HYPOTHESIS TESTING

Definition: Any statement made on the distribution of a population, which is either true or false, is
called a statistical hypothesis.

Remark: Unless one is given with the population, it cannot be determined with certainty whether a
hypothesis is true or false. Thus, in any test of hypothesis there are two contradicting statements. The
objective of the test is to determine which of these two statements is more acceptable as true.

Definition: The statement which is being tested is called the null hypothesis. In the event that the null
hypothesis is rejected, the alternative hypothesis is the one that is accepted in place of the null
hypothesis. The null and alternative hypotheses shall be denoted by Ho and Ha, respectively.

Remark:

1. The null and the alternative hypotheses must be non-overlapping so that whatever outcome of
the test of hypothesis, there will be no confusion so as to the state of the population that has
been accepted in the face of evidence coming from the sample.

2. Due to the fact that the true state of the nature is unknown, one is taking the risk of committing
an error or a wrong decision when conducting a test of hypothesis.

Steps in conducting a test hypothesis

1. State of null and alternative hypothesis

2. Determine the test statistics and the tail in the distribution of the test statistics where the
rejection region is to be located. From the direction of the inequality in the alternative
hypothesis, a test is classified as one-tailed if inequality is one-directional so that the rejection
region is only in one tail of the distribution. Otherwise, the test is two-tailed. That is, there is a
rejection region in each of the two tails of the distribution of the test statistic.

3. For a specified level of significance, state the decision rule. That is specify foe which values of
the test statistic Ho is to be rejected. This should be consistent with the alternative hypothesis.

4. Compute for the value of the test statistic for the available sample data

5. Make a decision in accordance with the decision rule.

6. Draw the conclusion, answering the original problem.

Independent Sample T-Test (Comparing 2 different/uncorrelated groups)
Example: Is there a significant difference in teaching performance of the faculty when grouped
according to gender?

Teaching Performance
MALE (x1) FEMALE (x2) (x1)2 (x2)2
4.35 4.18 18.9225 17.4724
4.06 4.49 16.4836 20.1601
3.93 4.49 15.4449 20.1601
4.22 4.74 17.8084 22.4676
4.00 4.31 16.0000 18.5761
4.39 4.48 19.2721 20.0704
4.19 3.93 17.5561 15.4449
3.67 4.69 13.4689 21.9961
3.67 3.07 13.4689 9.4249
3.62 4.42 13.1044 19.5364
∑x1=40.1 ∑x2=42.8 ∑(x1)2=161.5298 ∑(x2)2=185.309

Ho: There is no significant Male Female

difference between the Compute for Mean
teaching performance when
x=
x x=
x
grouped according to gender. n n
or 40.1 42.8
Ho: µ1=µ2 = =
10 10
Ha: Otherwise
= 4.01 = 4.28
α = 0.05
Critical Region: Compute for Variance
Reject Ho if the n x 2 − ( x ) n x 2 − ( x )
2 2

computed |t|>t (0.05,18)=2.101 s =

2
s =
2

n(n − 1) n(n − 1)
otherwise reserve judgment.
10(161.5298) - (40.1)(40.1) 10(185.309) − (42.8)(42.8)
= =
Test Statistic 10(9) 10(9)
t = −1.516 = 0.080978 = 0.236111
Solve for t – value
Decision / Conclusion: x1 − x 2
t=
Since the absolute value of (n1 − 1) s + (n 2 − 1) s 22
2
1 1
 + 
1
the computed t - value is less n1 + n 2 − 2  n1 n 2 
than 2.101, the null hypothesis
4.01 − 4.28
cannot be rejected. Hence, =
there is no significant 9(0.080978) + 9(0.236111)  1 1
 + 
difference between the 10 + 10 − 2 10 10 
teaching performance when t = −1.516
grouped according to gender.
EXERCISES:

1. In a Mathematics test, 25 grade six children from a city school registered a mean score of 78 and
a standard deviation of 8. In the same test a group of 30 children from the rural area got a mean
of 72 with a standard deviation of 5. Is there a difference between the performance of the two
groups?

2. A personality test was administered to 15 boys and 10 girls in a private school. With the data
given below, test whether the personality of boys and girls significantly differ. Higher scores
indicate better personality traits.
Boys Girls
13 18
16 19
16 22
19 24
20 25
20 25
20 26
21 32
24 36
24 36
26
27
28
30
32
3. The following data are grades of the students from A.M and P.M class. Do their academic
performance vary significantly?

A.M. Class (X1) P.M. Class (X2)

93 90
80 89
80 87
78 86
78 85
71 85
70 84
70 83
66 83
62 81
62 80
57 79
54 76
54 75
53 74
52 73
51 73
50 73
49 71
48 69
Solving Paired-Samples T-test (Comparing paired/correlated data)

Example:
A new teaching strategy is being checked if it is effective by testing the students before and
after conducting the teaching strategy. The data are as follows:
Pretest PostTest di d i2
3 5 -2 4
0 1 -1 1
6 5 1 1
7 7 0 0
4 10 -6 36
3 9 -6 36
2 7 -5 25
1 11 -10 100
4 8 -4 16
∑ di = –33 ∑ di2= 219

Ho: There is no significant Test statistic

difference between the scores in Compute for Mean Difference
the pretest and posttest.
d =
d
or n
Ho: µd=0 − 33
Ha: Otherwise =
9
α = 0.05 = −3.667
Critical Region: Compute for the Variance of Difference
n d 2 − ( d )
Reject Ho if the computed 2

|t|>t (0.05,8)= 2.306 otherwise s =

2

n(n − 1)
d
reserve judgment.
9(219) − (−33)(−33)
=
Test Statistic 9(8)
t = −3.143
s d2 = 12.25
Decision / Conclusion: s d = 12.25
Since the absolute value of s d = 3 .5
the computed t - value is greater Solve for t-value
than 2.306, the null hypothesis is d − d0
rejected. Hence, there is a t=
sd / n
significant difference between
the pretest and posttest scores of − 3.667
=
the students. This means that the 3.5 / 9
new teaching strategy is more effective t = −3.143
than the old one.
EXERCISES:

1. Do scores before and after lectures significantly differ?

Before After Difference

78 78
24 24
64 62
45 48
64 68
52 56
30 25
50 44
64 56
50 40
78 68
22 36
84 68
40 20
90 58
72 32
2. Managers at fitness facility want to determine whether their weight – loss program is effective.
They randomly select 15 people to participate in the program and weigh each subject before
and after participating.

Weight before Weight after Difference

(kg) (kg)
53 54
69 64
50 49
89 72
63 60
71 72
94 81
56 54
50 53
84 81
67 64
53 49
49 50
76 72
87 82
3. Suppose a sample of 20 students were given a diagnostic test before studying a
particular module and then again after completing the module. Do modules improve
students’ knowledge?

Pre-Module Post-Module
Difference
Score Score
18 22
21 25
16 17
22 24
19 16
24 29
17 20
21 23
23 19
18 20
14 15
16 15
16 18
19 26
18 18
20 24
12 18
22 25
15 19
17 16
ONE-WAY Analysis of Variance (ANOVA)
(Comparing 3 or more different groups)

School1 School2 School3

Anova Table
4.35 3.67 4.19
4.06 3.67 3.07 Source of Sum of Squares Degrees Mean f-value
4.18 4.74 3.95 Variation of Square
freedom
4.49 3.62 4.26
Between SSC=∑Ti2/n – T ..2/nk k–1 MSC=SSC
4.49 4.31 3.87 groups k–1 f = MSC
3.93 4.48 4.42 Within SSE = SST – SSC nk – k MSE=SSE MSE
4.22 3.73 3.23 groups nk – k
4.00 3.93 3.98 Total SST = ∑∑x2 – T..2/nk nk – 1
4.39 4.69 3.96
4.19 4.31 3.91
T1=42.3 T2=41.15 T3=38.84 T.. = 122.29

Ho: there is no significant Computations for Anova

difference among the teaching
performance of faculty from 3 SST = 4.352 + 4.062 + 4.182 + . . . + 3.912 – 122.292/10(3)
= 4.3591
different schools
Ha: there is a significant difference SSC = 42.32 + 41.152 + 38.842 _ 122.292
among the teaching performance of 10 30
= 0.6210
faculty from 3 different schools.
α=0.05 SSE = SST – SSC
= 4.3591 – 0.6210
Critical Region: Reject Ho if the = 3.7381
computed F-value>F(0.05,2,27) = 3.35,
otherwise, do not reject Ho.
Test Stat: Anova Table
f = 2.243
Source of Sum of Degrees Mean f-value
Decision/Conclusion: of
Variation Squares Square
freedom
Since the computed f = 2.243 is
Between
less than 3.35, the null hypothesis 0.6210 2 0.310503
groups
2.243
cannot be rejected. Hence there is Within
3.7381 27 0.138448
no significant difference among groups

the teaching performance of the Total 4.3591 29

faculty.
EXERCISES:
1. Compare the grades of the students from different sections.

Section A (X1) Section B (X2) Section C (X3)

98 91 94
94 88 83
94 88 86
89 87 88
89 87 87
84 87 86
80 86 84
79 86 84
76 84 81
76 80 80
2. Compare the grades of the students from different schools.

School 1 School 2 School 3 School 4

97 93 94 95
95 91 83 85
89 91 83 85
86 90 82 75
78 88 82 75
77 85 79 74
74 85 77 73
73 84 77 72
71 83 77 71
68 83 75 71
The following data represent the number of packages of five popular brands of cigarettes sold by
a supermarket on 5 randomly selected days:

BRAND
A B C D E
21 35 45 32 45
35 12 60 53 29
32 27 33 29 31
28 41 36 42 22
14 19 31 40 36

Determine whether or not the five brands sell, on the average, the same number of cigarettes at
the given supermarket.
Pearson-r (Relationship between 2 variables)

Teaching Emotional
x2 y2 xy
Performance (x) Quotient (y)

4.35 85 18.9225 7225 369.75

4.06 72 16.4836 5184 292.32
4.18 75 17.4724 5625 313.5
4.49 90 20.1601 8100 404.1
4.49 88 20.1601 7744 395.12
3.93 70 15.4449 4900 275.1
4.22 78 17.8084 6084 329.16
4.00 71 16 5041 284
4.39 86 19.2721 7396 377.54
4.19 76 17.5561 5776 318.44
∑x = 42.3 ∑y = 791 ∑x2 =179.2802 ∑y2 = 63075 ∑xy = 3359.03

Ho: there is no significant relationship

Solve for r:
between teaching performance and
n xy −  x  y emotional quotient
r=
n x 2 2

− ( x ) n y 2 − ( y )
2
 Ha: Otherwise
10(3359.03) − (42.3)(791)
α=0.05
=
10(179.2802) − (42.3) 10(63075) − (791) 
2 2
Critical Region: Reject Ho if
r = 0.982 r > r (0.025, 8)= 0.632

Test Statistic:
The pearson-r value of 0.982
indicates that there is a very high/strong r=0.982
and positive relationship between the
Teaching performance(x) and Emotional Decision/Conclusion:
quotient(y). A positive relationship Since 0.982>0.632, reject Ho and
indicates that as the Teaching conclude that there is a significant
performance increases, the Emotional relationship between the teaching
quotient also increases. And if the performance and emotional
Teaching performance decreases, the quotient.
Emotional quotient also decreases, and
vice versa.
EXERCISES:
1. Nine students held their breath, once after breathing normally and relaxing for one
minute, and once after hyperventilating for one minute. The table indicates how long (in
sec) they were able to hold their breath. Is there a association between the two
variables?

Normal Hypervent
56 87
56 91
65 85
65 91
50 75
25 28
87 122
44 66
35 58
2. Is there a significant relationship between entrance exam and board exam results.

Entrance Board
Exam Exam
49 75
47 72.2
69 81
65 78.4
49 75
58 72.4
76 65.4
79 75
79 76.2
56 75.8
63 78.8
74 78.4
74 80.8
57 80.2
56 75.4
3. A dietetics student wanted to look at the relationship between calcium intake and knowledge
about calcium in sports science students. Is there a significant relationship between the 2
variables?
Knowledge Score Calcium Intake
10 450
42 1050
38 900
15 525
22 710
32 854
40 800
14 493
26 730
32 894
38 940
25 733
48 985
28 763
22 583
45 850
18 798
24 754
30 805
43 1085