HYPOTHESIS TESTING

T-TEST

The basic assumptions to be established to obtain valid test result using t-test are as follows:
1. Sample size is small (n < 30).
2. Samples, which are randomly selected, come from a normally distributed population.
3. For inference based on a single sample, the population standard deviation (σ) is
unknown.

STEPS IN HYPOTHESIS TESTING:

1. State the hypotheses.
2. Choose a Level of Significance to be used.
3. Decide the test statistics to be used.
4. Compute for the value of the statistics.
5. Make a decision.
a. If the computed value of the test statistics is greater than the critical or tabular
value: Reject Ho and Accept Ha.
b. If the computed value of the test statistics is less than or equal to the critical or
tabular value: Accept Ho and Reject Ha.
6. State the conclusion/implication.

A. T-TEST FOR INDEPENDENT SAMPLES

The t-test is a test of difference between two independent groups. The means are being
compared x̄1 against x̄2

When do we use the t-test for independent samples?

The t-test for independent samples is used when we compare means of two independent
groups.
The distribution is normally distributed, Sk = 0 and Ku = 0.265.
We use the interval or ratio data.
The sample size is less than 30.
Why do we use the t-test for independent sample?
The t-test is used for independent sample because it is more powerful compared with
other tests of difference of two independent groups.

FORMULAS:
A. FOR ONE GROUP OF SAMPLE:
( x̄−μ) √ n
t=
s
Where: x̄ = sample mean
µ = population mean
n = number of samples
s = sample standard deviation
For Degrees of Freedom (df):
df = n - 1

B. FOR TWO GROUPS OF SAMPLES:

x̄ 1− x̄ 2
t=
SS 1+ SS 2 1 1
√( )
( + )
n1 +n2−2 n1 n2

Where: x̄ 1 = mean of group 1

x̄ 2 = mean of group 2
SS1 = the sum of squares of group 1
SS2 = the sum of squares of group 2
n1 = the number of observations in group 1
n2 = the number of observations in group 2
Σ x1
x̄ 1= the mean of group 1
n1
Σ x2
x̄ 2= the mean of group 2
n2
 SS
1=¿ Σ x 12−
(Σx ¿¿1 )2
¿¿ sum of squares of group 1
n1

SS
2=¿ Σ x 22−
( Σx¿¿ 2)2
¿¿ sum of squares of group 2
n2

For Degrees of Freedom (df):

df = n1 + n2 - 2

EXAMPLE 1:
The average life of an android phone in the Philippines is 30 months. It is known that the
standard deviation of the said phones is 4 months. A test is conducted to validate whether the
claim is true. 20 units are randomly chosen to be tested and it was found that the mean life is 34
months. Test the hypothesis using a level significance of 5%.

SOLUTION:
GIVEN: x̄ = 34 months
µ = 30 months
n = 20 units
s = 4 months

STEP 1: HYPOTHESES
H0: The average life of an android phone is equal to 30 months. (µ = 30)
Ha: The average life of an android phone is not equal to 30 months. (µ ≠ 30)

STEP 2: LEVEL OF SIGNIFICANCE (α)

α = 0.05
df = n1 – 1 = 20 – 1 = 19
critical/tabular value (t.v.) : t0.05 = 2.093

STEP 3: TEST STATISTICS

t-test for independent samples (one group)

STEP 4: COMPUTATION
 ( x̄−μ) √ n (34−30) √20 ( 4)( 4.472) 17.888
t= = = = =4.472
s 4 4 4

STEP 5: DECISION RULE

Since the computed value of 4.472 is greater than the critical/tabular value of 2.093,
reject the H0.

STEP 6: CONCLUSION:
The average life of an android phone is not equal to 30 months.

EXAMPLE 2:
The following are the scores of 10 male and 10 female BS students in Statistics. Test the
null hypothesis that there is no significant difference between the performance of male and
female BS students in the test. Use t-test at 0.05 level of significance.
MALE (X1) FEMALE (X2)
14 12
18 9
17 11
16 5
4 10
14 3
12 7
10 2
9 6
17 13

SOLUTION:
STEP 1: HYPOTHESES
H0: There is no significant difference between the performance of the male and female BS
students in Statistics. (x̄ 1= x̄ 2)
Ha: There is a significant difference between the performance of the male and female BS
students in Statistics. (x̄ 1 ≠ x̄ 2 ¿
STEP 2: LEVEL OF SIGNIFICANCE (α)
α = 0.05
df = n1 + n2 – 2 = 10 + 10 – 2 = 18
critical/tabular value (t.v.) : t0.05 = 2.101

STEP 3: TEST STATISTICS

t-test for independent samples

STEP 4: COMPUTATION
SOLUTION:

MALE FEMALE
2 2
X1 X1 X2 X2
14 196 12 144
18 324 9 81
17 289 11 121
16 256 5 25
4 16 10 100
14 196 3 9
12 144 7 49
10 100 2 4
9 81 6 36
17 289 13 169
TOTAL 131 1891 78 738

Σ X1 = 1891 Σ X2 = 738

n1 = 10 n2 = 10

x ̄ 1 = 131/10 x ̄ 2 = 78/10
x ̄ 1 = 13.1 x ̄ 2 = 7.8

SS (131 )2
1=¿ Σ x 12−
(Σx ¿¿1 )2
¿¿ = 1 , 891− =1 , 891−1,716.10=174 . 9
n1 10
SS ( 78 )2
2
2=¿ Σ x 2 −
( Σx¿¿ 2)2
¿¿ = 738 - = 738 – 608.4 = 129.6
n2 10
x̄ 1− x̄ 2
t=
SS 1+ SS 2 1 1
√( )
( + )
n1 +n2−2 n1 n2

13.1−7.8
t=
(√ 174.9+ 129.6 1 1
10+10−2 ) 10 10
( + )

5.3
5.3 5.3 5.3
t= 304.5 1 = = = =2.881
√ (
18
)( ) √ (16.92)(0.2) √3.384
5
1.8395

STEP 5: DECISION RULE

Since the computed value of 2.881 is greater than the critical/tabular value of 2.101,
reject the H0.

STEP 6: CONCLUSION:
Since the t-computed value of 2.881 is greater than t-tabular value 2.101 at 0.05 level of
significance with 18 degrees of freedom, the null hypothesis is disconfirmed in favor of the
research hypothesis. This means that there is a significant difference between the performance of
the male and the female BS students in Statistics, implying that the male students performed
better than the female students considering that the mean/average score of the male students of
13.1 is greater compared to the average score of the female students of only 7.8.
B. T-TEST FOR DEPENDENT/CORRELATED SAMPLES

What is the t-test for correlated/dependent samples?

The t-test for correlated/dependent samples is another parametric test applied to one
group of samples. It can be used in the evaluation of a certain program or treatment. Since this is
another parametric test, conditions must be met like the normal distribution and the use of
interval or ratio data.

When do we use the t-test for correlated/dependent samples?

The test for correlated/dependent samples is applied when the mean before and the mean
after are being compared. The pretest (mean before) is measured the treatment of the intervention
is applied and then the posttest (mean after) is likewise measured. Then the two means (pretest
vs posttest) are compared.

Why do we use the t-test for correlated/dependent samples?

The t-test for correlated/dependent samples is used to find out if the difference exists
between the before and after means. If there is a difference in favor of the posttest then the
treatment or intervention is effective. However, if there is no significant difference then the
treatment is not effective.
This is the appropriate test for evaluation of government programs. This is used in an
experimental design to test the effectiveness of a certain technique or method or program that
had been developed.
FORMULA:

Dm
t=
( Σ D)2

√ Σ D 2−
n(n−1)
n

Where: Dm = the mean difference between the pretest and posttest.

ΣD2 = the sum of the squares of the difference between the pretest
and the post test.
ΣD = the summation of the difference between the pretest and the
posttest.
n = the sample size.

EXAMPLE 3:
An experimental study was conducted on the effect of programmed materials in English
on the performance of 20 selected college students. Before the program was implemented the
pretest was administered and after 5 months the same instrument was used to get the posttest
result. The following is the result of the experiment. Use α = 0.05 level.
Pre te s t Pos tte s t
Stude nt
X1 X2
1 20 25
2 30 35
3 10 25
4 15 25
5 20 20
6 10 20
7 18 22
8 14 20
9 15 20
10 20 15
11 18 30
12 15 10
13 15 16
14 20 25
15 18 10
16 40 45
17 10 15
18 10 10
19 12 18
20 20 25
STEP 1: HYPOTHESES
H0: There is no significant difference between the pretest and posttest, OR
The use of the programmed materials did not affect the students’ performance in English.
Ha: The posttest result is higher than the pretest result.

STEP 2: LEVEL OF SIGNIFICANCE (α)

α = 0.05
df = n1 – 1 = 20 – 1 = 19
critical/tabular value (t.v.) : t0.05 = 1.729 (the Ha is directional, so it’s a one-tailed test)

STEP 3: TEST STATISTICS

t-test for correlated samples

STEP 4: COMPUTATION
Prete st Posttest D
Student 2
X1 X2 (X1 - X2 ) D
1 20 25 -5 25
2 30 35 -5 25
3 10 25 -15 225
4 15 25 -10 100
5 20 20 0 0
6 10 20 -10 100
7 18 22 -4 16
8 14 20 -6 36
9 15 20 -5 25
10 20 15 5 25
11 18 30 -12 144
12 15 10 5 25
13 15 16 -1 1
14 20 25 -5 25
15 18 10 8 64
16 40 45 -5 25
17 10 15 -5 25
18 10 10 0 0
19 12 18 -6 36
20 20 25 -5 25
-81 947

Σ D = - 81 Σ D 2 = 947

Dm 81/20

Dm -4.05
 Dm −4.05
t= −4.05 −4.05
( Σ D)2 (−81)2 −4.05

−4.05
√
2
ΣD −
n(n−1)
n
=

√ 947−
20
20 (20−1)
=
√ 947−328.05 =
20(19) √618.95 =
380
√1.6288
=

=−3.173
1.2762

STEP 5: DECISION RULE

Since the computed absolute value of -3.173 is greater than the critical/tabular value of
1.729 reject the H0.

STEP 6: CONCLUSION:
The t-computed value of -3.173 is beyond the t-critical value of 1.729 at 0.05 level of
significance with 19 degrees of freedom. The null hypothesis is therefore disconfirmed in favor
of the research hypothesis. This means that the posttest result is higher than the pretest result. It
implies that the use of the programmed materials in English is effective.
References:
Belecina, R., Baccay, E., & Mateo, E. (2016). Statistics and Probability. Manila: Rex Book
Store, Inc.
Broto, A. (2007). Simplified Approach to Inferential Statistics. Mandaluyong City: National
Bookstore.

Calaca, N., Uy, C, Noble, N. & Manalo, R. (2016). Statistics and Probability. Quezon City: Vibal
Group, Inc.
Carpio, J., Guillermo, J., Angeles, R., De Guzman, R. (2015). Statistics and Probability.
Mandaluyong City: Books Atbp. Publishing Corp.
Gomez, S., (2020). Data Analytics Module. Trinity University of Asia.
Mercado, J. & Orines, F. (2016). Next Century Mathematics: Statistics and Probability. Manila:
Phoenix Publishing House, Inc.
Ramos, Josephine, etal. (2020). Statistics and Probability: Grade 11, Quarter 4 Module 1 –
Testing Hypothesis. Pasig City: Department of Education.

Zorilla, R., Esller, B., Mendoza, V., De Castro, L.M., Abenilla, G. (2016). Statistics and
Probability for Senior High School. Malabon City: Mutya Publishing House, Inc.

https://www.investopedia.com/terms/s/sampling-distribution.asp

https://courses.lumenlearning.com/boundless-statistics/chapter/sampling-distributions/

https://link.quipper.com/en/organizations/556417baab63a7752c006f37/curriculum#curriculum

http://homepages.math.uic.edu/~bpower6/stat101/Sampling%20Distributions.pdf

https://stats.libretexts.org/Bookshelves/Introductory_Statistics/Book
%3A_Introductory_Statistics_(Shafer_and_Zhang)/06%3A_Sampling_Distributions/6.02%3A_
The_Sampling_Distribution_of_the_Sample_Mean

https://onlinestatbook.com/2/calculators/normal_dist.html
https://www.analytics-toolkit.com/glossary/parametric-test/
https://www.google.com/url?sa=i&url=https%3A%2F%2Ffanyv88.com%3A443%2Fhttps%2Fwww.pinterest.com%2Fpin
%2F326299935485508656%2F&psig=AOvVaw2PwWdz_R4BUJ4WHAOgGZkH&ust=161665
2427190000&source=images&cd=vfe&ved=2ahUKEwjyhbreocjvAhV_0IsBHX-
RBi8Qr4kDegUIARCtAQ

https://www.graphpad.com/support/faq/what-is-the-difference-between-ordinal-interval-and-
ratio-variables-why-should-i-care/#:~:text=An%20interval%20scale%20is%20one,score
%20(300%2D850).