Professional Documents
Culture Documents
1
After studying, you should be able to
carry out a one & two sample t-test;
identify the features that influence the power
of a test.
2
One Sample Mean Test
Two (Independent) Sample Mean Test
Paired (Dependent) T-Test
3
TEST STATISTIC
HYPOTHESIS
TESTING
T-test
(sample
mean)
5
Decision Types of Error
H0 is true H0 is false
6
t- distribution
Symmetrical about the mean
Degree of freedom
More variable than standardized normal
distribution
n < 30
test the difference between means
7
(X – μ)
(X–μ)√n
t = ------------
or t = ----------------
s
s
√n
df = n – 1
n < 30
80 75 72 74 89 70 65 73
67 82 100 68 66 79 89 78
10
H0: The patient’s pulse rate is within or above the
accepted norm.
H1: The patient’s pulse rate is below the accepted norm.
α = 0.05 (X–μ)√n
t = ----------------
One-sample t-test: SPSS s
16 20 18 16 15 17 21 16 24 15
13 15 12 14 19 20 13 17 14 21
α = 0.05 ( 16.8 – 19 ) √ 20
t = ------------------- = - 3.08
One-sample t-test: SPSS 3.19
14
Compare the means of two independent
groups
Data are normally distributed
Data are expressed in interval and ratio.
N < 30
15
(X1 – X2)
t = -------------------------
1 1
s12 (n1 - 1) + s22 (n2 - 1)
Sp ------ + -------
n1 n2 Sp = --------------------------------
n1 + n2 - 2
or
(X1 – X2)
t =----------------------------------------------------------
s12 (n1 - 1) + s22 (n2 - 1) 1 1
---------------------------- ------ + -------
n1 + n 2 - 2 n1 n2
Where:
X1 = mean of one sample df = n1 + n2 - 2
X2 = mean of the other sample
S12 = sample variance of one sample
S22 = sample variance of the other sample
n1 = size of the sample
n2 = size of the other sample 16
Two (Independent) Sample Mean Test
247 230 220 245 250 256 243 227 220 237
220 236 240 229 239 248 230 240 255 229
246 228 260 256 218 219 237 245 240 222
236 248 230 205 235 210 207 246 239 234
247 225 239 240 236
Cryoprecipitate
Packed Red Blood Cell
17
H0: There is no significant difference between the mean
content of packed RBC from two blood bank.
H1: There is a significant difference between the mean
content of packed RBC from two blood bank.
α = 0.05
Two-sample t-test: SPSS
19
H0: There is no significant difference between the
mean cholesterol level of SV and SD ward.
H1: There is a significant difference between the
mean cholesterol level of SV and SD ward.
α = 0.05
Two-sample t-test: SPSS
Critical region: approximately 2.101 & t-test: 0.447
p-value: 0.660 > 0.05
Decision: Accept null hypothesis
21
1.0 1.5
9.8 12.0 Two groups of experimental
13.2 7.4 rats were injected with a
11.2 9.8 tranquilizer at 1.0 mg and 1.5
9.5 11.5 mg dose respectively. The
13.0 13.0 time given in seconds that
took them to fall asleep is
12.1 12.5
hereby given. Use 0.01
9.8 9.8
significance to test the null
12.3 10.5
hypothesis that the difference
7.9 13.5
in dosage has no effect on the
10.2 length of time it took them to
9.7 fall asleep.
samples are drawn from the same population
difference exists between the before and after
means.
used in experimental design to test the
effectiveness of a certain technique or method
or program that had developed.
d
t = ------------ ∑d2 - (∑d)2 / n
sd / √ n sd = ---------------------
-
n-1
or
D = the mean
d(√n) difference between
t = ------------------------ the pretest and the
2 2
∑d - (∑d) / n posttest.
----------------------
n–1
D = ∑d/n
df = n - 1
24
Paired (Dependent) t- tests
Wt. in lb Wt. in lb
before after
A random sample of 1 160 156
10 female adults were
taken, to test the 2 144 142
effectiveness of a 3 137 134
reducing pill. Their 4 152 150
weight before taking
the pill and the weight 5 159 154
10 days after the pill 6 158 153
had been taken were
recorded as follows: 7 149 146
8 164 159
9 147 145
10 156 152
25
H0: There is no significant difference between the mean
weight before & after taking the reducing pills.
H1: There is a significant difference between the mean
weight before & after taking the reducing pills.
α = 0.05
Paired t-test: SPSS
Critical region: approximately 2.262 & t-test: 8.720
p-value: 0.000 < 0.05
Decision: Do not accept null hypothesis.