Examples of continuous

probability distributions:

The normal and standard

normal
 The Normal Distribution

f(X) Changing μ shifts the

distribution left or
right.
Changing σ
increases or
σ decreases the
spread.

μ X
The Normal Distribution:
as mathematical function
(pdf)

1 x  2
1  (

)
f ( x)  e 2
 2
This is a bell shaped
Note constants: curve with different
=3.14159 centers and spreads
e=2.71828 depending on  and

 The Normal PDF
It’s a probability function, so no matter what the
values of  and , must integrate to 1!
 1 x  2
1  ( )

 2
e 2  dx 1
Normal distribution is
defined by its mean and
standard dev.
 1 x  2
E(X)= = 1  ( )


x  2
e 2  dx

 1 x  2
1  ( )
Var(X)=2 = ( 

x2
 2
e 2  dx)   2

Standard Deviation(X)=
**The beauty of the normal curve:

No matter what  and  are, the area between - and

+ is about 68%; the area between -2 and +2 is
about 95%; and the area between -3 and +3 is
about 99.7%. Almost all values fall within 3 standard
deviations.
68-95-99.7 Rule

68% of
the data

95% of the data

99.7% of the data

68-95-99.7 Rule
in Math terms…
  1 x  2
1  ( )

  
 2
 e 2  dx .68

  2 1 x  2
1  ( )

  
2 2
 e 2  dx .95

 3 1 x  2
1  ( )

  
3 2
 e 2  dx .997
 How good is rule for real data?

Check some example data:

The mean of the weight of the women = 127.8
The standard deviation (SD) = 15.5
 68% of 120 = .68x120 = ~ 82 runners
In fact, 79 runners fall within 1-SD (15.5 lbs) of the mean.

112.3 127.8 143.3

25

20

P
e 15
r
c
e
n 10
t

0
80 90 100 110 120 130 140 150 160
POUNDS
 95% of 120 = .95 x 120 = ~ 114 runners
In fact, 115 runners fall within 2-SD’s of the mean.

96.8 127.8 158.8

25

20

P
e 15
r
c
e
n 10
t

0
80 90 100 110 120 130 140 150 160
POUNDS
 99.7% of 120 = .997 x 120 = 119.6 runners
In fact, all 120 runners fall within 3-SD’s of the
mean.

81.3 127.8 174.3

25

20

P
e 15
r
c
e
n 10
t

0
80 90 100 110 120 130 140 150 160
POUNDS
Example
 Suppose SAT scores roughly follows a normal
distribution in the U.S. population of college-
bound students (with range restricted to 200-
800), and the average math SAT is 500 with a
standard deviation of 50, then:
 68% of students will have scores between 450 and
550
 95% will be between 400 and 600
 99.7% will be between 350 and 650
 Example
 BUT…
 What if you wanted to know the math SAT

score corresponding to the 90th percentile

(=90% of students are lower)?
P(X≤Q) = .90 
Q 1 x  500 2
1  ( )
(50)
200
2
e 2 50 dx .90

Solve for Q?….Yikes!

 The Standard Normal (Z):
“Universal Currency”
The formula for the standardized normal
probability density function is

1 Z 0 2 1
1  ( ) 1  ( Z )2
p( Z )  e 2 1
 e 2
(1) 2 2
 The Standard Normal Distribution (Z)
All normal distributions can be converted into
the standard normal curve by subtracting the
mean and dividing by the standard deviation:
X
Z


Somebody calculated all the integrals for the standard

normal and put them in a table! So we never have to
integrate!
Even better, computers now do all the integration.
Comparing X and Z
units

100 200 X ( = 100,  = 50)

0 2.0 Z ( = 0,  = 1)
 Example
 For example: What’s the probability of getting a math SAT score of 575 or less, =500 and =50?

575  500
Z 1.5
50
i.e., A score of 575 is 1.5 standard deviations above the mean

575 1 x  500 2 1.5 1

1  ( ) 1  Z2
 P ( X 575)  (50)
200
2
e 2 50 dx   

2
e 2 dz

Yikes!
But to look up Z= 1.5 in standard normal chart (or enter
into SAS) no problem! = .9332
Practice problem
If birth weights in a population are
normally distributed with a mean of 109
oz and a standard deviation of 13 oz,
a. What is the chance of obtaining a birth
weight of 141 oz or heavier when
sampling birth records at random?
b. What is the chance of obtaining a birth
weight of 120 or lighter?
 Answer
a. What is the chance of obtaining a birth
weight of 141 oz or heavier when
sampling birth records at random?
141  109
Z 2.46
13

From the chart or SAS  Z of 2.46 corresponds to a right tail (greater

than) area of: P(Z≥2.46) = 1-(.9931)= .0069 or .69 %
 Answer
b. What is the chance of obtaining a birth
weight of 120 or lighter?
120  109
Z .85
13

From the chart or SAS  Z of .85 corresponds to a left tail area of:
P(Z≤.85) = .8023= 80.23%
 Looking up probabilities in
the standard normal table
What is the area to
the left of Z=1.51 in
a standard normal
curve?

Area is 93.45%
Z=1.51

Z=1.51
Normal probabilities in
SAS
data _null_;
The “probnorm(Z)” function gives
theArea=probnorm(1.5); you the probability from negative
put theArea; infinity to Z (here 1.5) in a
run; standard normal curve.
0.9331927987

And if you wanted to go the other direction (i.e., from the area to the Z score
(called the so-called “Probit” function 
data _null_;
The “probit(p)” function gives you
theZValue=probit(.93); the Z-value that corresponds to a
put theZValue; left-tail area of p (here .93) from a
run; standard normal curve. The probit
1.4757910282 function is also known as the
inverse standard normal function.
Probit function: the
inverse
(area)= Z: gives the Z-value that goes with the probability you want
For example, recall SAT math scores example. What’s the score that corresponds to the
90th percentile?
In Table, find the Z-value that corresponds to area of .90  Z= 1.28
Or use SAS
data _null_;
theZValue=probit(.90);
put theZValue;
run;
1.2815515655
If Z=1.28, convert back to raw SAT score 
1.28 =
X – 500 =1.28 (50)
X=1.28(50) X+  500
500 = 564 (1.28 standard deviations above the mean!)
50
`
Are my data “normal”?
 Not all continuous random
variables are normally distributed!!
 It is important to evaluate how well
the data are approximated by a
normal distribution
 Are my data normally
distributed?
1. Look at the histogram! Does it appear bell
shaped?
2. Compute descriptive summary measures—
are mean, median, and mode similar?
3. Do 2/3 of observations lie within 1 std dev of
the mean? Do 95% of observations lie within
2 std dev of the mean?
4. Look at a normal probability plot—is it
approximately linear?
5. Run tests of normality (such as Kolmogorov-
Smirnov). But, be cautious, highly influenced
by sample size!
Data from our class…

Median = 6
Mean = 7.1
Mode = 0

SD = 6.8
Range = 0 to 24
(= 3.5 σ)
Data from our class…

Median = 5
Mean = 5.4
Mode = none

SD = 1.8
Range = 2 to 9
(~ 4 σ
Data from our class…

Median = 3
Mean = 3.4
Mode = 3

SD = 2.5
Range = 0 to 12
(~ 5 σ
Data from our class…

Median = 7:00
Mean = 7:04
Mode = 7:00

SD = :55
Range = 5:30 to
9:00
(~4 σ
 Data from our class…

7.1 +/- 6.8 =

0.3 13.9 0.3 – 13.9

Data from our class…

7.1 +/- 2*6.8

=
0 – 20.7
Data from our class…

7.1 +/- 3*6.8

=
0 – 27.5
Data from our class…

5.4 +/- 1.8 =

3.6 – 7.2

3.6 7.2
 Data from our class…

5.4 +/- 2*1.8

=
1.8 – 9.0
1.8 9.0
 Data from our class…

5.4 +/- 3*1.8

=
0– 10
0 10
 Data from our class…

0.9 5.9 3.4 +/- 2.5=

0.9 – 7.9
 Data from our class…

0 8.4 3.4 +/-

2*2.5=
0 – 8.4
 Data from our class…

0 10.9 3.4 +/-

3*2.5=
0 – 10.9
Data from our class…

6:09 7:59 7:04+/- 0:55

=
6:09 – 7:59
 Data from our class…

5:14
8:54 7:04+/-
2*0:55 =
5:14 – 8:54
 Data from our class…

4:19
9:49 7:04+/-
2*0:55 =
4:19 – 9:49
The Normal Probability
Plot
 Normal probability plot

Order the data.

Find corresponding standardized normal
quantileth values: i
i quantile  ( )
n 1
where  is the probit function, which gives the Z value
that corresponds to a particular left - tail area


Plot the observed data values against normal
quantile values.

Evaluate the plot for evidence of linearity.
Normal probability plot
coffee…

Right-Skewed!
(concave up)
Normal probability plot
love of writing…

Neither right-
skewed or left-
skewed, but big
gap at 6.
Norm prob. plot Exercise…

Right-Skewed!
(concave up)
Norm prob. plot Wake up
time

Closest to a
straight line…
Formal tests for normality
 Results:
 Coffee: Strong evidence of non-
normality (p<.01)
 Writing love: Moderate evidence of non-
normality (p=.01)
 Exercise: Weak to no evidence of non-
normality (p>.10)
 Wakeup time: No evidence of non-
normality (p>.25)
Normal approximation to
the binomial
When you have a binomial distribution where n
is large and p is middle-of-the road (not too
small, not too big, closer to .5), then the
binomial starts to look like a normal
distribution in fact, this doesn’t even take a
particularly large n

Recall: What is the probability of being a smoker

among a group of cases with lung cancer is .6,
what’s the probability that in a group of 8
cases you have less than 2 smokers?
Normal approximation to
the binomial
When you have a binomial distribution
where n is large and p isn’t too small
(rule of thumb: mean>5), then the
binomial starts to look like a normal
distribution

Recall: smoking
.27
example… Starting to have a normal
shape even with fairly small
n. You can imagine that if n
got larger, the bars would get
thinner and thinner and this
would look more and more
0 1 2 3 4 5 6 7 8 like a continuous function,
with a bell curve shape. Here
np=4.8.
 Normal approximation to
binomial
.27

0 1 2 3 4 5 6 7 8

What is the probability of fewer than 2 smokers?

Exact binomial probability (from before) = .00065 + .008 = .00865

Normal approximation probability:

=4.8
=1.39
2  (4.8)  2.8
Z   2 P(Z<2)=.022
1.39 1.39
A little off, but in the right ballpark… we could also use the
value to the left of 1.5 (as we really wanted to know less than but
not including 2; called the “continuity correction”)…

1.5  (4.8)  3.3

Z   2.37
1.39 1.39
A fairly good approximation of
P(Z≤-2.37) =.0069
the exact probability, .00865.
Practice problem
1. You are performing a cohort study. If the
probability of developing disease in the
exposed group is .25 for the study duration,
then if you sample (randomly) 500 exposed
people, What’s the probability that at most
120 people develop the disease?
 Answer
By hand (yikes!):
P(X≤120) = P(X=0) + P(X=1) + P(X=2) + P(X=3) + P(X=4)+….+ P(X=120)=
 500   500   500   500 
120
  (.25) (.75)
 120 
380
+
2
  (.25) (.75)
 2
498
+ 1
  (.25) (.75)
 1
499
+
0
  (.25) (.75)
 0
500
…

OR Use SAS:
data _null_;
Cohort=cdf('binomial', 120, .25, 500);
put Cohort;
run;
0.323504227

OR use, normal approximation:

=np=500(.25)=125 and 2=np(1-p)=93.75; =9.68
120  125
Z  .52 P(Z<-.52)= .3015
9.68
Proportions…
 The binomial distribution forms the basis
of statistics for proportions.
 A proportion is just a binomial count
divided by n.

For example, if we sample 200 cases and find
60 smokers, X=60 but the observed
proportion=.30.
 Statistics for proportions are similar to
binomial counts, but differ by a factor of
n.
 Stats for proportions
For binomial:  x np Differs by
2 a factor
 x np(1  p) of n.

 x  np(1  p)
Differs
by a
factor
 pˆ  p of n.
For proportion:
np(1  p ) p (1  p )
 pˆ 2  2

n n
P-hat stands for “sample p (1  p )
proportion.”  pˆ 
n
It all comes back to Z…
 Statistics for proportions are based
on a normal distribution, because
the binomial can be approximated
as normal if np>5