Вы находитесь на странице: 1из 3

Class One:

Non-Parametric test:
Data is not normally distributed
There is no assumption required about mean and variance of the population.
T-test and F-test:
A t-test compares means between samples, while an f-test compares variances
between samples.
T-Test:
1. One sample t-test
2. Two sample t-test
One sample t-test:

Chapter 10: Question No 46.


Stata:
One sample t-test:
Data:
y

27

25

26

25

28

28

27

24

26

25

29

25

27

24

Statistics=>summaries, tables, test=>classical test=>one sample test=>select


variable=>put hypothesized mean=>set confidence level=>ok
. ttest y == 25
One-sample t test
Variable
Obs
Mean Std. Err.
Std. Dev.
[95% Conf. Interval]
y
14
26.14286
.4174378
1.561909
25.24104 27.04468
mean = mean(y)
t = 2.7378
Ho: mean = 25
degrees
of freedom =
13
Ha: mean < 25
Ha: mean !=25
Ha: mean > 25
Pr(T < t) = 0.9915 Pr(T > t) =0.0169
Pr(T > t) = 0.0085
If the problem is about more than or less than then the first one is to be considered is
alternative hypothesis must be written. However, null hypothesis must be written first
is there is no difference.

Generally if t calculated is 2 or more than 2 then the null hypothesis is rejected. Here
therefore the null hypothesis is rejected.
Two sample test| Chapter 11, Question No. 36
Little river
Mirreells

31.7
3
22.9
3

28.7
7
23.9
2

29.5
3
26.9
2

22.0
8
27.2

29.4
7
26.4
4

18.6
25.6
2

32.9
2
30.6
1

25.1
8
29.4
4

29.8
2
23.0
9

26.4
9
23.0
1

26.6
9

22.3
1

Stata: Statistics=>summaries, tables, test=>classical test=> two samples mean


comparison test=>select variables=>put hypothesized mean=>set confidence
level=>ok
. ttest littleriver == mirreells, unpaired
Two-sample t test with equal variances
Variable
Obs
Mean
Std. Err.
Std. Dev.
[95% Conf. Interval]
little~r
10
27.459
1.403323
4.437698
24.28446
30.63354
mirree~s
12
25.68167
.777261
2.692511
23.97093
27.39241
combined
22
26.48955
.7704143
3.613564
24.88738
28.09171
diff
1.777333 1.534825 -1.424257 4.978922
diff = mean(littleriver) - mean(mirreells) t = 1.1580
Ho: diff = 0
degrees of freedom = 20
Ha: diff < 0
Ha: diff != 0
Ha: diff > 0
Pr(T < t) = 0.8697
Pr(T > t) = 0.2605
Pr(T > t) = 0.1303

So the null hypothesis is accepted and the alternative hypothesis is rejected.

Paired test:
Stata: Statistics=>summaries, tables, test=>classical test=> two samples mean
comparison test=>select variables=>put hypothesized mean=>set confidence
level=>ok
ttest littleriver == mirreells, unpaired
Two-sample t test with equal variances
-----------------------------------------------------------------------------Variable |
Obs
Mean Std. Err. Std. Dev. [95% Conf. Interval]
---------+-------------------------------------------------------------------little~r |
10
27.459 1.403323 4.437698 24.28446 30.63354
mirree~s |
12 25.68167
.777261 2.692511 23.97093 27.39241
---------+-------------------------------------------------------------------combined |
22 26.48955 .7704143 3.613564 24.88738 28.09171
---------+-------------------------------------------------------------------diff |
1.777333 1.534825
-1.424257 4.978922
-----------------------------------------------------------------------------diff = mean(littleriver) - mean(mirreells)
t = 1.1580
Ho: diff = 0
degrees of freedom =
20
Ha: diff < 0
Pr(T < t) = 0.8697

Ha: diff != 0
Ha: diff > 0
Pr(|T| > |t|) = 0.2605
Pr(T > t) = 0.1303

Variance Test:
Stata: Statistics=>summaries, tables, test=>classical test=> one sample variance
comparison test =>select variables=>put hypothesized standard deviation=> ok
. sdtest y == 5.8
One-sample test of variance
Variable
Obs
Mean
Std. Err.
Std. Dev.
[95% Conf. Interval]
y
14 26.14286
.4174378 1.561909
25.24104 27.04468
sd = sd(y)
c = chi2 = 0.9428
Ho: sd = 5.8
degrees of freedom =
13
Ha: sd < 5.8
Ha: sd != 5.8
Ha: sd > 5.8
Pr(C < c) = 0.0000
2*Pr(C < c) = 0.0000
Pr(C > c) = 1.0000

Вам также может понравиться