You are on page 1of 30

Inferential Statistics Overview

Jose Jurel M. Nuevo, RMT, MA CHEM, MSMT,FRIMTECH, PhD, DrPH


Parametric and Nonparametric

Tests
Parametric Test – Rely on
assumptions about the shape of
the distribution, assume Normal Normal
Distribution in the underlying Distribution
population and about the form of means being
parameters (mean, sd) of SYMMETRIC
assumed distribution. and
 Nonparametric Tests – Rely on MESOKURTIC
no or few assumptions about the at the same time
shape or parameters of the
population distribution from which
the sample was drawn.
Testing
Hypothesis
 1. Formulate the Null (Ho) hypothesis
and Alternative (H) hypothesis.
 2. Select the level of Significance (α).
 3. Determine the test statistic to be
used.
 4. Define the Area of Rejection.
 5. Compute for the values of the
Statistical Test.
 6. Draw conclusion.
Tests of Significance
• Parametric tests of significance – used if there are at
least 30 observations, the population can be
assumed to be normally distributed, variables are at
least in an interval scale
– Z tests are used with samples over 30. There are four
kinds (two samples or two categories)
– t-tests are used when samples are 30 or less.
• Single sample t-test (one sample)
• Independent t-test (two samples)
• Paired t-test (two categories
T-tests
• When to use them?
–Paired t-tests: When comparing the MEANS of a continuous variable in two non-
independent samples (i.e., measurements on the same people before and after a
treatment)

• Ex) Is diet X effective in lowering serum cholesterol levels in a sample of 12 people?


• Ex) Do patients who receive drug X have lower blood pressure after treatment then they
did before treatment?

–Independent samples t-tests: To compare the MEANS of a continuous variable


in TWO independent samples (i.e., two different groups of people)

• Ex) Do people with diabetes have the same Systolic Blood Pressure as people without
diabetes?
• Ex) Do patients who receive a new drug treatment have lower blood pressure than those
who receive a placebo?

Tip: if you have > 2 different groups, you use ANOVA, which compares the means of 3 or more groups
T-tests
• What does a t-test tell you?
– If there is a statistically significant difference between the
mean score (or value) of two groups (either the same
group of people before and after or two different groups
of people)
• What do the results look like?
– Student’s t
• How do you interpret it?
– By looking at corresponding p-value
• If p < .05, means are significantly different from each other
• If p > 0.05, means are not significantly different from each
other
Example.

Using 5% alpha
Solutions: The use of the module does not
 Ho: leads to improvement.

The use of the module leads to


 Ha: improvement.
 Alpha: 5%
 Tail: 1 tailed test 3.231
 Test Statistic: t-Test
 Critical Value: +1.729
 Area of Rejection
Decision:
REJECT Ho

1.729
Analysis
: 1. Comparison ofModules
Table Achievement Score of Using

Test Mean Diff CV Comp Decision Remark


(α=5%) t
Pre 18.40 2.05 1.729 3.231 Reject Significant
Post 20.45 Ho
Table 1 presents the comparison of the pretest and posttest
result in using a module. Based on the table, a difference
of 2.05 between the average achievement score of pretest
(x=18.40) and posttest (x=20.45), is proven statistically
significant since, the computed t of 3.231 is greater than
the critical value of 1.729 at 5% level of significance. Thus,
there is strong evidenced that on average, the module leads
to the improvement of achievement. With the results it is
implied that the use of Module……….. Moreover, it
supports the study of Ruiz (2015), he claimed
Example.
 A course in Mathematics is taught to
students by a usual strategy. A second group
12
of 10 students was given the same course by
means of strategy x. At the end of the
semester the same examination was given to
each group, below were the results. Test the
hypothesis using 0.01 level of significance.
Usual 82 83 80 81 82 80 82 81 83 82 80 83

Strategy 83 83 84 80 85 87 86 85 83 82
x
Solutions: There is no significant difference… .
 Ho: There is a significant difference
between the achievements of
using usual strategy and strategy
 Ha: x.
 Alpha: 1%
 Tail: 2 tailed test
 Test St a tistic: t-
Test
 Critical Value: +2.845
 Area of Rejection
-3.19
- 2.845
2.845 Decision:
REJECT Ho
Analysi : Table 2. Comparison of Mean Scores of Using
s
Method Mean Diff
Usual and Strategy X
CV Comp Decision Remark
(α=5%) t
Usual 81.58 2.22 +2.84 -3.19 Reject Significant
Strategy 5 Ho
Strategy 83.80
X

Table 2 presents the comparison of mean scores of using usual


and strategy x. Based on the table, a difference of 2.22 between
the average mean score of usual (x=81.58) and strategy X
(x=83.80), is proven statistically significant since, the computed t
of -3.19 is less than the critical value of -2.845 at 1% level of
significance. Thus, there is a significant difference between the
achievements of using usual strategy and strategy x. The results
show a strong evidenced that teaching strategy used has an
effect towards the achievements of students. With the results it
is implied that ……………….
Analysis of Variance

(ANOVA)
It is a parametric test used to test the
significant difference of three or
more groups.
 a method of testing the equality of three or
more population means by analyzing
sample variations
Case 1 – Equal no. of respondents in each
group
Case 2 – Unequal no. of respondents in
each group
Example
 Twelve
. (12) overweight participated in a study to
subjects the weight reducing regiments. Subjects are grouped
compare
according to initial weight and each of the subjects was randomly
assigned to one of the three reducing regiments. At the end of the
experimental period the following weight losses, in pounds were
recorded.
 After eliminating differences due to initial weight, do these data
provide sufficient evidence to indicate a difference in regiment
effects? Use 1% level of significance.
Regiment A Regiment B Regiment C
12 14 15
13 14 14
15 11 12
12 10 19
CLICK DATA

INPUT DATA
CLICK
DATA
ANALYSI
S
CLICK
ANOVA
SINGLE
FACTOR,
THEN OK
CLICK
REGIMENT A
DATA WILL
APPEAR

HOLD SHIFT KEY AND ARROWKEYS


GOING TO THE RIGHT AND DOWN TO
COVER ALL THE DATA
DATA WILL
APPEAR
CLICK

EDLIATBTE
OLS1% CLICK
CLICK WINDOW
OUTPUT
RANGE
CLICK
OK

DATA WILL
APPEAR
DECISION: ACCEPT
NULL HYPOTHESIS
Solutions: There is no significant difference
in the regiment effect.
 Ho: µ1=µ2=µ3
 Ha: µ1≠µ ≠µ 2 3
There is a significant difference
in the regiment effect.
 Alpha: 1%
 Tail: 1 tailed test
 Test Statistic: ANOVA
 Critical Value: 8.022
 Area of Rejection
1.626

Decision:
ACCEPT Ho
8.022
An lysis
Table 1. Comparison of the Effects of the Regiments
a : Mean CV Comp Decision Remark
Regiments
(α=1%) f
A 13.00 8.022 1.626 Accept Not
B Ho Significant
12.25
C 15.00
Table 1 presents the comparison of the effects of the three
regiments in losing weights. Based on the table, the
average weight lose of regiment A (x=13.00), Regiment B
(x=12.25) and regiment C (x=15.00), is proven statistically
not significant since, the computed f of 1.626 is less than
the critical value of 8.022 at 1% level of significance. Thus,
there is strong that the three (3)
evidenced have reducing effect in the
regimentit is the
results impliedsame body. With
that……….. the
Moreover, it supports
the study of Ruiz (2015), he claimed
that………………………
Solution:
Solutions: There is no significant difference
in the number of cavities produce.
 Ho: µ1=µ2=µ3
 Ha: µ1≠µ ≠µ 2 3
There is a significant difference in
the number of cavities produce.
 Alpha: 1%
 Tail: 1 tailed test
 Test Statistic: ANOVA
 Critical Value: 6.927
 Area of Rejection 22.558

Decision:
REJECT Ho
6.927
Analysis
Table 1. Comparison of the Acquired Number of Cavities
:
Toothpaste Mean CV
(α=1%)
Comp Decision
f
Remark

#1 31 6.927 22.558 Reject Ho Significant


#2 18
#3 16
Table 1 presents the comparison of the acquired number of
cavities. Based on the table, the average acquired number
of cavities in using toothpaste #1 (x=31), #2 (x=18) and #3
(x=16), is proven statistically significant since, the
computed f of 22.558 is greater than the critical value of
6.927 at 1% level of significance. Thus, there is strong
evidenced that there is a variation in the effect of the three
(3) toothpaste in preventing cavities. With the results it is
implied that ……….. Moreover, it supports the study of
Ruiz (2015), he claimed that………………………
POSTH
OC
 Scheffe’s Test - (Posthoc) –This
is used to find out where the
differences lie.
𝟐
𝒙𝟏 − 𝒙𝟐
𝑭′ = = 𝟐𝟔. 𝟎𝟒
Computation 𝒔𝟐𝟐(𝒏𝟏 + 𝒏𝟐൯
𝒏 𝟏 𝒏𝟐
n Mean
#1 vs #2, Reject Ho,
SIGNIFICANT, thus toothpaste #2
is better than #1
Toothpaste # 1 6 30.33333 2
𝑥1 − 𝑥3
𝐹′ = = 37.53
𝑠2 (𝑛 1 + 𝑛3ሻ
Toothpaste # 2 4 17.25 𝑛1𝑛3
#1 vs #3, Reject Ho,
Toothpaste # 3 5 SIGNIFICANT, thus #3 is better
15.6
than #1
2
𝑥2 − 𝑥 3
𝐹′ = = 0.38
F’critical = (F Crit)(n-1)
= (6.927)(2) 𝑠2 (𝑛 2 + 𝑛3ሻ
𝑛2 𝑛3
= 13.854
#2 vs #3, Accept Ho, NOT
SIGNIFICANT, thus #2 and #3
have the same effect
Example

.
Consider a 10-year study in which a sample of 15 people
has been observed while using toothpaste #1, #2, and #3,
respectively. Let us assume that the following participants
have been randomly assigned to each of the statement and
that the study has provided the data given in the table
below:

 Test at 0.01 level of significance, whether the difference


among the number of cavities are significant. (Assuming
Normal Distribution)

You might also like