Unit 7: Statistical Analysis of Quantitative Data

Lesson 2: Testing Difference of Two

Means

Practical Research 2
Senior High School Applied - Academic
How can we ensure that
the differences between
groups are tested
accurately?

ENGAG EXPLOR EVALUAT

EXPLAIN EXTEND 22
 ● Identify the different
Learning parametric and nonparametric
Objectives tests in comparing the
difference between two means.
At the end of the
lesson, you should be
● Compare different groups by
able to do the
following:
using appropriate statistical
tests.

ENGAG EXPLOR EVALUAT

EXPLAIN EXTEND 33
Student Average Number of Minutes Average Number of Minutes
Spent on Social Media During Spent on Social Media During
Weekdays Weekends

1 60 minutes 90 minutes

2 30 minutes 180 minutes

3 45 minutes 120 minutes

4 90 minutes 90 minutes

5 60 minutes 60 minutes
ENGAG EXPLOR EVALUAT
EXPLAIN EXTEND 44
Exploring
Differences Bivariate analysis is defined as
Between the analysis of two variables.
Two Means

ENGAG EXPLOR EVALUAT

EXPLAIN EXTEND 5
What types of research are
appropriate to test differences
between means?

6
 ● Parametric continuous data
● Dependent variable at an
Assumption
interval or ratio level
s of t-test
● Random sampling
● Normal distribution of the data

ENGAG EXPLOR EVALUAT

EXPLAIN EXTEND 7
 Independent Groups t-test

Used when the two sets of data have come

from two independent sources or two
different groups.

ENGAG EXPLOR EVALUAT

EXPLAIN EXTEND 88
 Independent Groups t-test

Example:
Testing the difference between plants grown
with organic fertilizer and plants grown with
conventional fertilizer

ENGAG EXPLOR EVALUAT

EXPLAIN EXTEND 99
Formula for Independent Groups t-test

ENGAG EXPLOR EVALUAT

EXPLAIN EXTEND 10
10
Tip

Some formulae may look tedious at

first, but the key to answering
problems involving long formulae is
to break it down into smaller pieces.

11
Independent Groups t-test

Martina is an agriculturist who wants to find

out whether the growth of okra plants sprayed
with biopesticides is significantly different from
okra plants which are sprayed with
conventional pesticides. She took a total of 16
random samples, with eight samples for each
group, and measured the length of the plants.
ENGAG EXPLOR EVALUAT
EXPLAIN EXTEND 12
12
Independent Groups t-test

Biopesticides Group Conventional Pesticides

(in cm.) Group (in cm.)
3.0 2.5
4.5 1.5
5.0 3.0
5.2 3.2
3.5 2.0
4.0 2.9
2.8 1.9
ENGAG 4.7
EXPLOR
EXPLAIN EXTEND 1.6
EVALUAT 13
13
Independent Groups t-test

1. State the null and alternative hypothesis.

ENGAG EXPLOR EVALUAT

EXPLAIN EXTEND 14
14
Independent Groups t-test

2.a Calculate the degrees of freedom (df).

ENGAG EXPLOR EVALUAT

EXPLAIN EXTEND 15
15
Independent Groups t-test

2.b Set alpha level, identify the critical values,

and state a decision rule.
● Refer to the critical values table.
● Decision rule: “Reject H0 if t < -2.145 or >
2.145.”

ENGAG EXPLOR EVALUAT

EXPLAIN EXTEND 16
16
Independent Groups t-test

2.c Calculate the test statistic: t-test for

independent groups.

t = 4.43
ENGAG EXPLOR EVALUAT
EXPLAIN EXTEND 17
17
Independent Groups t-test

2.d Make a decision whether to reject or accept

the null hypothesis.

ENGAG EXPLOR EVALUAT

EXPLAIN EXTEND 18
18
Independent Groups t-test

3. Draw a conclusion from the analysis.

“The average length of okra plants sprayed
with biopesticides (M = 4.09) is significantly
different from the average length of okra
plants sprayed with conventional pesticides
(M = 2.33), t(14) = 4.43, p < .05.”

ENGAG EXPLOR EVALUAT

EXPLAIN EXTEND 19
19
 Dependent Groups t-test

Used when two sets of data come from the

same group of people.

ENGAG EXPLOR EVALUAT

EXPLAIN EXTEND 20
20
 Dependent Groups t-test

Example:
Administering a science diagnostic test to the
same group of students during the start of the
semester and then administering again at the
end of the semester

ENGAG EXPLOR EVALUAT

EXPLAIN EXTEND 21
21
Formula for Dependent Groups t-test

ENGAG EXPLOR EVALUAT

EXPLAIN EXTEND 22
22
Dependent Groups t-test

Maureen wants to know if the average daily

intake of water during the month of March is
significantly different from the average daily
intake of water during the month of December.
She has 10 participants for the study and
recorded their average daily intake of water
during the month of February and March.
ENGAG EXPLOR EVALUAT
EXPLAIN EXTEND 23
23
Dependent Groups t-test

March (in glasses) December (in glasses)

8.00 7.00
8.50 6.50
9.00 8.00
10.00 6.00
9.50 8.50
9.25 8.25
8.75 7.75
11.00 8.00
10.80 6.80
ENGAG EXPLOR
12.25 EXPLAIN EVALUAT
8.25
EXTEND 24
24
Dependent Groups t-test

1. State the null and alternative hypothesis.

There is no difference between the average

daily intake of water during the month of
March and the month of December.

ENGAG EXPLOR EVALUAT

EXPLAIN EXTEND 25
25
Dependent Groups t-test

1. State the null and alternative hypothesis.

There is a significant difference between the

average daily intake of water during the
month of March and the month of
December.
ENGAG EXPLOR EVALUAT
EXPLAIN EXTEND 26
26
Dependent Groups t-test

2.a Calculate the degrees of freedom (df).

ENGAG EXPLOR EVALUAT

EXPLAIN EXTEND 27
27
Dependent Groups t-test

2.b Set alpha level, identify the critical values,

and state a decision rule.
● Critical value: +/- 2.262
● Decision rule: Reject H0 if t < -2.262 or >
2.262

ENGAG EXPLOR EVALUAT

EXPLAIN EXTEND 28
28
Dependent Groups t-test

2.c Calculate the test statistic: t-test for

dependent groups.

t = 4.97
ENGAG EXPLOR EVALUAT
EXPLAIN EXTEND 29
29
Dependent Groups t-test

2.d Make a decision whether to reject or accept

the null hypothesis.

ENGAG EXPLOR EVALUAT

EXPLAIN EXTEND 30
30
Dependent Groups t-test

3. Draw a conclusion from the analysis.

“The average daily intake of water during the
month of March (M = 9.71) is significantly
different from the average daily intake of
water during the month of December (M =
7.51), t(9) = 4.97, p < .05.”
ENGAG EXPLOR EVALUAT
EXPLAIN EXTEND 31
31
 ● do not rely on the
distribution of the data
from the sampled
Nonparametric
Tests of Difference participants

● Mann-Whitney U test
and Wilcoxon test
ENGAG EXPLOR EVALUAT
EXPLAIN EXTEND 32
32
Tip

Nonparametric tests are

recommended to use when the
sample size is small, and there is an
issue with the data set’s symmetry.

33
 Mann-Whitney U-test

● Used to examine differences between

independent groups
● The nonparametric counterpart of the
independent groups t-test
● Used when the data obtained is based on
ranks
ENGAG EXPLOR EVALUAT
EXPLAIN EXTEND 34
34
Formula for Mann-Whitney U-test

ENGAG EXPLOR EVALUAT

EXPLAIN EXTEND 35
35
Mann-Whitney U-test

Ariel is a student researcher who wants to

know whether the stress levels of female
college students are significantly different than
that of male college students.

ENGAG EXPLOR EVALUAT

EXPLAIN EXTEND 36
36
Mann-Whitney U-test

To measure the stress level of the students,

she administered the University Students
Stress Scale (USSS) to 20 participants (10 male
and 10 female). The questionnaire can be
scored from 0-100 (higher scores represent
higher stress levels).
ENGAG EXPLOR EVALUAT
EXPLAIN EXTEND 37
37
Mann-Whitney U-test

Female Male
79 47
41 76
57 60
88 51
82 57
66 57
82 44
79 47
57 38
ENGAG EXPLOR 69 28
EVALUAT
EXPLAIN EXTEND 38
38
Mann-Whitney U-test

1. Rank all the scores together.

Female Rank Male Rank
79 16.5 47 5.5
41 3 76 15
57 9.5 60 12
88 20 51 7
82 18.5 57 9.5
66 13 57 9.5
82 18.5 44 4
79 16.5 47 5.5
57 9.5 38 2
ENGAG EXPLOR
69 14 EVALUAT
EXPLAIN EXTEND28 1 39
39
Mann-Whitney U-test

2. Look for the sum of the ranks.

In the given problem, the rank sum for
the female group is 139 and 71 for the
male group.

ENGAG EXPLOR EVALUAT

EXPLAIN EXTEND 40
40
Mann-Whitney U-test

3. Find the U score.

ENGAG EXPLOR EVALUAT

EXPLAIN EXTEND 41
41
Mann-Whitney U-test

4. Refer to the U tables to find out whether

the U score obtained is significant. Critical
value = 23.

ENGAG EXPLOR EVALUAT

EXPLAIN EXTEND 42
42
Mann-Whitney U-test

5. Conclusion: There is a significant

difference between the stress levels of
the students, with females having
higher levels of stress.

ENGAG EXPLOR EVALUAT

EXPLAIN EXTEND 43
43
 Wilcoxon Test

● Used when you want to examine differences

between related sets of data
● The nonparametric counterpart of the
dependent groups t-test
● Used when the data obtained is based on
ranks
ENGAG EXPLOR EVALUAT
EXPLAIN EXTEND 44
44
Wilcoxon Test

Anna wants to find out if there is a significant

difference between the weight of people
before and after taking a new diet pill called
Slimming Z. Anna had six participants in the
study and recorded their weight (in kg) before
and after taking Slimming Z for three months.

ENGAG EXPLOR EVALUAT

EXPLAIN EXTEND 45
45
Wilcoxon Test

Before Slimming Z After Slimming Z

48 45
60 60
50 52
53 52
72 73
58 58
ENGAG EXPLOR EVALUAT
EXPLAIN EXTEND 46
46
Wilcoxon Test

1. Calculate the difference between each pair

of scores, then rank.

ENGAG EXPLOR EVALUAT

EXPLAIN EXTEND 47
47
Wilcoxon Test

Before After Difference Rank

Slimming Z Slimming Z
48 45 3 5
60 61 1 1
50 52 -2 4
53 52 1 1
72 73 -1 1
58 54 4 6
ENGAG EXPLOR EVALUAT
EXPLAIN EXTEND 48
48
Wilcoxon Test

2. Find the value of T, which is the sum of the

ranks for the differences with the less
frequent sign:
T=4+1
=5
In this problem, there are only two negative
signs which have ranks 4 and 1.
ENGAG EXPLOR EVALUAT
EXPLAIN EXTEND 49
49
Wilcoxon Test

3. Find the critical value of T on the Wilcoxon

table.
Critical value: 0

ENGAG EXPLOR EVALUAT

EXPLAIN EXTEND 50
50
Wilcoxon Test

4. Findings:
There is no significant difference between
the weight of the participants before and
after taking the Slimming Z (Wilcoxon T =
3, two-tailed p < 0.05).

ENGAG EXPLOR EVALUAT

EXPLAIN EXTEND 51
51
Remember

It is important to note that the t

used for parametric t-tests is in
lowercase, whereas the T used in
Wilcoxon test is in uppercase.

52
 Wrap Up
The t-test is a parametric test that aims to
test whether the differences between the
means of two groups or conditions are
statistically significant.

The independent group t-test is used

when the two sets of data have come from
two independent sources or two different
groups.
53
 Wrap Up
The dependent group t-test is used
when two sets of data come from the
same group of people.

The Mann–Whitney U-test is a

nonparametric test used when you want
to examine differences between
independent groups.

54
 Wrap Up

The Wilcoxon test is the nonparametric

counterpart of the dependent groups t-
test and is also used when the data
obtained is based on ranks.

55
Wrap Up