You are on page 1of 39

Hypothesis Testing:

a logical supposition and educated guess


based on observation. It provides tentative
explanation for certain events, phenomena
or behaviors. It is a suggested solution for
an unexplained occurrence that does not fit
into current accepted scientific theory.
Kinds of Hypothesis
1. ()- expresses the idea of non-
significance. The statement opposes the perceived
conclusion in a research

. () expresses the
idea of significance. It generally represents the
hypothetical statement that the researcher wants to
prove.
Examples: State the null and alternative
hypothesis for the following statements.
1. Significant difference in the leadership performance
of a sectarian and non-sectarian school administrator.
Ho:

Ha:

Significant relationship between the attitude score and


academic performance.
Ho:

Ha:
TYPE I and TYPE II ERRORS
Decision Condition Ho is true Condition Ha is true
Reject Ho Type I Error Correct decision
Do not reject Ho Correct decision Type II Error

The first type of error committed is


when the null hypothesis is rejected wherein it is the
true solution to the problem

The second type of error committed is


when the alternative hypothesis is rejected wherein it
is the true solution to the problem.
Level of Significance
The significance level is the criterion used for
rejecting the null hypothesis. Choice of levels is largely
subjective and most common choices for level of
significance are = . and = .

Setting at . indicates that the researcher is


% confident that the is TRUE or accepted and the
researcher has % chance of rejecting the .
Reducing the at . indicates that the
researcher is % confident that he is making the
RIGHT decision and 1% chance of committing ERROR.
Rejection & Acceptance Region

The is the region in which the


null hypothesis () is rejected if the test statistics falls
within it. By setting the level of significance, the rejection
or critical value can be determined. It serves as the
the is rejected or not rejected.

The location of the rejection region is determined


by the nature of the alternative hypothesis(), whether
the hypothesis is or a test.
One Tailed Test
A test of any statistical hypothesis, where the
alternative is one-sided. In a one sided test, it is a
directional statement of the alternative hypothesis
wherein only one meaning can be interpreted from it.
Example of One Tailed Test;
1.) Ho: The average income of a skilled worker in the
Philippines is not greater than 300Php.

Ha: The average income of a skilled worker in the


Philippines is greater than 300Php

For the above example the best test to


be used is one tailed test because the
alternative hypothesis (Ha) is stated
directional. Greater than 300Php
has only one meaning.
Two Tailed Test

A test of any statistical hypothesis, where the alternative


is two sided. It means that the statement is stated as
non-conditional. Non-directional means that there are
two ideas that could be interpreted from the very way
how the alternative hypothesis is stated.
Example of Two Tailed Test;
: There is no significant difference in the leadership
performance of a sectarian and non-sectarian school
administrator.

: There is a significant difference in the leadership


performance of a sectarian and non-sectarian
administrator.

The above alternative hypothesis (Ha) is stated non-


directional. Significant difference in the leadership
performance can be interpreted that the sectarian
perform well than the non-sectarian or the other way
around. Two meanings can be interpreted from the
alternative hypothesis (Ha)
Steps in Hypothesis Testing
1. State the null hypothesis (Ho) and alternative
hypothesis (Ha).
2. Set the (level of significance).
3. Choose that statistical test appropriate to test the
hypothesis.
4. Determine the tabular value for the test.
5. Compute the value of the statistical test.
6. Determine the significance of the computed value
from the tabular value.
Continuation
7. We have the cases in formulating the alternative
hypothesis testing. , is the direction of the Ha is to
the left of the hypothesized value, the computed value
(Cv) is less tha or equal to the tabular value (Tv).
, if the direction of the Ha is to the right of the
hypothesized value, the computed value (Cv) is greater
than or equal to the tabular vale (Tv). , if the
direction of the Ha is either to the left of to the right of
the hypothesized value, the absolute computed value
(Cv) is greater than tabular value (Tv). Do not reject Ho
if the assumptions in step 7 are not satisfied.
8. Formulate conclusion.
Critical Values of Z-Test
Type of One Tailed Test Two-Tailed
Test Test
Level of Left Tailed Right Tailed
Significance

= 0.05 Reject if z 1.645 Reject Reject if 1.96 or


1.645 Reject if = 1.96

= 0.01 Reject if 2.33 Reject if Reject if 2.575or


2.33 Reject if 2.575

= 0.10 Reject if 1.28 Reject if Reject if 1.645or


1.28 Reject if 1.645
Z-test and T-test: The Z-test and T-test is
used to determine whether there are any significant
differences between the means of two groups.
Formula for Z-test
1. Sample mean compared with population
mean( )

=

where:
= sample mean
= pop mean
= total # of samples
=pop standard deviation
continuation 4. Comparing two sample
2. Comparing two proportions
independent mean
2 & 2 1 2
1 2 0 = 1 2
= +
1 2
2 2
1 + 2
3. Comparing one sample In determining the tabular
proportions value for z-test use table
0 for critical values as z
= based on the area under
0 0
the normal curve.
Example 1:
A record from local Solution:
college shows that the mean 1. : The 80 students
score in the qualifying are not better
examination of sophomore than the rest of
students was 45 with a the sophomore
standard deviation of 2.5.
students.
samples of 80 students were
taken and found to have a : The 80 students
mean score of 48. are the 80 are better than the
students better than the rest of the
rest of the sophomore sophomore
students? Use 0.01 level of students.
significance.
continuation
2. = 0.01 48 45 80
3. z-test (one tailed) =
2.5
4. = 2.326 Tvalue
5. Computation using the = 10.73
formula for sample mean 6. Decision: Since Comp z
compared with population value of 10.73 is greater tha
mean the tabular value of 2.326.
therefore, reject null
= hypothesis (Ho).

Substitute the given value
= 48 7. Formulate conclusion: The
=45 80 students are better than
the rest of the sophomore
=80 students.
=2.5
Example 2:
A researcher wishes to find out whether or not
there is a significant difference between the academic
performance of CRIM and EDUC students in his
school. By random sampling, he took a sample of
respondents for CRIM. These respondents were
found to have a . The researcher also took
a sample of respondents for EDUC students. They
were found to have a and it is given that
the population for CRIM and
EDUC students are . Is there a
significance difference between the two samples at
. level of significance?
Solution:
1. Ho: There is no significant difference between the
academic performance crim and educ students.
Ha: There is a significant difference between the
academic performance crim and educ students.
2. = 0.05
3. (two tailed)
4. 1 = 1.96 Tabular value
5. Computation using the formula for two sample means
1 2 0
=
2 2
1 + 2
continuation
Substitute the given values in
the formula 6. : Since the
1 = 92 computed z values of 8.24 is
greater than the tabular value
2 = 80 of 1.96 therefore reject Ho.
1 = 5
2 = 6
7. : There is a
1 = 54 significant difference
2 = 75 between the academic
performance of crim and
9280 educ students.
=
52 62
+
54 75

= 8.24
Example 3:
In a market survey, 31 out of 126 males and 68 out of 190
females indicate a preference for a particular brand of mobile
phones. Do the data warrants the conclusion that there is sex
difference exists in brand preference for mobile phones? Use
0.05 level of significance.

Solution:
1. Ho: Sex difference does not exist in brand
preference for mobile phones.
Ha: Sex difference exists in brand preference for
mobile phones.
2. = 0.05
3. z-test (two-tailed)
4. = 2.576 Tabular Value
continuation
= # of units possessing
the attribute of interest in
5. Computation using the the first group of sample
formula for comparing two
sample proportions. = # of units possessing
1 2 the attribute of interest in
= 1 2 the second group sample
+
1 2
where: = total # of
1 + 2 respondents in the 1st
=
1 + 2 group

= total # of
respondents in the 2nd
group
continuation
Substitute the given values in
the formula
5. : Since the
1 =31/126 computed z value of 2.09 is
2 =68/190 less than tabular value of
1 =126 2.58 therefore we cannot
2 = 190 reject Ho.
31 68
126 190
= 6. : Sex
99 217 1 1 difference does not exist in
. +
316 316 126 190 brand preference for
mobile phones.
= 2.09 = 2.09
Formula for T-test
1. Sample mean compared with population mean

=

where:
=
=
n =
s =
Second Formula
T-test for independent samples:

This is used to compare the values of the means


from two samples and test whether it is likely that the
samples are from populations having different mean values
and it is assumed that the variance of the two populations
are equal 2 = 2
1 2
=
1 1 1 2 + 2 1 2 2 1 1
+
1 + 2 2 1 2
where:
1 = mean of the 1st group of sample
2 = mean of the 2nd group of sample
1 = total # of the 1st group of sample
2 = total # of the 2nd group of sample
1 = standard deviation of the 1st group of sample
2 = standard deviation of the 2nd group of sample
3. T-test of correlated or paired samples:
Compare the means between two related groups obtained
from the same sample.
0
General formula: =

where:
=mean diff bet the matched pairs of measures
0 = hypothesized diff bet the matched pairs of
meas
= sd of the diff bet the matched pairs of meas
= total number of samples
If 0 = then we can use this formula:
1 2
=
2 1 2 2 1 2
1
where:
2 1 = difference between the matched pairs of meas
= total number of observation
To determine the tabular value for t-test compute first the
degree of freedom (df), and look for the tabular value from the
table of t-distribution.

= 1 for one sample group


= 1 + 2 2 for two sample group
Example 1
Last year an average life span of 69 years with standard
deviation of 7.8 years was shown by random sample of 100
recorded deaths in the Philippines. Using 0.05 level of
significance does it seem to show that the average life span today
is greater than 72 years ago?
Solution:
1. Ho: The average life span today is NOT greater than 72
years
Ha: The average life span today is greater than 72 years
2. = 0.05
3.t-test ( one tailed)
4. = 1, = 100 1 = 99
= 1.66 Tabular Value
continuation
5. Computation using the =
6972 100
formula for sample mean 8.9
compared with population = 3.37 = 3.37
mean 6. : Since the t
computed value of 3.37 is
= greater than tabular value

Substitute the given values of 1.66. therefore, we reject
in the formula our null hypothesis (Ho).
= 69 7. Formulate conclusion:
= 72 The average life span today
= 100 is greater than 72 years.
= 8.9
Example 2:
As part of an industrial training program, some
trainees are instructed by method 1 and some by method
2. Both methods involve computer-administered lessons,
nut method 1 also involves personal attention of an
instructor. The assignment of trainees to instructional
methods is by random allocation. Random samples of size
10 are taken from large groups of trainees instructed by
the two methods.
The following values are their scores
on a relevant achievement test:
Metho Score on Achievement Test Mean SD
d
1 81 71 79 83 76 75 84 90 83 78 80 5.40

2 59 65 62 59 57 64 60 56 66 62 61 3.37

The claim has been made of that personal attention of


an instructor will improve a trainees score. Test the
claim at the 0.05 level of significance
Solution : Personal attention of an instructor
will NOT improve a trainees score
: Personal attention of an instructor
will improve a trainees score

2. = 0.05
3. t-test (one-tailed)
4. = 1 + 2 2,
= 10 + 10 2,
= 1.73 Tabular Value
5. Computation using the formula for T-test for independent
samples
1 2
=
1 1 1 2 + 2 1 2 2 1 1
+
1 + 2 2 1 2
continuation
where:
1 = 10 1 = 80 1 = 5.40
2 = 10 2 = 61 2 = 3.37
80 61
= = 9.44
10 1 5.40 2 + 10 1 3.372 1 1
+
10 + 10 2 10 10

6. : Since computed value of 9.44 is greater


than tabular value of 1.73. Therefore, we our null
hypothesis ()
7. : Personal attention of an
instructor will improve a trainees score.
Example 3:
A doctor is interested in determining whether a film
about exercise will change 10 persons attitudes about
exercise. The results of this questionnaire are shown. A
higher numerical value shows a more favourable attitude
towards exercise. Is there enough evidence to support the
claim that there was a change in attitude? Test the claim at
0.05 level of significance for the difference of the two
means:
Solution:
1. : There is NO enough evidence to support the claim that
there was a change in attitude about exercise.
: There is enough evidence to support the claim that
there was a change in attitude about exercise.
2. = 0.05
3. t-test (two-tailed)
4. = 1 1, = 10 1 = 9
= 2.26 Tabular Value
5. Computation using the formula for t-test of correlated or
paired samples
1 2
=
2 1 2 2 1 2
1
continuation
Before1 12 11 14 9 8 6 8 5 4 7 T
After2 13 12 10 9 8 8 7 6 5 5
2 1 -1 1 -4 0 0 2 -1 1 1 -2 -1
2
2 1 1 1 6 0 0 4 1 1 1 1 29

2 1 1 2 1 2 29
1
= = 0.18 = 0.18
10 29 1 2
10 1
6.: Since t comp value of 0.18 is less than tabular value of
2.26. Therefore, we cannot reject the null hypothesis (Ho)
7. : There is not enough evidence to
support the claim that there was a change in attitude about exercise
The End!
Use Microsoft Excel for easy
computation.
Answer Exercises 8.1-8.10
Good Luck! !

You might also like