
Consider tthe five estimators of θ: Θ̂1 = Y1 , Θ̂2 = 2 , Θ̂3 = 3 ,
Θ̂4 = min{Y1 , Y2 , Y3 }, Θ̂5 = Y .
(i) Which of these estimators are unbiased.
(ii) Among the unbiased estimators which has the smallest variance.
4. Let Y v B(n, p).
Y
(i) Prove that n is an unbiased estimator of p.
(ii) To estimate the variance of Y we generally use n( Yn )(1 − Y
n ). Show that this is a biased
estimator of Var(Y).
(iii) Modify n( Yn )(1 − Y
n) slightly to form an unbiased estimator of Var(Y).
5. Let Y1 , Y2 , . . . , Yn denote a random sample of size n drawn from a population with density
given by  α−1
 αy
, 0 ≤ y ≤ θ,
f (y) = θα
0, otherwise
where α > 0 is a known fixed value but θ is unknown. This is called the power family
distribution. Consider the estimator θ̂ = max{Y1 , Y2 , . . . , Yn }
(i) Show that θ̂ is a biased estimator of θ.
(ii) Find a multiple of θ̂ that is a unbiased estimator of θ.
(iii) Derive MSE(θ̂).

6. A large study was undertaken on the possible effects of trace elements in drinking water on
kidney stone disease. The accompanying table presents data on age, amount of calcium in
home drinking water (measured in parts per millions), and smoking activity. These data
were obtained with indiviguals with recurrant kidney stone problems, all of whom lived in
state A or state B.

State A State B
Sample size 467 191
Mean age 45.1 46.4
Standard deviation of age 10.2 9.8
Mean calcium component (ppm) 11.3 40.1
Standard deviation of calcium 16.6 28.4
Proportion now smoking 0.78 0.61

(a) Estimate the average calcium content in drinking water for kidney-stone patients in
state A. Place a bound on error of estimation.
 (b) Estimate the difference in mean ages for kidney-stone patients in State A and in State
B. Place a bound on error of estimation.
(c) Estimate and place a two-standard-deviation bound on the difference in proportions of
kidney-stone patients from State A and State B who were smokers at the time of study.
7. A survey conducted in 1922 on 1251 indiviguals and in 1983 on 1250 indiviguals on health
habits of the population of a certain country yeilded the following data:

Health Issue 1983 survey 1992 survey

Ate recommended amount of fibrous food 0.59 0.53
Avoided fat 0.55 0.51
Avoided excess salt 0.53 0.46
Used seat belts 0.19 0.70
Used smoke detectors 0.67 0.90

(a) Give a point estimate the difference in proportions of population who ate the
recommended amount of fibrous foods in 1983 and 1992. Provide a bound on the error
of estimation.
(b) Based on your results from (a) do you think that there has been a demonstrable
decrease in the proportion of population who ate the recommended amount of fibrous
foods in 1983 and 1992.
8. We can place a two-standard-deviation bound on the error of estimation with any estimator
for which we can find a reasonable estimate of the standard error. Suppose that
Y1 , Y2 , . . . , Yn denote a random sample from a Poisson distribution with mean λ. We know
that Var(Yi ) = λ and hence E(Ȳ ) = λ, Var(Yi ) = nλ . How will you employ Y1 , Y2 , . . . , Yn to
estimate λ. How would you estimate the standard error of your estimator.
 Probability exercises XVI
1. In exercise XVI problem 3 find the efficiency of Θ̂1 relative to Θ̂5 , Θ̂2 relative to Θ̂5 and Θ̂3
relative to Θ̂5 .
2. Let Y1 , Y2 , . . . , Yn denote a random sample of size n from a population with µ and σ 2 .
Consider the following three estimators for µ:
Y1 + Y2 Y1 Y2 + Y3 + · · · + Yn−1 Yn
(i) µ̂1 = , µ̂2 = + + , µ̂3 = Y .
2 4 2(n − 2) n
(ii) Show that each of the three estimators are unbiased.
(iii) Find the efficiency of µ̂3 relative to µ̂2 and µ̂3 relative to µ̂1 .
3. Let Y1 , Y2 , . . . , Yn denote a random sample of size n from an exponential distribution with
probability density function given by
 
 1 e− yθ for y > 0
f (y) = θ
0 otherwise
