Unit 12

Probability and Calculus II

Continuous Distributions
So far, we have only been concerned with discrete random variables. Re-
call that a discrete random variable is one that can take any one of (only)
a nite number of values. In this section, we study another kind of random
variable, called a continuous random variable.
Denition 12.1. A Continuous Random Variable is one which may assume
any of an innite number of values on some interval of the real number line.
Because a continuous R.V. can have any of an innite number of values,
events are dened as sub-intervals in which the value of the R.V. could fall
(i.e. we never dene events as specic values of the R.V.).
Example 1. Consider the experiment: randomly pick a number between 0
and 1 inclusive. Let X be the number picked, so that X is a continuous
random variable. Identify the sample space, S, for this experiment and some
events on that sample space.
Solution: Here, X can take on any of the innitely many numbers which
lie in the interval [0, 1]. So it is this interval that is the sample space, i.e.,
S = {0 X 1} or S = {x|0 x 1}.
Notice: The notation {x| some condition } says the set of all xs such
that the specied condition is satised. That is, the | means such that.
Any sub-interval of the sample space is a possible event. For instance, we
might be interested in the event A, that the number chosen is bigger than
.5, or in event B, that a number between
1
4
and
1
3
is chosen. We have
A = {0.5 < X 1} and B =
_
1
4
X
1
3
_
.
1
For a continuous random variable, because there are so many possible
values, the probability of any exact particular value occurring is eectively
0. That is:
Theorem: For a continuous R.V., X, prob{X = x} = 0 for all values x.
(This is why we dont bother with events which correspond to specic values
of X.)
Example 2. Identify the following random variables as either Discrete or Con-
tinuous.
(a) X is the sum of 2 dice when rolled ;
(b) Y is the height of a randomly selected calculus student.
Solution:
(a) This is a discrete random variable because there are only 11 possible
values (the integers 2 through 12).
(b) A persons height, when measured precisely, can have any real value
within some range. That is, we could identify limits on the possible height
values which might be observed - an interval outside of which a persons
height could not possibly be. For instance, it would appear reasonable to
assume that no calculus student could ever be less that 1 foot tall, or more
than 10 feet tall. However, within a certain range, any value could be ob-
served. So in this case, Y is a continuous random variable.
Notice: This means that the probability of any particular height is 0. For
instance, measuring in feet, prob{X = 6} = 0, i.e., the probability that some-
one is exactly 6 feet tall, not 6.001 feet or 5.999 feet or 5.9999982 feet, is 0.
However, the probability that someone is within, say, 0.1 feet of being 6 feet
in height, i.e., prob{5.9 X 6.1} is not 0.
Denition 12.2. Associated with any Continuous Random Variable X,
there is a function, f(x), called its probability density function, which satises
prob{c X d} =
_
d
c
f(x)dx
Notice: This probability is the area under the curve y = f(x), between x = c
and x = d.
2
In order to be a probability density function (often abbreviated as pdf),
f(x) must satisfy both of the following conditions:
1. f(x) 0 for all values of x and
2. The total area under the curve y = f(x) must be 1 i.e.
_

f(x)dx = 1
That is, we know that probabilities must always be between 0 and 1. In
order to ensure that
_
d
c
f(x)dx is non-negative for all possible c and d values,
the curve y = f(x) must never fall below the x-axis. Also, we must have (as
always) Prob(S) = 1, so the total area under the curve must be 1.
Note: If f(x) is a probability density function, we consider its domain to
be all real values of x. Typically, f(x) is dened as some particular func-
tion in some interval (the sample space), and is dened as f(x) = 0 elsewhere.
Notice also: Recognizing that the probability of an event involving a contin-
uous R.V., X, is an area helps us to understand why prob{X = x} = 0. Even
if c is a possible value of X, we have prob{X = c} =
_
c
c
f(x)dx = 0. That
is, the probability that X is exactly c is the area under the curve y = f(x)
right at x = c. But this region is just a line segment, and its area is 0.
And notice: Since prob{X = x} = 0 for all x, then there is no dierence be-
tween prob{X x} and prob{X < x}. That is, with continuous R.V.s, we
do not need to distinguish between the events {X x} and {X < x}. This
is not true for discrete random variables. If X is a discrete random variable
and x is one of its possible values, then these 2 events are very dierent.
Example 3. Show that f(x) =
3
8
x
2
when 0 x 2 and f(x) = 0 otherwise
could be a pdf for some random variable.
Solution: We need to show that this function has the 2 properties required
of a probability density function. First, we observe that
3
8
x
2
0 for all x
in the interval [0, 2]. So we have f(x) = 0 on (, 0) and on (2, ), and
f(x) 0 on [0, 2] and hence f(x) 0 for all values of x, as required.
We must also verify that the total area under y = f(x) is 1. Of course, we
have:
_

f(x)dx =
_
0

f(x)dx +
_
2
0
f(x)dx +
_

2
f(x)dx
3
So then
_

f(x)dx =
_
0

0dx +
_
2
0
3
8
x
2
dx +
_

2
0dx
= 0 +
3
8
_
2
0
x
2
dx + 0
=
3
8
_
x
3
3
_
2
0
=
3
8
_
2
3
0
3
3
_
= 1
We see that both of the requirements of a pdf are satised, so this function
could be the pdf for some random variable.
Example 4. Find the value of the constant k such that the function f(x) = kx
on the interval [3,5] and f(x) = 0 elsewhere is a probability density function.
Solution: We need to have f(x) 0 for all values of x. Clearly, we need k to
be some non-negative constant, in order to have kx 0 on [3,5], so we know
that k 0. To nd the specic value of k, we use the fact that we need the
total area under y = f(x) to be 1, i.e. we need
_

f(x)dx = 1. Of course,
since f(x) = 0 everywhere outside the interval [3,5], then we have
_

f(x)dx =
_
5
3
kxdx
so we actually need
_
5
3
kxdx = 1
We get:
_
5
3
kxdx = k
_
5
3
xdx = k
_
x
2
2
_
5
3
=
k
2
_
5
2
3
2
_
=
16k
2
= 8k
Thus we see that we need 8k = 1, so k =
1
8
. That is, f(x) is a pdf only when
k =
1
8
.
4
Example 5. The continuous random variable X has pdf f(x) =
4
225
(x
3
x)
on the interval [1,4] and f(x) = 0 elsewhere. Find the probability that X
has a value between 2 and 3.
Solution: We see that X has possible values which lie on the interval [1,4].
Also, we know that the probability that X has a value between some num-
bers a and b which are in the interval on which the pdf is non-zero is given
by prob{a X b} =
_
b
a
f(x)dx. We are looking for prob{2 X 3}, and
both 2 and 3 lie in [1, 4], so f(x) =
4
225
(x
3
x) on [2, 3]. Thus we have
prob{2 X 3} =
_
3
2
4
225
(x
3
x)dx
=
4
225
_
x
4
4

x
2
2
_
3
2
=
4
225
__
3
4
4

3
2
2
_

_
2
4
4

2
2
2
__
=
4
225
_
81
4

9
2

16
4
+
4
2
_
=
4
225
_
55
4
_
=
11
45
We see that prob{2 X 3} =
11
45
.
Example 6. For the continuous random variable X in the preceding example,
nd prob{X 2}.
Solution: Here, the possible values of X are the numbers in the interval
[1,4], so {X 2} means {1 X 2}. That is, technically the event
{X 2} is the event { X 2}, but since f(x) = 0 on (, 1) then
prob{ X 1} = 0 and we see that
prob{X 2} = prob{ X 2}
= prob{ X 1} + prob{1 X 2}
= 0 +prob{1 X 2}
= prob{1 X 2}
5
Thus we have
prob{X 2} =
_
2
1
4
225
(x
3
x)dx =
4
225
_
x
4
4

x
2
2
_
2
1
=
4
225
__
2
4
4

2
2
2
_

_
1
4
4

1
2
2
__
=
4
225
__
16
4

4
2
_

_
1
4

1
2
__
=
4
225
_
2
_

1
4
__
=
4
225
_
9
4
_
=
9
225
Denition 12.3. Let X be a Continuous Random Variable with probability
density function f(x). Then , the mean of X, is given by
=
_

x f(x)dx
and
2
, the variance of X, is given by

2
=
_

(x )
2
f(x)dx =
_

x
2
f(x)dx
2
Also, the standard deviation of X is dened to be the positive square root
of the variance,
2
, and is denoted by . That is, we dene the standard
deviation of X as:
=

2
Notes:
1. Notice the similarities to the denitions of mean and variance of a
discrete random variable. Summing over all discrete possible values
becomes integrating over all continuous possible values, i.e. over the
entire interval over which the density function is dened.
2. Just as in the case of discrete random variables, we have 2 formulas
for variance here. The rst is, technically, the denition of variance,
while the second (which can easily be shown to be equivalent) is easier
to use.
3. Because of the way that variance is dened, variance is always non-
negative. That is, (x )
2
0 and also f(x) 0, so it must be true
that
_

(x )
2
f(x)dx 0. Therefore, =

2
is always dened,
for any continuous random variable.
6
4. As we have seen before, if the pdf of X is only non-zero on some interval
[a, b], then for both mean and variance we need only integrate from a
to b, rather than from to .
Example 7. Let X be a continuous random variable with pdf f(x) =
3
8
x
2
on
[0, 2] and f(x) = 0 elsewhere. Find the mean and standard deviation of X.
Solution: We have =
_

xf(x)dx, but f(x) = 0 everywhere outside of the

interval [0, 2]. Thus we get
=
_
2
0
x
_
3
8
x
2
_
dx =
3
8
_
2
0
x
3
dx =
3
8
_
x
4
4
_
2
0
=
3
8
_
2
4
4

0
4
4
_
=
_
3
8
_
(4) =
3
2
To nd the standard deviation, , we must rst nd the variance,
2
. We
have

2
=
_

x
2
f(x)dx
2
=
_
2
0
(x
2
)
_
3
8
x
2
_
dx
_
3
2
_
2
=
3
8
_
2
0
x
4
dx
9
4
=
3
8
_
x
5
5
_
2
0

9
4
=
3
8
_
2
5
5

0
5
5
_

9
4
=
3
8
_
32
5
_

9
4
=
12
5

9
4
=
48 45
20
=
3
20
We see that the variance of X is
2
=
3
20
and so the standard deviation of X
is =

2
=
_
3
20
.
Example 8. X is a continuous random variable with pdf f(x) =
1
2
sin x on
[0, ] and f(x) = 0 elsewhere. Find the mean and variance of X.
Solution: First of all, notice that f(x) =
1
2
sin x on [0, ] and f(x) = 0 el-
swhere does meet the requirements of a pdf. From x = 0 to x = (i.e.,
in the rst and second quadrants), the function y = sin x increases from 0
to 1 (at x =

2
) and then decreases back to 0. Therefore for the function
f(x) =
1
2
sin x on [0, ] and f(x) = 0 elswhere, we have f(x) 0 everywhere.
7
As well, we see that
_

f(x)dx =
_
0

0dx +
_

0
1
2
sin xdx +
_

0dx
=
1
2
_

0
sin xdx =
1
2
[cos x]

0
=
1
2
[cos cos 0] =
1
2
[(1) (1)] =
1
2
(2) = 1
Now, we need to nd the mean of X. We have
=
_

xf(x)dx =
_

0
x
_
1
2
sin x
_
dx =
_

0
xsin x
2
dx
To evaluate this, we use integration by parts, letting u =
x
2
and dv = sin xdx
so that du =
1
2
dx and v = cos x. We get:
=
_

0
xsin x
2
dx =
xcos x
2

_

0
cos x
2
dx
=
_

0
cos x
2
dx
xcos x
2

0
=
sin x
2

xcos x
2

0
=
_
(sin sin 0) ( cos 0 cos 0)
2
_
=
_
0 0 (1) + 0(1)
2
_
=

2
Therefore the mean of X is =

2
. For the variance of X we need to nd

2
=
_

x
2
f(x)dx
2
=
_

0
x
2
_
1
2
sin x
_
dx
_

2
_
2
To evaluate the integral here, we must use integration by parts twice. First,
let u =
x
2
2
and dv = sin xdx, so that du = xdx and v = cos x. Then we get
_
x
2
sin x
2
dx =
x
2
cos x
2

_
(xcos x)dx =
x
2
cos x
2
+
_
xcos xdx
Now, to nd
_
xcos xdx, we use u = x and dv = cos x, so that du = dx and
v = sin x. We get
_
xcos xdx = xsin x
_
sin xdx = xsin x(cos x)+C = xsin x+cos x+C
8
Thus we have
_
x
2
sin x
2
dx =
x
2
cos x
2
+ xsin x + cos x + C
This gives

2
=
_

0
x
2
sin x
2
dx
_

2
_
2
=
_

x
2
cos x
2
+ xsin x + cos x
_

2
_
2
=
__

2
cos
2
+ sin + cos
_

0
2
cos 0
2
+ 0 sin 0 + cos 0
__

2
4
=
__

2
(1)
2
+ (0) + (1)
_

_
0(1)
2
+ 0(0) + (1)
__

2
4
=
_

2
2
2
_

2
4
=

2
4
2
The Exponential Distribution
Next, we look at two useful types of continuous distributions which often
arise in practical situations. For the rst one, consider the following question.
Example 9. For what values of k is the function f(x) = ke
kx
on the interval
[0, ) and f(x) = 0 elsewhere a probability density function?
Solution: For f(x) to be a probability density function, we need f(x) 0 for
all x. Since e
kx
> 0 for all values of k and all values of x, then ke
kx
0
only for k 0. That is, we cannot have k < 0.
As well, we require that
_

f(x)dx = 1, so we must have

_

0
ke
kx
dx = 1.
If k = 0, then
_

0
ke
kx
dx =
_

0
0dx = 0, so we cannot have k = 0.
For any k > 0, we have
_

0
ke
kx
dx = lim
b
_
b
0
ke
kx
dx
9
And so
_

0
ke
kx
dx = lim
b
k
_
e
kx
k
_
b
0
= lim
b
_
e
kx

b
0
= lim
b
_
e
kb
+ e
0
_
= 1 lim
b
e
kb
But of course lim
b
e
kb
= lim
b
1
e
kb
, and since k > 0 then as b we have
kb and so e
kb
so that
1
e
kb
0.
Therefore
_

0
ke
kx
dx = 1 0 = 1 for any k > 0 and so the given function
is a probability density function for any k > 0.
This gives us an important family of continuous random variables.
Denition 12.4. A random variable, X, whose probability density function
has the form
f(x) = ke
kx
on [0, ) for some k > 0
= 0 on (, 0)
is called an exponentially distributed random variable. The value k, which
distinguishes one member of this family from another, is called a parameter.
Example 10. If the continuous random variable X has pdf f(x) = 2e
2x
on
[0, ) and f(x) = 0 elsewhere, nd prob{X 1}.
Solution: Notice that X is an exponential random variable. The parameter
value is k = 2. Here, as with any exponential random variable, f(x) is only
non-zero for x 0. So the event {X 1} is actually the event {0 X 1}.
Thus we get:
prob{X 1} = prob{0 X 1}
=
_
1
0
f(x)dx =
_
1
0
2e
2x
dx
10
So we have
prob{X 1} = 2
_
e
2x
2
_
1
0
= e
2x

1
0
= e
2
(e
0
) = e
2
+ 1
= 1
1
e
2
.86466
Notice: Since
_
ke
kx
dx = e
kx
+ C, in general for any exponential R.V.
X, we have
prob{0 X a} =
_
a
0
ke
ax
= e
kx

a
0
= e
ak
+ e
0
= 1 e
ak
Fact: Exponentially distributed R.V.s arise in situations involving random
arrivals. That is, if X measures the time between occurrences of some
event that occurs randomly, then X is a random variable with an exponen-
tial distribution. Examples: arrivals of customers dialling up to an internet
provider, arrivals of patients at an emergency room. (We refer to the time
between 2 consecutive arrivals as the interarrival time.)
Theorem 12.5. If X is a Random Variable with probability density function
f(x) = ke
kx
on [0, ) and f(x) = 0 elsewhere, then the mean and standard
derivation of X are respectively:
=
1
k
, =
1
k
That is, if X has an exponential distribution, then the mean and the
standard deviation of X are both given by
1
k
, where k is the parameter ap-
pearing in the pdf of X. (Note: Exponential R.V.s are the only R.V.s for
which = .)
The fact that =
1
k
can be turned around to tell us what the parameter
value is, if we know the average time between arrivals. For instance, if we
know that the average time between phone calls arriving at a switchboard
is 4 minutes (where the phone calls are assumed to arrive at random), then
we have = 4. But we know that =
1
k
, so then k =
1

=
1
4
. So if X is
11
the time between arrivals of phone calls, then X is exponentially distributed
with pdf f(x) =
1
4
e

1
4
x
on [0, ) (and f(x) = 0 elsewhere). Notice that for
exponentially distributed interarrival times, the parameter value is the mean
arrival rate. For instance, if the mean time between calls is 4 minutes, then
calls arrive at an average rate of
1
4
per minute.
Example 11. During a particularly busy period, patients arrive at the emer-
gency room at University Hospital at an average rate of 1 patient every
5 minutes. Assuming that the time between arrivals is exponentially dis-
tributed, nd the probability that 2 patients arrive within 1 minute.
Solution: Let T be the time between (consecutive) arrivals of patients. Then
T is an exponentially distributed random variable. Were told that the aver-
age time between arrivals is 5 minutes, so we have = 5. And we know that
=
1
k
, so we have
1
k
= 5 k =
1
5
= .2 (that is, patients arrive at an average
rate of .2 per minute). Therefore the pdf of T is f(t) = .2e
.2t
on [0, ) and
f(t) = 0 elsewhere. We need to nd prob{T 1}.
We can use our previous observation that for an exponentially distributed
R.V. with parameter k, we have prob{0 X a} = 1 e
ak
. In this case,
for the exponentially distributed R.V. T with parameter value k = .2, we get
prob{T 1} = prob{0 T 1}
= 1 e
(1)(.2)
= 1
1
e
.2
.18127
Example 12. For the period described in the previous example, what fraction
of patients will arrive at the emergency room at UH between 3 and 6 minutes
after the previous patient arrived?
Solution: Notice that What fraction of patients will arrive between 3 and
6 minutes after the previous patient arrived? means what fraction of in-
terarrival times are between 3 and 6 minutes?, which is another way of
saying what is the probability that the time between consecutive arrivals is
between 3 and 6 minutes?. That is, we have T being the random variable
dened in the previous example and we want to nd prob{3 T 6}.
12
We get:
prob{3 T 6} =
_
6
3
f(t)dt =
_
6
3
.2e
.2t
dt
= e
.2t

6
3
= e
.2(6)
(e
.2(3)
)
= e
.6
e
1.2
.2476
We see that approximately
1
4
of patients will arrive between 3 and 6 minutes
after the previous patient arrived.
The Normal Distribution
Denition 12.6. A Random Variable X whose probability density function
has the form
f(x) =
1

2
e

1
2
(
x

)
2
, on (, )
for some values of and , with > 0, is called a normally distributed ran-
dom variable.
The Normal Distribution is the most important, i.e., the most commonly
used, of all probability distributions. This is because many phenomena in
the real world are either normally distributed or at least approximately nor-
mally distributed. Hence the normal distribution gives at least a very good
approximation of the probabilities involved in many real situations.
Notice that the pdf of the normal distribution involves 2 parameters,
shown above as and . The reason that we use and to represent these
2 parameters is, of course, because the parameter values turn out to be the
mean and the standard deviation, respectively, of the distribution. That is,
it can be shown that:
Theorem 12.7. A normally distributed random variable X, with the pdf:
f(x) =
1

2
e

1
2
(
x

)
2
on (, )
has mean and standard derivation .
13
There are a whole family of Normal Distributions, with dierent values
of and . The pdf of a normally distributed random variable is the fa-
miliar bell-curve or bell-shaped curve. The graph is symmetric about
the value , with the height and thickness of the bell determined by . (A
smaller value of gives a taller, thinner bell-shape, while a larger value of
gives a shorter, squatter bell.) Of course, the total area under the curve (i.e.
under the bell) is always 1.
There is one very special particular Normal Distribution which is very
useful the one which has mean = 0 and standard deviation = 1. A
normal R.V. with this particular distribution is useful because any other
normally distributed R.V. can easily be expressed in terms of this Standard
Normal Random Variable. This allows probabilities for any normally dis-
tributed R.V. to be easily calculated, as long as we know probability values
for the standard normal random variable. By convention, this standard nor-
mal random variable is represented by Z.
Denition 12.8. The standard normal random variable Z is normally dis-
tributed with = 0, = 1. Thus, Z has the probability density function:
f(z) =
1

2
e

1
2
(z
2
)
on (, )
The graph of y = f(z) is shown in Figure 1. (Notice: In spite of appear-
ances in this picture, the tails of f(z) do not actually touch the z-axis, but
continue out to innity, getting closer and closer to the axis.) We see that
the graph is a bell-shaped curve which is symmetric about the vertical axis
(i.e., symmetric about the line z = , and for Z we have = 0). As for any
pdf, the probability that the observed value of Z lies in the interval [c, d] is
given by the area under the curve between z = c and z = d. For instance, the
highlighted region of the graph in Figure 1 corresponds to prob{0 Z z}
for some value z.
0 Z
Figure 1: The pdf of the standard normal random variable Z, with area
corresponding to prob{0 Z z} shaded.
14
Looking at the functional expression for the pdf of Z quickly shows that
these areas involve an integral which we do not know how to evaluate. That
is, we have
prob{c Z d} =
_
d
c
1

2
e

1
2
(z
2
)
dz
and we nd that neither substitution nor integration by parts is helpful in
evaluating this.
Instead of evaluating this integral, we use a table to look up probabilities
for the standard normal random variable Z. The values of prob{0 Z z}
for z-values correct to 2 decimal places have been tabulated. We use these
values, and also the facts that the pdf is symmetric about the mean and that
therefore half of the area lies to the left of the mean (and half to the right)
to calculate any probabilities involving Z that we need. The table of these
probabilities (areas) is given on the next page (and also on p. 547 of the text).
The table gives values for prob{0 Z a.bc}, where a.bc represents a
number with 2 decimal places, from 0.00 to 3.09. The table has rows for a.b
and columns for 0.0c. That is, to nd Prob{0 Z a.bc} from the table,
we go to the row corresponding to the integer-part and rst decimal place,
and then nd the column which corresponds to the second decimal place.
The table entry in that row and column is prob{0 Z a.bc}.
Example 13. Find prob{0 Z 1.51}.
Solution: In the table, nd the row Z = 1.5 and look in the column 0.01
(because we have a.bc = 1.51 in this case).
Z 0.00 0.01 0.02 0.03
.
.
.
1.4 0.4192 0.4207 0.4222 0.4236
1.5 0.4332 0.4345 0.4357 0.4370
1.6 0.4452 0.4463 0.4474 0.4484
.
.
.
We see that prob{0 Z 1.51} = 0.4345.
15
The Standard Normal Random Variable, Values Give prob{0 Z z}
z 0.00 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09
0.0 0.0000 0.0040 0.0080 0.0120 0.0160 0.0199 0.0239 0.0279 0.0319 0.0359
0.1 0.0398 0.0438 0.0478 0.0517 0.0557 0.0596 0.0636 0.0675 0.0714 0.0753
0.2 0.0793 0.0832 0.0871 0.0910 0.0948 0.0987 0.1026 0.1064 0.1103 0.1141
0.3 0.1179 0.1217 0.1255 0.1293 0.1331 0.1368 0.1406 0.1443 0.1480 0.1517
0.4 0.1554 0.1591 0.1628 0.1664 0.1700 0.1736 0.1772 0.1808 0.1844 0.1879
0.5 0.1915 0.1950 0.1985 0.2019 0.2054 0.2088 0.2123 0.2157 0.2190 0.2224
0.6 0.2257 0.2291 0.2324 0.2357 0.2389 0.2422 0.2454 0.2486 0.2517 0.2549
0.7 0.2580 0.2611 0.2642 0.2673 0.2704 0.2734 0.2764 0.2794 0.2823 0.2852
0.8 0.2881 0.2910 0.2939 0.2967 0.2995 0.3023 0.3051 0.3078 0.3106 0.3133
0.9 0.3159 0.3186 0.3212 0.3238 0.3264 0.3289 0.3315 0.3340 0.3365 0.3389
1.0 0.3413 0.3438 0.3461 0.3485 0.3508 0.3531 0.3554 0.3577 0.3599 0.3621
1.1 0.3643 0.3665 0.3686 0.3708 0.3729 0.3749 0.3770 0.3790 0.3810 0.3830
1.2 0.3849 0.3869 0.3888 0.3907 0.3925 0.3944 0.3962 0.3980 0.3997 0.4015
1.3 0.4032 0.4049 0.4066 0.4082 0.4099 0.4115 0.4131 0.4147 0.4162 0.4177
1.4 0.4192 0.4207 0.4222 0.4236 0.4251 0.4265 0.4279 0.4292 0.4306 0.4319
1.5 0.4332 0.4345 0.4357 0.4370 0.4382 0.4394 0.4406 0.4418 0.4429 0.4441
1.6 0.4452 0.4463 0.4474 0.4484 0.4495 0.4505 0.4515 0.4525 0.4535 0.4545
1.7 0.4554 0.4564 0.4573 0.4582 0.4591 0.4599 0.4608 0.4616 0.4625 0.4633
1.8 0.4641 0.4649 0.4656 0.4664 0.4671 0.4678 0.4686 0.4693 0.4699 0.4706
1.9 0.4713 0.4719 0.4726 0.4732 0.4738 0.4744 0.4750 0.4756 0.4761 0.4767
2.0 0.4772 0.4778 0.4783 0.4788 0.4793 0.4798 0.4803 0.4808 0.4812 0.4817
2.1 0.4821 0.4826 0.4830 0.4834 0.4838 0.4842 0.4846 0.4850 0.4854 0.4857
2.2 0.4861 0.4864 0.4868 0.4871 0.4875 0.4878 0.4881 0.4884 0.4887 0.4890
2.3 0.4893 0.4896 0.4898 0.4901 0.4904 0.4906 0.4909 0.4911 0.4913 0.4916
2.4 0.4918 0.4920 0.4922 0.4925 0.4927 0.4929 0.4931 0.4932 0.4934 0.4936
2.5 0.4938 0.4940 0.4941 0.4943 0.4945 0.4946 0.4948 0.4949 0.4951 0.4952
2.6 0.4953 0.4955 0.4956 0.4957 0.4959 0.4960 0.4961 0.4962 0.4963 0.4964
2.7 0.4965 0.4966 0.4967 0.4968 0.4969 0.4970 0.4971 0.4972 0.4973 0.4974
2.8 0.4974 0.4975 0.4976 0.4977 0.4977 0.4978 0.4979 0.4979 0.4980 0.4981
2.9 0.4981 0.4982 0.4982 0.4983 0.4984 0.4984 0.4985 0.4985 0.4986 0.4986
3.0 0.4987 0.4987 0.4987 0.4988 0.4988 0.4989 0.4989 0.4989 0.4990 0.4990
For z = 4.0, 5.0 and 6.0, the probabilities are 0.49997, 0.4999997, and 0.499999999.
16
To nd other probabilities involving Z, we use our understanding of the
symmetry properties of the graph of the pdf and other basic properties of
areas. For instance, for any c > 0 and d > c, prob{c Z d} corresponds to
the area under f(z) from z = c to z = d. But this area can also be expressed
as (the area under f(z) from z = 0 to z = d) minus (the area under f(z)
from z = 0 to z = c). That is, we have
prob{c Z d} = prob{0 Z d} prob{0 Z c}
Also, because the graph is symmetric about z = 0,
prob{a Z 0} = prob{0 Z a}
for any positive value a. That is, the area under the curve from z = 0 to a
z-value a units away from 0 is the same, whether we go a units to the right
or a units to the left. And since the total area under the pdf is 1, we see that
prob{Z 0} = prob{Z 0} = 0.5. (Notice that in the table, as the z-value
gets large, the probability prob{0 Z z} approaches .5 .)
With these simple relationships we can nd any probabilities involving Z
which we need, using the tabulated values of prob{0 Z z}.
Example 14. If Z is the standard normal random variable, nd the following:
(a) prob{1 Z 2} (b) prob{Z 0.5}
(c) prob{0.32 Z 1.06} (d) prob{Z > 2.19}
Solution:
(a) For prob{1 Z 2}, we are looking for the probability of the event
{1 Z 2}. As noted above, we have
prob{1 Z 2} = prob{0 Z 2} prob{0 Z 1}
On the right hand side of this equation we just have 2 probabilities which we
can look up in the Z-table. We look up the table entries in row 2.0, column
0,00 and row 1.0, column 0.00 to see that
prob{1 Z 2} = .4772 .3413 = .1359
(b) To nd prob{Z 0.5}, we use the symmetry of the Z-distribution to
see that prob{Z 0.5} = prob{Z 0.5}. (That is, the area under the
curve to the right of z = 0.5 is the same as the area under the curve to
the left of z = 0.5.) And of course {Z 0.5} = {Z < 0} {0 Z 0.5}
17
so that prob{Z 0.5} = prob{Z < 0} + prob{0 Z 0.5}. Using the fact
that prob{Z 0} = .5, and looking up prob{0 Z 0.5} in the table (row
0.5, column 0.00), we get
prob{Z 0.5} = prob{Z 0.5}
= prob{Z 0} + prob{0 Z 0.5}
= .5 + .1915 = .6915
Notice: In general, for any a > 0, we have
prob{Z a} = prob{Z a} = .5 + prob{0 Z a}
(c) To nd prob{0.32 Z 1.06}, we need to realize that
{0.32 Z 1.06} = {0.32 Z 0} {0 Z 1.06}
so that
prob{0.32 Z 1.06} = prob{0.32 Z 0} + prob{0 Z 1.06}
Of course, we know that prob{0.32 Z 0} = prob{0 Z 0.32}, so
we have
prob{0.32 Z 1.06} = prob{0.32 Z 0} + prob{0 Z 1.06}
= prob{0 Z 0.32} + prob{0 Z 1.06}
= .1255 +.3554 = .4809
Notice: In general, for a > 0 and b > 0, we have
prob{a Z b} = prob{0 Z a} + prob{0 Z b}
(d) One way that we can nd prob{Z > 2.19} is to nd the probability of
the complementary event. We have
{Z > 2.19}
C
= {Z 2.19}
and of course for any event A, prob(A
c
) = 1 prob(A), so we have
prob{Z > 2.19} = 1 prob{Z 2.19}
= 1 (prob{Z < 0} + prob{0 Z 2.19})
= 1 (.5 + .4857) = 1 .9857 = .0143
18
Alternatively, we could nd this by recognizing that
prob{Z > 2.19} = prob{Z 0} prob{0 Z 2.19}
so we have prob{Z > 2.19} = .5 .4857 = .0143.
We can also nd probabilities involving other normal random variables,
besides the standard normal random variable, by converting values of X,
a normal R.V. with mean and standard deviation , to values of Z, the
standard normal R.V., using the relationship below.
Theorem 12.9. Given a normal random variable X with mean and stan-
dard deviation , we can standardize values of X using the formula
Z =
X

How do we use this formula? Suppose we have some normal random vari-
able X, with mean known to be some value and standard deviation known
to be some value , and we want to nd prob{a X b}.
Consider the event {a X b}. (Recall: is always positive)
a X b a X b (subtract from each part)

(divide through by )
But we know that
X

= Z, so we see that the event {a X b} is

equivalent to the event {
a

Z
b

}, and so we have:
Theorem 12.10. If X is a normal random variable with mean and stan-
dard deviation , then for any values a and b, with a b,
prob{a X b} = prob
_
a

Z
b

_
Example 15. Find prob{2.3 X 6.1} where X is a normal random vari-
able with = 5 and = 2.
19
Solution:
prob{2.3 X 6.1} = prob
_
2.3

6.1

_
= prob
_
2.3 5
2
Z
6.1 5
2
_
= prob
_

2.7
2
Z
1.1
2
_
= prob{1.35 Z 0.55}
= prob{1.35 Z 0} + prob{0 Z 0.55}
= prob{0 Z 1.35} + prob{0 Z 0.55}
= .4115 +.2088 = .6203
Example 16. The actual weight of a 180 gram bag of a certain brand of potato
chips is normally distributed, with mean = 188 grams and standard devia-
tion = 5 grams. What is the probability that a bag contains less than 180
grams of chips?
Solution: Let X be the weight of a bag of chips. Then X is a normal random
variable with = 188 and = 5. We need to nd prob{X < 180}. Of
course, since prob{X = 180} = 0, then this is the same as prob{X 180}.
(Recall: for any continuous random variable, the event {X < x} is identical
to the event {X x} because prob{X = x} is innitesimally small, i.e. is 0.)
So we have
prob{X < 180} = prob
_
X

180

_
= prob
_
Z
180 188
5
_
= prob
_
Z
8
5
_
= prob{Z 1.60}
= prob{Z 1.60} = prob{Z 0} prob{0 Z 1.60}
= .5 .4452 = .0548
We see that almost 5.5% of the bags contain less than the stated weight.
20