Math 501 Assignment 6: due Tuesday 11/12 at 11:59 p.m.

submitted
via Canvas in pdf format

Scoring: This assignment is worth 180 points (10 + 10 + 25 + 10 + 30 + 30

+ 65 = 180).

1 Academic integrity and collaboration policy

for this assignment
You may work together on the reading assignment, which includes dozens of
solved problems, examples, proofs, detailed discussions, theorems, etc., and you
are encouraged to do so, but you must complete the written assignment entirely
on your own. Learn the material well with the reading assignment; prove you
have learned the material well with the written assignment.
You may consult our TAs or me regarding questions about written exercises
if you have worked hard on them and are very stuck, but you may not consult
anyone else until after all parties have submitted the assignment.
You are to …gure the exercises out entirely on your own.
If you …nd solutions to one or more of the written exercises in Weiss’s book,
in Hsu’s book, or in documents I myself have previously posted to Canvas for
our course this semester, you may make use of them, but the write-up should
be your own, and you must cite the source.
If you …nd solutions to one or more of the written exercises in any source
other than Weiss’s book, Hsu’s book, or documents I myself have previously
posted to Canvas for our course this semester, you may not make use of them
at all, and doing so would be an academic integrity violation.

2 Reading assignment
Class notes: Read the class notes up to and including Chapter 37 on
random vectors (Chapters 1 through 36 were previously assigned.).
Weiss textbook: Sections 7.1, 7.2, 7.3, 10.1, 10.2, 6.1, 6.2, 6.3, 6.4, 9.1,
9.2, 9.3 (Chapters 1 through 5, and Sections 8.1, 8.2, 8.3, 8.4, and 8.5 were
previously assigned.)
Schaum’s book (Hwei Hsu): Skip the “notes” (use our notes and Weiss’s
textbook). However, carefully work through exercises 2.27 – 2.36 (all),
2.44 – 2.50 (all), 2.57, 2.58. Do not turn these in, but make sure you
understand the details of the calculations and how to solve these exercises.

1
3 Written Assignment (to be turned in)
3.1 Exercise 1 (general random variables and cdfs) (10
points –5 points per part)
3.1.1 Background
Read Proposition 8.2 from Weiss’s book carefully. In class, we derived this
result. In this exercise, you will make use of it. Here is an example. Study it
carefully, and proceed similarly for the assigned problem below. You will have
to make some adjustments, since the cdf for the assigned problem is di¤erent.

Example 1 Solve Exercise 8.25a from the textbook. Also, solve Exercise 8.25e
from the textbook but only for the case a = 1; b = 2 (as in part a).

Solution: We calculate

P (f1 < W 2g) = P (fW 2g) P (fW 1g)

= FW (2) FW (1)
= 0:875 0:5 = 0:375:

P (f1 < W < 2g) = P (fW < 2g) P (fW 1g)

= FW (2 ) FW (1)

= lim FW (w) FW (1)

w!2
= 0:5 0:5 = 0:

P (f1 W < 2g) = P (fW < 2g) P (fW < 1g)

= FW (2 ) FW (1 )

= lim FW (w) lim FW (w)

w!2 w!1
= 0:5 0:125 = 0:375:

P (1 W 2) = P (fW 2g) P (fW < 1g)

= FW (2) FW (1 )

= FW (2) lim FW (w)

w!1
= 0:875 0:125 = 0:75:

Next, we’ll …nd these probabilities by using the pmf and the Fundamental
Probability Formula. The values a for which P (fW = ag) > 0 are 0; 1; 2; and 3;

2
and so P (fW 2 Ag) = 1; where A = f0; 1; 2; 3g : When using the Fundamental
Probability Formula for these exercises, we can intersect the interval B with the
set A; as follows:

P (f1 < W 2g) = pW (2) = 0:375;

P (f1 < W < 2g) = 0;
P (f1 W < 2g) = pW (1) = 0:375;
P (1 W 2) = pW (1) + pW (2) = 0:75:

3.1.2 The exercises

a) Solve Exercise 8.28a from the textbook.

b) Solve Exercise 8.28f from the textbook.

3.2 Exercise 2) (10 points –5 points per part) (cdfs)

3.2.1 The exercises
a) Let X be the random variable from HW 5 Exercise 2. Give the cdf, F; of X
as a formula. Be sure that your formula gives the values of F (x) for all real x:
Hint: Proceed as in Example 8.3 on pp. 407-408 of Weiss’s book. Study that
example carefully before proceeding. Also study the many related examples in
Hsu’s book from the HW 5 reading assignment.
Recall that the graph of the cdf of a random variable X can jump only at
points a for which P (fX = ag) > 0: For any random variable, the set of such
points (those points are the points of discontinuity of the cdf) is countable, and
the amount by which the cdf jumps at such points is F (a) F (a ) ; which
equals P (fX = ag) : Moreover, when X is discrete (and only in that case), the
values of P (fX = ag) add up to 1:

b) Let X be the random variable from HW 5 Exercise 2. Graph the cdf, F; of

X:
Hint: See the hint for part a).

3.3 Exercise 3) (25 points –5 points per part) (absolutely

continuous random variables)
3.3.1 The exercises
Suppose Y is a random variable whose cdf is given by
8
>
> 0; if y < 0
< 1
5 y; if 0 y<3
FY (y) = 1 :
>
> (y + 3) ; if 3 y<7
: 10
1; if y 7

3
a) Prove that Y is an absolutely continuous random variable, and …nd a pdf for
Y; proceeding as in the class notes.
Suggestion: I am not requiring that you do this on this particular problem,
but I recommend that you quickly verify that the function you …nd as a pdf
is non-negative and integrates to 1; that’s a good way to catch most careless
errors.

b) Calculate P (f2 Y < 6g) by using the pdf you found in part a.

c) Calculate P (f2 Y < 6g) by using the cdf.

d) Calculate P (fY < 6g j fY 2g) :

e) Are the events fY < 6g and fY 2g independent?

3.4 Exercise 4) (10 points) (cdfs and probability)

3.4.1 The exercise
Suppose that X is a random variable whose cdf is given by
8
>
> 0; if x < 2
<
(x 1) =4; if 2 x < 3
F (x) = 2 ;
>
> (1=2) x 3x + 5; if 3 x < 4
:
1; if x 4
and suppose we attempt to calculate P (f1 X 7=2g) in two ways as follows:
Calculation 1: Setting f = F 0 in each open interval, and setting f = 0 at the
endpoints, we get the pdf
8
>
> 0; if x < 2
>
>
< 1=4; if 2 < x < 3
f (x) = x 3; if 3 < x < 4
>
>
>
> 0; if x > 4
:
0; if x 2 f2; 3; 4g
8
< 1=4; if 2 < x < 3
= x 3; if 3 < x < 4 ;
:
0; otherwise
and so by the Fundamental Probability Formula we have
Z 7=2
P (f1 X 7=2g) = f (x) dx
1
Z 2 Z 3 Z 7=2
= f (x) dx + f (x) dx + f (x) dx
1 2 3
Z 2 Z 3 Z 7=2
1
= 0dx + dx + (x 3) dx
1 2 4 3
= 3=8:

4
Calculation 2: Using the cdf, we have

P (f1 X 7=2g) = F (7=2) F (1 )

= (5=8) lim F (x)
x!1
= (5=8) 0
= 5=8:

These calculations yield unequal answers, so clearly one or both of them is

incorrect. Study them both.
For each calculation, comment on whether it is completely correct or not. If
it is not completely correct, explain exactly what went wrong. Be precise, and
explain clearly.

3.5 Exercise 5) (30 points – 5 points per part) (…nding

cdfs and using them to …nd pdfs)
3.5.1 Background
Given the cdf of a random variable, in certain cases (see the class notes) we can
conclude that the random variable is absolutely continuous, and we can …nd a
pdf from the cdf. Therefore, it is important to know how to …nd the cdf of a
random variable.
One way is to use the de…nition of the cdf, F (x) = P (fX xg) ; together
with the de…nition of the random variable. If X represents a non-negative
quantity, then F (x) = P (fX xg) will be 0 for all x < 0: In such situations,
it is helpful to proceed by cases, for instance …nding F (x) for all x < 0 …rst, or
even for all x 0; and then …nding F (x) for all x > 0: Once F (x) is known, we
can test to see whether or not we can conclude that it is absolutely continuous.
Study Examples 8.5, 8.6, and 8.8 from pp. 409-411, as well as Example 8.19
from pp. 465-466 of Weiss’s textbook carefully before proceeding.
It will also help to review the chain rule for di¤erentiation, from …rst semester
calculus. Using the current notation, if

FY (y) = FX (g (y))

for some open interval of y values, and if FY is di¤erentiable at y; then we can

5
let u = g (y) and proceed as follows:

0 d
FY (y) = FY (y)
dy
d
= FX (g (y))
dy
d
= FX (u)
dy
d du
= FX (u)
du dy
du
= fX (u)
dy
= fX (g (y)) g 0 (y) ;

assuming fX is continuous at u:

3.5.2 The exercises

Important For several of these exercises, the “univariate transformation the-
orem” from Weiss’s book (in Section 8.7) could in principle be used to
help solve the problem. Do not use it. Instead, proceed as indicated be-
low (which in various cases corresponds to the “cdf method”from Weiss’s
book). You may use the univariate transformation theorem to check your
work if you wish, but not in place of the methods indicated below (which
are the ones on which you are being graded).

a) A point is selected at random from a solid ball (in 3-dimensional space)

having radius 10: Let X denote the distance from the point to the outer edge
of the ball. Find the cdf of X: Is X absolutely continuous? If so, explain why,
and …nd a pdf of X: If not, explain why not.
Hint: Remember to use the meaning of X: For instance, fX 1g is the set of
all points within a distance 1 of the outer spherical boundary, and hence it is the
set of points whose distance from the center of the sphere is between 9 and 10;
it is a spherical shell. Its volume can be found using geometry. Its volume can
be used to …nd P (fX 1g) ; which is F (1) : Values of F (x) for other x values
can be found similarly, but be sure to proceed as described in the background
section above.

b) Let T be a triangle (the boundary and interior) in the plane, having base
length 4 and height 13: As it turns out, the shape of the triangle itself will not
matter (but its base length and height will).
A point ! is chosen at random from the interior of T:
We de…ne a function Y : ! R as follows: for each ! 2 ; we let Y (!)
denote the perpendicular distance from ! to the base of the triangle.
Find FY ; the cdf of Y:

6
Hint: First, the area of any triangle is half the length of its base multiplied by
its height. Now proceed as follows:
1) On a sheet of paper draw a triangle with base 4 and height 13:
2) Draw a “random” point inside the triangle. Call the point A:
3) Draw a dotted line segment connecting the point A to the base (extend
the base parallel to itself with dotted lines if necessary), making sure that the
dotted line segment is perpendicular to the base (this is crucial). Let y denote
the distance, along this line segment, from the point A to the base.
4) Draw a dotted line segment parallel to the base and passing through the
point A; and extend it so that it meets both sides (other than the base) of the
triangle.
5) The dotted line segment you drew in 4) divides the triangle T into two parts.
It is easy to show with elementary geometry that the part of T on or above
that line segment is similar (look up “similar triangles” if you wish to review
the topic) to the entire triangle T: The part of T on or above the dotted line
from 4) is fY > yg : By formulas concerning similar triangles (namely, that the
ratio of their areas is the square of the ratios of their heights), we have
2
Area (fY > yg) (height of fY > yg)
= :
Area (T ) height of T

The height of the top triangle fY > yg can be found easily in terms of h
and/or y using your picture.
The height of T can be found easily in terms of h and/or y:
Area (T ) can be found easily in terms of h and/or y:
You can now easily solve for Area (fY > yg) : Then use it to …nd P (fY > yg) :
You can then use that to …nd P (fY yg) ; which you can use to get the cdf of
Y: Remember to proceed with cases, as described in the background above.

c) Let Y be the random variable from part b). Is Y absolutely continuous? If

so, explain why, and …nd a pdf of Y: If not, explain why not.

d) Let X Exp (5) ; and let Y = max (X; 7) : I.e., if X (!) 7 then Y (!) = 7;
while if X (!) > 7 then Y (!) = X (!) : Find the cdf of X: Is X absolutely
continuous? If so, explain why, and …nd a pdf of X: If not, explain why not.
Hint: Let FX and FY denote the cdfs for X and for Y; respectively. Finding
FX is easy, since there is a right tail probability formula for exponential random
variables. The strategy is to …nd FY in terms of FX : For any real number y;
consider the event fY yg : There are two cases: y 7 and y < 7:
Case 1: y < 7: In this case, Y y if and only if both X and 7 are less than
or equal to y: However, 7 cannot be less than or equal to y in Case 1, so Y y
cannot occur. I.e., fY yg = ;; and so FY (y) = P (fY yg) = 0:
Case 2: y 7: In this case, Y y if and only if both X and 7 are less than
or equal to y; which (since we’re in Case 2) is true if and only if X y: I.e.,

7
fY yg = fX yg ; and so FY (y) = P (fY yg) = P (fX yg) = FX (y) ;
which can be found once we …nd the cdf of X:

e) Solve Exercise 8.154 part a) from Weiss’s textbook.

f ) Suppose X N 3; 42 ; and let Y = (X 3) =4: Prove that Y N 0; 12

as follows: …nd a formula expressing FY in terms of FX : Find a formula for
fY : Show that your formula for fY agrees with the pdf of a standard normal
random variable except at most at a set of points having length zero. It then
follows that Y has the same distribution as a standard normal random variable.

3.6 Exercise 6) (30 points – 5 points per part) (normal

random variables)
3.6.1 Background
Normal random variables are among the most important kind of absolutely con-
tinuous random variable because they appear in many applications, especially in
statistics. They are among the most ubiquitous random variables in probability,
because of the Central Limit Theorem (which we will study later). They also
appear as approximations to binomial random variables, when p is near 1=2:
We say that a random variable X is a normal random variable, and we
write X N ; 2 ; provided it has a pdf of the form

1 (1=2)[(x )= ]2
f (x) = p e ;
2
where is a real number (called the mean or expected value of X) and is a
positive real number (called the standard deviation of X). These assumptions
about and are always to be made when we write X N ; 2 :
The quantity 2 ; the square of the standard deviation, is called the variance
of X:
With a fair amount of work (see the class notes; it is an exercise in integra-
tion), it can be shown that f is a legitimate pdf for each and as above.
The graph of f is bell-shaped and is called a “bell curve.” Its center (where
the highest point of the curve is located) is at x = ; and the graph is left-right
symmetric about the vertical line x = : Varying a¤ects how much of the
area will be located near the mean, : smaller values of correspond to taller,
narrower graphs with most of the area close to the line x = :
Your calculator has built-in features for approximate evaluation of integrals
of normal pdfs. Start a new document, select Add Calculator, and then press
MENU-5-5-2 to select the normal cdf option. Then enter the left endpoint, the
right endpoint, ; and in the designated …elds, and press ENTER. You will
get the probability that the normal random variable you speci…ed (with and
as you entered them) takes a value between the two endpoints you selected.
For practice, try setting 5 as the left endpoint, 11 as the right endpoint, = 8;
and = 3: You should get about 0:6827:

8
 Be sure you know how to use that feature and that you also can evaluate
integrals of normal pdfs by entering the pdf directly. I recommend you try both
methods for each of these problems to check your work.
Suppose X B (n; p) ; where n is a positive integer and p 2 (0; 1) :
Suppose that B is a union of intervals and/or integers, and suppose we wish
to estimate
P (fX 2 Bg) :
When p is su¢ ciently close to 0 or to 1; we can use a Poisson approximation
(switching the roles of success and failure if p is close enough to 1), but what
about when p is not close to either?
It turns out that when p is far enough away from both 0 and 1 we may use
a normal approximation to the binomial! Our speci…c criteria in practice are

np 5 and n (1 p) 5 simultaneously.

We let Y N ; 2 ; with and 2

chosen so that X and Y will have
matching means and variances, namely
p
= np; = np (1 p):

There is one major issue when using this approximation, however: Y is

absolutely continuous, while X is discrete!
As a result, we make use of the so-called continuity correction: we treat
each integer as though it were an interval having length one centered at that
integer, which is perhaps the most natural thing to do.
Thus:

We will approximate P (fX = 10g) by P (f9:5 < Y < 10:5g) :

We will approximate P (f10 X 11g) by P (f9:5 < Y < 11:5g) :

We will approximate P (f10 X < 20g) ; which equals P (f10 X 19g) ;

by P (f9:5 < Y < 19:5g) :
We will approximate P (f5 X 7g [ f12 X 15g) by P (f4:5 < Y < 7:5g)+
P (f11:5 < Y < 15:5g) :

3.6.2 The exercises

a) On the same graph, draw the pdfs for N 2; 122 ; N 2; 62 ; and N 2; 42
random variables. You may use your calculator to sketch them, but copy them
over onto a single graph and label the x and y axes. Be sure to identify which
is which in some clear way. The reason I want you to draw them on the same
graph is that I want you to see clearly the e¤ect of keeping the mean the same
but varying : Make sure that your graphs make this clear.

9
b) Normal random variables are so frequently used that there is a “rule” for
estimating certain probabilities involving normal random variables. The Em-
pirical Rule says that the probability

P (f k <X< + k g)

that a normal random variable X takes a value within k standard deviations

of its mean is approximately (rounded to 4 decimal places) 0:6827; 0:9545; and
0:9973 when k = 1; 2; and 3; respectively. This is true regardless of the value
of the mean and standard deviation of X: Of course, since normal random
variables are continuous (they have pdfs, so they are absolutely continuous,
hence in particular continuous), we have

P (f k X + k g) = P (f k <X< + k g)

as well.
For example, a N 0; 12 random variable is called a standard normal
random variable. It is easy to see with a calculator that the probability X
takes a value within 3 standard deviations of its mean (i.e., between 0 3 1 = 3
and 0 + 3 1 = 3) is
Z 3
1 (1=2)x2
p e dx = 0:99730020 : : :
3 2

Rounding to four decimal places, …nd the values of P (f 10 < X < 14g) for
N 2; 122 ; N 2; 62 ; and N 2; 42 random variables.

c) Suppose X N ; 2 ; let ( ; F; P ) denote the probability space over which

X is de…ned, and de…ne a new random variable Z by the formula

X (!)
Z (!) = ; whenever ! 2 :

Show that the events

R = fa < X < bg
and
a b
S= <Z<

are the same. This implies that P (R) = P (S) :

Hint: ! satis…es X (!) < b if and only if it satis…es X (!) <b ; which
is true if and only if it satis…es (X (!) ) = < (b ) = ; and therefore

b
fX < bg = Z< :

Proceed the same way but with double inequalities at each step.

10
d) Solve Exercise 8.104a from the textbook as follows. Let X be the measured
BAC with units of percents. We are told that X is a normal random variable
with = 0:11 and = 0:005: Find a pdf of X; and use it and the Fundamental
Probability Formula to calculate the required probability. Remember that X is
in terms of percents, so do not divide any percents by 100.

e) Solve Exercise 8.104b from the textbook by proceeding as in part d.

f ) In the normal approximation to the binomial, we approximate a B (n; p)

random variable with a N ; 2 random variable having = np and 2 =
np (1 p) : In this way, the random variables have equal means, and the random
variables have equal variances.
Normal pdfs are symmetric about their means. However, binomial pmfs are
approximately symmetric (which is essential for a good approximation by a
normal) only when p is far from either 0 or 1:
Use the normal approximation to the binomial, as described above, to solve
Exercise 11.55b from the textbook. When doing so, be sure to demonstrate
that the conditions np 5 and n (1 p) 5 are satis…ed, as required for our
procedure.

3.7 Exercise 7) (65 points –5 points per part) (expected

value and variance)
3.7.1 Background
If we know the random variable type and the parameters of a random variable X;
we can in many cases look up its expected value and variance. Weiss’s textbook
has several tables with this information:

E (X) V AR (X)
X is discrete Table 7.4, p. 332 Table 7.10, p. 358
X is absolutely continuous Table 10.1, p. 568 Table 10.2, p. 588

Additionally, using linearity properties of summation and of integration, we

can derive formulas for the expected value, variance, and standard deviation of
an a¢ ne transformation Y = aX + b (with a 6= 0) of a random variable X:
Speci…cally,
E (aX + b) = aE (X) + b;
provided X has …nite expectation, and also

V AR (aX + b) = a2 V AR (X) ;
ST DDEV (aX + b) = jaj ST DDEV (X) ;

provided X 2 has …nite expectation.

11
3.7.2 Instructions and hints
For each of the following random variables, …rst indicate what type of random
variable X is (its type and parameter(s), if this information is not given). For
these exercises only (to save you time), you do not need to justify your claims
about what type of random variable X is. You can just assert X B (20; 0:4) ;
or X G (0:9) ; or whatever, without proof or justi…cation.
If you can use the tables in the book to quickly evaluate expected values,
variances, or standard deviations for these exercises then do so.
If you can use the a¢ ne transformation formulas above to quickly evaluate
expected values, variances, or standard deviations for these exercises then do
so.
If no other way Zworks, you can use the de…nition of expected value (as
X
x p (x) ; or as x f (x) dx), or the de…nition of variance, to …nd the
expected value or variance of the given random variable (and with the variance
you can easily …nd the standard deviation).
In some cases, you will be able to proceed quite quickly as follows. If you
are able to determine, for instance, that X B (10; 0:6) ; then you can use the
tables to quickly …nd Ep(X) = 10 0:6p= 6; V AR (X) = 10 0:6 0:4 = 2:4;
and ST DDEV (X) = V AR (X) = 2:4: If you need V AR (2X 7) ; that
is simply 22 V AR (X) = 9:6; according to the a¢ ne transformation formulas
above. Many of the problems below can be solved quickly in this fashion. In
other cases, you will have to work harder.
In each case, use exact distributions –do not approximate X with any other
type of random variable, even if such an approximation would be reasonable to
make.
Also, saying “X has a binomial distribution” is the same as saying “X is a
binomial random variable.”
Some of the random variable types might be types we have not yet worked
with. They are all in the book, as is the relevant information about their
expected values and variances.
All …nal answers for E (X) and V AR (X) must be numerical and cannot
have any unknown parameters in them. Enough information is given to allow
you to solve for all of the parameters exactly.
The formulas for E (X) and V AR (X) can help us determine the parameters
for X: For instance, if we are told X has a binomial distribution with n = 20; and
if we are further told that X has an average value of 8; then since E (X) = np
we can substitute E (X) = 8 and n = 20 to …nd that p = 0:4; and we can use
that information to …nd V AR (X) :

3.7.3 The exercises

a) An experiment consists of rolling a 10-sided die, with sides numbered 0, 1,
2, 3, 4, 5, 6, 7, 8, and 9. We will roll the die 4 times. Let X denote the number
of rolls which are 6 or less. Find E (X) and V AR (X) :

12
b) The Economic Security Index (ESI) is one measure of the percentage of
Americans who have experienced a “signi…cant …nancial loss.” For a more pre-
cise de…nition of this index, and for more details, you may consult economicse-
curityindex.org if you wish (it is not necessary to do so to solve this problem).
State by state results are also available. One of the reports from that site can
be found here (it is not necessary to read this to solve this problem):
http://economicsecurityindex.org/upload/media/esiupdate_11_16_13.pdf
The ESI (nationwide) for 2012 was estimated to be about 17.8. I.e., in
the technical sense of the de…nition of the ESI, about 17.8% of Americans had
experienced a signi…cant …nancial loss. This was an improvement compared to
2011 when the number was 18.9%.
A random sample of 10 Americans is selected. Let X denote the number of
Americans in our selected sample who had experienced a signi…cant …nancial
loss in the technical sense of the de…nition of the ESI. Find E (X) and V AR (X) ;
assuming there were 315,000,000 Americans in 2012.

c) Suppose the number of cars getting onto the Key Bridge from Virginia be-
tween 3 p.m. and 3:30 p.m. on a Monday, Tuesday, Wednesday, or Thursday
in November has a Poisson distribution with an average of 14 cars per minute.
Let X be the number of cars getting onto the Key Bridge on a given Wednesday
in November, between 3:07 p.m. and 0 seconds and 3:08 p.m. and 0 seconds.
Find E (X) and V AR (X) :

d) Suppose a certain type of anti-ballistic missile (ABM) can e¤ectively inter-

cept a particular type of incoming missile about 40% of the time. Suppose
ABMs are launched in an attempt to intercept an incoming missile, and let X
be the number of missiles used altogether in the interception process (i.e., if the
…rst two ABMs fail and the third succeeds, then X = 3). (For the purposes of
this problem, suppose su¢ cient independence that we may regard the launches
of the ABMs as Bernoulli trials.)
Find E (X) and V AR (X) :

e) A certain pharmaceutical company is interested in developing a vaccine for

a given disease. One researcher for the company believes that a successful
vaccine can be manufactured using blood components from individuals having
a particular variety of antibody in their blood. It is estimated that 2% of
individuals have this variety of antibody in their blood. Furthermore, since
there are limits on how much blood may be safely drawn from a donor, it has
been determined that 30 individuals with this particular variety of antibody
must be found in order to develop a su¢ cient supply of the vaccine for testing
and clinical trials.
Since they have not found any way to know in advance whether a given
person has this variety of antibody or not without doing a blood test, the com-
pany advertises that they are willing to pay anyone willing to undergo a blood
test, and willing to allow their blood (if it turns out they have this variety of
antibody) to be used for testing and clinical trials.

13
 People are tested one at a time until 30 individuals with the crucial variety of
antibody have been found, and then no one else is tested. Let X be the number
of individuals who receive blood tests as a result of these e¤orts. E.g., if the
2000th person tested is the 30th person to have this variety of antibody then
X = 2000: (For the purposes of this problem, suppose su¢ cient independence
that we may regard the blood tests as Bernoulli trials.)
Find E (X) and V AR (X) :

f ): Suppose X has a chi-square distribution and that V AR (X) = 10: What is

E (X)?

g): Suppose X has an exponential distribution with E (X) = 4: What is

V AR (X)?

h): Suppose X has a Gamma distribution with E (X) = 2; V AR (X) = 1=2:

What are its parameters and ?

i) Let X B (4; 0:2) : Find V AR (2X 3) by any legitimate method.

j) Let X Exp (5) : Find V AR (4X + 1) by any legitimate method.

k) Let X be a discrete random variable that takes the values 7; 3; 1; and 4

with probabilities 3=18; 2=18; 5=18; and 8=18; respectively. Find V AR (4X + 1) :

l) Let X U ([1; 4] [ [10; 22]) : Find V AR (3X 9) by any legitimate method.

2
Hint: First, explain why X and X have …nite expectation. Find E (X) : Find
E X 2 : Then …nd V AR (X) : Finally, …nd V AR (3X 9) :

m) Let X U (( 7; 3)) ; and let Y = jXj : Find V AR (Y ) :

14