Most of the hypothesis-testing and confidence interval

procedures discussed in previous chapters are based on the

assumption that we are working with random samples from
normal populations.

These procedures are often called parametric methods

In this chapter, nonparametric and distribution free methods will

be discussed.

We usually make no assumptions about the distribution of the

underlying population.
 9.9.1 The Sign Test

The sign test is used to test hypotheses about the median of

a continuous distribution.

Let R+ represent the number of differences

X ~
i 0
that are positive.
 Null Hypothesis
H 0 :
=
0

One-
One-Sided Test

Alternate
Hypothesis H1 :
<
0 H1 :
>
0

1 1
P-Value P R + r + when p = P R + r + when p =
2 2
 Two-
Two-Sided Test:
Null Hypothesis:
H 0 :
=
0
Alternate Hypothesis:
H1 :

0

+ n + n
If r < , If r > ,
2 2
1 1
P = 2 P R + r + when p = P = 2 P R + r + when p =
2 2
Example 1

Montgomery, Peck, and Vining (2006) report on a study in which a rocket

motor is formed by binding an igniter propellant and a sustainer propellant
together inside a metal housing. The shear strength of the bond between the
two propellant types is an important characteristic. The results of testing 20
randomly selected motors are shown in Table 1. We would like to test the
hypothesis that the median shear strength is 13790 kN/m2, using = 0.05.
Solution
Parameter of Interest:
Interest:
The parameter of interest is the median of the distribution of
propellant shear strength.
Null hypothesis, H0:
H 0 :
= 13790 k N m 2
Alternative hypothesis, H1:
H1 :
13790 k N m 2
Test statistic:
statistic:
The test statistic is the observed number of plus differences in Table 1,
or r+ = 14.
Reject H0 if:
if:
We will reject H0 if the P-value corresponding to r+ = 14 is less than or
equal to = 0.05.
Computations:
Computations:
n 20
Since r+ = 14 is greater than = = 10, we calculate P-value from
2 2
1
P = 2 P R + when p =
2
20
20
= 2 ( 0.5 ) ( 0.5 )
r 20 r
= 0.1153
r =14 r
 Table 1 Propellant Shear Strength Data
Observation, Shear Strength, Differences, Observation, Shear Strength,
Strength, Differences,
Sign Sign
i xi xi - 13790 i xi xi - 13790
1 14883.71 +1093.71 + 11 14928.53 +1138.53 +
2 11570.44 -2219.56 - 12 16544.31 +2754.31 +
3 15968.26 +2178.26 + 13 12271.29 -1518.71 -
4 14212.16 +422.16 + 14 16111.32 +2321.32 +
5 15220.18 +1430.18 + 15 12171.31 -1618.69 -
6 11778.31 -2011.69 - 16 14158.38 +368.38 +
7 12305.07 -1484.93 - 17 16646.70 +2856.7 +
8 17754.69 +3964.69 + 18 15171.91 +1381.91 +
9 16257.15 +2467.15 + 19 18300.06 +4510.06 +
10 15559.40 +1769.4 + 20 12091.33 -1698.67 -
 Table 1 Propellant Shear Strength Data
Observation, Shear Strength, Differences, Observation, Shear Strength,
Strength, Differences,
Sign Sign
i xi xi - 13790 i xi xi - 13790
1 14883.71 +1093.71 + 11 14928.53 +1138.53 +
2 11570.44 -2219.56 - 12 16544.31 +2754.31 +
3 15968.26 +2178.26 + 13 12271.29 -1518.71 -
4 14212.16 +422.16 + 14 16111.32 +2321.32 +
5 15220.18 +1430.18 + 15 12171.31 -1618.69 -
6 11778.31 -2011.69 - 16 14158.38 +368.38 +
Conclusion:
7 12305.07 -1484.93 - 17 16646.70 +2856.7 +
8 Since P17754.69
= 0.1153 is not less than
+3964.69 + = 0.05,
18 we cannot reject H+1381.91
15171.91 0 that is the +

9
median16257.15
shear strength
+2467.15
kN/m2. 19
is 13790+kN/m Another way to say is +4510.06
18300.06
that the observed
+
number +
of plus sign+1769.4
r = 14 was+not large20or small12091.33
enough to indicate that -
10 15559.40 -1698.67
median shear strength is different from 13790 kN/mkN/m2 at the = 0.05 level of
significance.
Ties in the Sign Test

We will find a tie when Xi exactly equal to
0 .

When ties occur, they should be set aside and the sign test
applied to the remaining data.
 Normal
Approximation
R + 0.5n
for Sign Test Z0 =
Statistics 0.5 n
Example 2
We will illustrate the normal approximation procedure by applying it to the
problem in Example-1. Recall that the data for this example are in Table 1.
The seven-step procedure follows:

Solution
Parameter of Interest:
Interest:
The parameter of interest is the median of the distribution of
propellant shear strength.
Null hypothesis, H0:
H 0 :
= 13790 k N m 2
Null hypothesis, H1:
H1 :
13790 k N m 2
Test Statistics:
Statistics:
r + 0.5n
z0 =
0.5 n

We reject H0 if:
if:
We will reject H0 if z0 > z0.025 = 1.96 or z0 < z0.025 = 1.96

Computations:
Computations:
14 0.5 ( 20 )
z0 = = 1.789
0.5 20
Conclusions:
Conclusions:
Since 1.96 < 1.789 < 1.96, we failed to reject H0. Thus, our conclusion is
identical to those in Example-1.
 9.9.2 The Wilcoxon Signed-
Signed-Rank Test

The Wilcoxon signed-

signed-rank test applies to the case of symmetric
continuous distributions.

Under this assumption, the mean equals the median.

The null hypothesis is H0: = 0

 We are interested in testing H 0 : = 0 against the usual alternatives.
Assume that X 1 , X 2 , , X n is a random sample from a continuous and
symmetric distribution with mean (and median) .
Steps:

Compute the differences

Rank the absolute differences
X i 0 , i = 1, 2, , n.
in ascending order, and then give the ranks X i 0 , i = 1, 2, , n
the signs of their corresponding differences.

W + ~ sum the positive ranks
W ~ absolute value of the sum of the negative ranks
(
and let W = min W + ,W )
Example 3

We will illustrate the Wilcoxon signed-rank test by applying it to the propellant

shear strength data from Table 1. Assume that the underlying distribution is
a continuous symmetric distribution. The seven-step procedure is applied as
follows:
Parameter of Interest:
Interest:
The parameter of interest is the mean (or median) of the distribution of propellant
shear strength.
Null Hypothesis:
Hypothesis:
H 0 : = 13790 k N m 2

Alternative Hypothesis:
Hypothesis:
H1 : 13790 k N m 2
Test statistic:
statistic:
The test statistics is

w = min ( w+ , w )

Reject H0 if:
if:
*
We will reject H0 if w w0.05 = 52 from Appendix Table IX.

Computations:
Computations:
The sum of the positive ranks is:

w+ = (1 + 2 + 3 + 4 + 5 + 6 + 11 + 13 + 15 + 16 + 17 + 18 + 19 + 20 ) = 150
The sum of the absolute values of the negative ranks is:

w = ( 7 + 8 + 9 + 10 + 12 + 14 ) = 60
Therefore, w = min (150, 60 ) = 60
The signed ranks from Table 1 are shown in the following display:
Differences, Differences,
Observation Signed-
Signed-Rank Observation Signed-
Signed-Rank
xi 13790 xi 13790
16 +368.38 +1 10 -1769.4 -11
4 +422.16 +2 6 -2011.69 -12
1 +1093.71 +3 3 +2178.26 +13
11 +1138.53 +4 2 -2219.56 -14
18 +1381.91 +5 14 +2321.32 +15
5 +1430.18 +6 9 +2467.15 +16
7 -1484.93 -7 12 +2754.31 +17
13 -1518.71 -8 17 +2856.7 +18
15 -1618.69 -9 8 +3964.69 +19
20 -1698.67 -10 19 +4510.06 +20

Conclusions:
Conclusions:
Since w = 60 is not less than or equal to the critical value w0.05 = 52, we cannot
reject the null hypothesis that the mean (or median, since the population is
assumed to be symmetric) shear strength is 13790 kN/m2.
Ties in the Wilcoxon Signed-
Signed-Rank Test

If several observations have the same absolute magnitude, they

are assigned the average of the ranks that they would receive if
they differed slightly from one another.
 Normal
Approximation A test of H 0 : = 0 can be based on the
for Wilcoxon statistics:
Signed-
Signed-Rank W + W +
Statistics Z0 =
W +
where
n ( n + 1)
W + =
4
and
2 n ( n + 1)( 2n + 1)
W + =
24

* The sample size is moderately large, say, n > 20.