Testing Two Related Means
Uploaded by
Sylvia CheungTesting Two Related Means
Uploaded by
Sylvia CheungHypothesis Testing :
Two Related Means
Comparing Two Related Means
Before-after study :
blood pressure of a patient before/after intake of a drug
performance of staff before/after receiving training
Name
Before
After
A
B
C
D
E
F
210
185
215
198
187
225
196
192
204
193
181
233
Comparing Two Related Means
Repeated measure study :
blood pressure of a patient measured several times in a day
Name
A
B
C
D
:
Y
Z
Blood
Blood
Pressure (am) Pressure (pm)
135
136
121
121
:
116
117
126
129
124
119
:
123
120
Comparing Two Related Means
Experiment with natural pairing :
comparing the heights of husband and wife
comparing the IQ of twins
Pair
Husband
Wife
1
2
3
4
5
6
2.1
1.8
1.5
1.6
1.8
1.7
1.9
1.9
1.4
1.3
1.8
2.3
Notations & Assumptions
Pair
1st sample
2nd sample
Difference
1
2
:
n
X11
X12
:
X1n
X21
X22
:
X2n
D1 = X11 X21
D2 = X12 X22
:
Dn = X1n X2n
The two random samples come from normal populations
The differences form a random sample from a
normal population with mean D and variance
D2
Comparing Two Related Means
Hypotheses :
H0: D
H1: D< (lower-tail test)
H0: D
H1: D > (upper-tail test)
H0: D =
H1: D (two-tail test)
D is estimated by D where
n
D (X
i1
i1
1i
X 2i )
X1 X 2
Paired Z Test
Used when :
populations normal
D known
Test statistic :
D
Z
D n
Paired t Test
Used when :
populations normal
D unknown
Test statistic :
D
t
SD
n
where
SD
which has a t distribution with (n-1) df
2
(
D
D
)
i
i 1
n 1
Paired t Test : Example
Is the new computer system faster at the 0.05 level ? You collect
the processing times of 10 jobs, assumed coming from normal
populations :
User Old System (1)
C.B.
9.98 s
T.F.
9.88
M.H.
9.84
R.K.
9.99
M.O.
9.94
D.S.
9.84
S.S.
9.86
C.T.
10.12
K.T.
9.90
S.Z.
9.91
New System (2)
9.88 s
9.86
9.75
9.80
9.87
9.84
9.87
9.98
9.83
9.86
Difference
0.10
0.02
0.09
0.19
0.07
0.00
- 0.01
0.14
0.07
0.05
Paired t Test : Example
(continued)
Step 1: Define hypotheses
H 0 : D 0
(new system is not faster)
H1 : D > 0
(new system is faster)
Paired t Test : Example
(continued)
Step 2: Determine the rejection region
- Significance level is 0.05
- Critical Value: t , n 1 t0.05, 10-1 1.8331
- Reject H0 if t 1.8331
Rejection
Region
t0.05,101 1.8331
Paired t Test : Example
(continued)
Step 3: Compute test statistic
D
SD
0.072
D D
i
n 1
0.06215
Test Statistic
D0
0.072 0
t
3.66
S D n 0.06215 10
Paired t Test : Example
Step 4: Make decision
(continued)
Since t = 3.66 is in
the rejection region,
we reject the null
hypothesis.
There is sufficient
evidence that the
new system is faster
than the existing
system
Rejection
region
-4 -3.5 -3 -2.5 -2 -1.5 -1 -0.5
0.5
1.5
2.5
t0.05,101 1.8331
3.5
t 3.66
Test Statistic 3.66 is in the Rejection Region
Paired t Test Using EXCEL
Paired t Test Using EXCEL
Paired t Test Using SPSS
Paired t Test Using SPSS
Paired t Test Using SPSS
Relation Between S D , S1 , S2 and
2
(
D
D
)
S D2 i
n 1
i 1
n
(( X 1i X 1 ) ( X 2i X 2 )) 2
=
n 1
i 1
n
n
( X 1i X 1 ) 2 n ( X 2i X 2 ) 2
( X 1i X 1 )( X 2i X 2 )
2
n 1
n 1
n 1
i 1
i 1
i 1
n
S12 S 22 2 S12
S12 S 22 2rS1S 2
