STATISTICAL TOOLS P2

MEASURES
OF
DISPERSION
 DISPERSION
 Dispersion
is a way of describing
how spread out a set of data is.

 When a data set has a large

value, the values in the set are
widely scattered or
heterogeneous; when it is small
the items in the set are tightly
clustered or homogeneous.
The spread of a data set can be
described by a range of
descriptive statistics:

1. Range;
2. Mean Absolute Deviation;
3. Variance; and
4. Standard Deviation
 RANGE
❑ Range is the difference between
the highest and lowest values in
the given set of data.
❑ Thisstatistical tool is
appropriate for interval and ratio
scale.
In a pre-experiment on proper nutrition, two groups of
randomly selected kindergarten learners were weighed to
determine the body weights. The group of learners
without treatment have the following body weights: 18.5.
15.7, 20.2, 19.8, 20.0, and 17.8 kg while the group of
learners with treatment have the following weights: 19.9,
21.1, 16.5, 15.0, 19.9, 20.4 and 17.7 kgs. Describe the
weights of the kindergarten learners.

Group Lowest Highest value Range

value
Without Treatment 15.7 kg 20.2 kg 4.5 kg
With treatment 15.0 kg 21.1 6.1 kg

Since the range of the weights of the group without treatment is less than
the range of the group with treatment, then the weights of the group
without treatment is relatively homogenous than of the group with
treatment.
EXERCISE
 Table 1

Mean and Score Range of the Division-wide Assessment in Mathematics

School Mean score, 𝒙

ഥ Score range
A 78.5 4.5
B 79.3 6.6
C 80.6 5.0
D 74.5 6.0

Describe and interpret the results of the division-wide assessment in

mathematics.

School C obtained the highest mean of 80.6among the three schools

while school D got the lowest mean score of 74.5. School C performed
better than the other three schools in mathematics. School D did not
perform well in mathematics. However, school A has the lowest range
of 4.5 and school B has the largest range of 6.6. School A has
comparably homogenous scores in mathematics compared to the
other schools. Meanwhile, school B has comparably heterogenous
scores as to other three schools.
RANGE FOR GROUPED DATA

Range = Exact Lower Limit of the Topmost Class –

Exact Upper Limit of the Bottom most
Class

Range = Exact Upper Limit of the Topmost Class –

Exact Lower Limit of the Bottom most
Class
 Table 2
Exact lower Exact upper limits =
limits 126 +0.5= 126.5
Frequency Distribution of =Scores
118- 0.5 in
Licensure Examination with
= 117.5
corresponding Exact Limits

118, 127,
136, 145,
154, & 163
are lower
limits

126, 135,
144, 153,
162, & 171
are upper
limits

Use Range = Exact Lower Limit of the Topmost Class – Exact Upper Limit of the
Bottom most Class

The topmost class is 163 – 171 and Bottom most class 118 – 126
Exact lower limit of the top most class = 163 – 0.5 = 162.5
Exact upper limit of the bottom most class = 126 + 0.5 = 126.5

Range = 162.5 - 126.5 = 36

With topmost class of 163-171 and bottom most cals of 118-126 , the range of distribution is 36.
EXERCISE
The two groups of Grade 1 pupils were considered on a study about
reading.

Group 1 Group 2
Score f Score f
90 - 94 4 90 - 94 5
85 - 89 12 85 - 89 8
80 – 84 10 80 – 84 11
75 – 79 8 75 – 79 10
74 - 70 3 74 - 70 5

Compare the two groups in terms performance in reading and

homogeneity.
 MEAN ABSOLUTE DEVIATION
❑ Themean absolute deviation is the
average distance of each of the value in
the given set of data from the mean.

❑ Thisstatistical is appropriate for interval

or ratio scale
MEAN ABSOLUTE DEVIATION
𝟐 + 𝟑 + 𝟑 + 𝟑. 𝟓 + 𝟓 + 𝟒 + 𝟔. 𝟓 + 𝟕 + 𝟖 + 𝟖
ഥ=
𝒙
𝟏𝟎

ഥ=𝟓
𝒙
𝟑+𝟐+𝟐+𝟏.𝟓+𝟏+𝟎+𝟑+𝟑+𝟐+𝟏.𝟓
MAD = 𝟏𝟎
MAD = 1.9
THE DISPERSION OF THE SET OF SCORES HAS A MEAN ABSOLUTE DEVIATION OF 1.9.
Mean Absolute Deviation for Ungrouped Data

MAD =
 x−x
n
MAD = Mean Absolute Deviation of the Ungrouped Data

x = Measurement in the set of data

x = Mean of the set of the data

n = Number of measurements
EXAMPLE
 Table 3
Mean and Mean Absolute Deviation of the Pretest Scores in Science
Group Mean, 𝒙
ഥ Mean Absolute Deviation, MAD
Girl 45.3 3.21
Boy 43.5 1.34
Describe and interpret the table.

Between the girls and boys’ pretest scores, the girls obtained a
highest mean score of 45.3 while the boys got 43.5. The girls has
greater knowledge about science as compare to the boys. Between
mean absolute deviation of the scores, the boys has the lowest
mean absolute deviation of 1.34 while the girls has 3.21. The scores of
the boys comparably homogenous as the scores of the girls.
EXERCISE
Three groups of randomly selected Grade 11 students took part in
an experimental study on the impact of language in teaching mathematics.
A 25-point test was administered to the determine the effectiveness of
instruction using different languages.

Table 4
Mean Score and Mean Absolute Deviation of the Performance of Grade 11 Students
in Mathematics
Group Mean Score MAD
English-instruction 15.6 4.33

Ilokano-instruction 15.1 5.21

EngIlokano - instruction 16.2 3.98

Describe and compare the results of the tests.

Mean Absolute Deviation for Grouped Data

MAD =
 f x−x
f
MAD = Mean Absolute Deviation of the Grouped Data
x = Class mark of the class

x = Mean of the grouped data

f = Frequency of the class

EXERCISE
A quasi-experimental study was conducted to determine the effectiveness of
technology in teaching English. The following grades were earned by the
participants after the experiment:

Group without Treatment Group with Treatment

Score f Score f
95 - 99 1 95 - 99 3
90 - 94 4 90 - 94 5
85 - 89 12 85 - 89 8
80 – 84 10 80 – 84 11
75 – 79 8 75 – 79 0
74 - 70 3 74 - 70 5

Describe and draw your conclusion

 VARIANCE
❑ The average of the squared difference or
deviation from the mean
❑ Appropriate for interval or ratio scale
Variance for Ungrouped Data

s 2
=
 (x − x)
2

n −1
s2 = Variance of the ungrouped

x = Measurement in the set of data

x = Mean of the set of the data

n = Number of measurements
Variance for Grouped Data

 f (x − x )
2

s 2
=
n −1
s 2 = Variance of the Grouped Data
x = Class mark of the class

x = Mean of the Grouped Data

f = Frequency of the class

n = Total number of measurements

VARIANCE FOR GROUPED DATA

 n fu 2 − ( fu )2 
s2 = i2  
 n(n − 1) 
 
2
s = Variance
i of the grouped data

i = Class size

n = Total number of measurements

f = Frequency of the class

u = Unit code of the class

EXERCISE
A 20-item pretest in Calculus was administered to three groups of
randomly selected first year engineering students to determine their
preparedness for Mathematical Analysis 1. The results are as follows

Table 6
Mean and Standard Deviation of the Pretest Scores
Group ഥ
𝒙 𝒔𝟐
A 15.3 2.22

B 10.7 1.23

C 15.3 1.20

Describe and compare the results of the pretest.

From the result of the pre-test in calculus, group A and C are both very
ready for mathematical analysis 1 with mean score of 15.3. However,
the scores of group C is more homogeneous than of groups A and B.
STANDARD DEVIATION
❑ is the square root of the variance.

❑ Appropriate for interval or ratio scale.

STANDARD DEVIATION

s= s 2
EXERCISE
The impact of teacher-designed teaching strategy on the academic
performance in reading underwent experimental study. The academic
performance were based on the final grade of the learners and yielded the
following results:

Group ഥ
𝒙 s
Control 82.5 4.3
Experimental 83.1 2.7

Interpret the results and draw your conclusion.

From the results, the experimental group has better final grade af
83.1 as compared to the final grade of 82.5 of the control group.
Furthermore, the final grades of the experimental group comparably
homogenous with standard deviation 2.7 than that of the control
group with standard deviation of 4.3.
MEASURES
OF
SYMMETRY
Sk = 0 Sk > 0 Sk <0

Symmetrical is “ mirror-image”

Asymmetrical distribution – there is

no mirror igage
SKEWNESS
 Skewness is a measure of the asymmetry of the
probability distribution of a real-valued random variable
about its mean.
 Coefficient of Skewness sk is a measure of symmetry of a
distribution, or more precisely, the lack of symmetry.
 Coefficient of Skewness is an index of unevenness of the
distribution.
 If sk is > 0, the distribution is skewed to the right or
positively skewed. If sk < 0, the distribution is skewed to
the left or negatively skewed. And if sk =0, the distribution
is symmetric.
COEFFICIENT OF SKEWNESS FOR UNGROUPED DATA

sk =
 (x − x)
3

3
ns
sk = Coefficient of Skewness of distribution

x = measurement in the data set

x = mean of the data set

s = standard deviation of the data set

n = number of measurement
COEFFICIENT OF SKEWNESS FOR GROUPED DATA

 f (x − x )
3

sk = 3
ns
sk = Coefficient of Skewness of distribution
x = class mark of the class

x = mean of the distribution

s = standard deviation of the distribution

f = frequency of the class

n = number of distribution
 Table 7
Means and Coefficients of Skewness of the Final Grades in Chemistry
for Engineers of First Year Students

Section ഥ
𝒙 𝒔𝒌 Remarks
A 92 12.5 Positively skewed scores
B 𝟖𝟗 0 Normal distribution of scores
C 83 - 0.03 Negatively skewed

Complete the table and describe the results

From the results, section A performed better with a mean final grade
of 92 in Chemistry as compared to sections B and C with 89 and 83,
respectively. But section A has positively skewed distribution of final
grades with sk = 12.5 and section C has negatively skewed
distribution with sk = - 0.03. On the other section B has a normal
distribution with sk = 0.
Exercise
 Table 8
Mean and Coefficients of Skewness of the Scores in English

School ഥ
𝒙 𝒔𝐤 Remarks
Private 123 0.02

Public 179 0.07

Complete the table and describe the distribution

MEASURES
OF
KURTOSIS
 KURTOSIS
 Kurtosis is a measure of
“peakedness” or “flatness” of the
distribution.

 Coefficient of Kurtosis is a measure

of extent of peakedness of flatness
of distribution.

 If k > 3, the distribution is

leptokurtic. If k < 3, the distribution
is platykurtic. And if k = 3, the
distribution is mesokurtic.
COEFFICIENT OF KURTOSIS FOR UNGROUPED
DATA

k=
 (x − x)
4

4
ns

Coefficient of Kurtosis for Grouped Data

 f (x − x )
4

k= 4
ns
 Table 8
Mean and Coefficients of Kurtosis of the scores in English

School ഥ
𝒙 𝒌 Remarks
Private 123 3.33 Leptokurtic distribution
Public 179 - 0.04 Platykurtic distribution

Complete the table and describe the distribution

From the results, the private schools did not performed well with
mean of 123 as compared the performance of public schools with
mean of 179.
Parametric
Statistics
&
Non-parametric
Statistics
 PARAMETRIC STATISTICS
 Applicable for normal distributed continuous
data;
 Analyses to assess group means
Parametric Analyses Sample Size Requirement
1-sample t-test ❑ Two groups with Greater than 20
2-sample t-test ❑ Two groups with each group should have
more than 15 observations
❑ For 2 – 9 groups, each group should have
One-way ANOVA more than 15 observations
❑ For 10 – 12 groups, each group should have
more than 20 observations
NONPARAMETRIC STATISTICS
 Applicable for non normal distributed data;
 Applicable
when sample size does not meet
requirement of normal distributed data;
 Analyses to assess group medians;
 Appropriate
for all groups of data with
same dispersion or spread
GENERAL RULE
When data do not meet the
assumptions of the parametric
statistics, especially the
assumption about normally
distributed data,
nonparametric statistics are
used.
 Parametric Statistics Non-parametric Statistics
❖Pearson ❖ Spearman Rank Rho
Coefficient of Coefficient of Correlation, 𝝆
❖ Point Biserial Coefficient of
Correlation, r Correlation
❖ Chi Square Test for
Independence
❖1 - sample t - test ❖ 1 - sample Sign test,
❖ 1- sample Wilcoxon test
❖2 – sample t - test ❖ Mann – Whitney test
❖One – way ANOVA ❖ Kruskall Wallis test
❖ Mood’s Median test
HYPOTHESIS TESTING
❑ Hypothesis testing is a systematic
procedure for deciding whether the results
of a research study support a particular
theory which applies to a population.
❑ Hypothesis testing uses sample data
to evaluate a hypothesis about a
population.
❑A hypothesis test assesses how unusual the
result is, whether it is reasonable chance
variation or whether the result is too
extreme to be considered chance variation.
 HYPOTHESIS
Research hypothesis, Ha is the hypothesis that
the researcher proposes, also known as the
alternative hypothesis,

Null hypothesis, Ho and exis the opposite of the

research hypothesis presses that there is no
relationship between variables, or no
differences between groups
HYPOTHESIS TESTING PROCEDURE
1.Restate the research question in the
Statement of the Problem as research
hypothesis and a null hypothesis about the
populations.
Research question:
Is there significant relationship between the profile of
the respondents and their academic performance in
History?
Hypothesis:
There is no significant relationship between the profile
of the respondents and their academic performance in
History.
2.Determine the characteristics of the
comparison distribution.
Sex : Male and Female Respondents
Academic Performance in History of Respondents
correspondingly

3. Determine the cut off sample score

on the comparison distribution at
which the null hypothesis should be
rejected.
𝛂 = 𝟓% level of significance
 4.Determine the sample’s score on the
comparison distribution.
Computed coefficients and probability

5.Decide whether to reject the null hypothesis

Rule:
If computed probability 𝑝 ≤ 0.05 reject 𝐻0 ; otherwise
accept H0
 REFERENCES

Aladjec, V. Z. and V. N. Harinotov (2004). General Theory of Statistics. Estonia: Fultus Books

Dalgaard, Peter (2008). Introductory Statistics with R. United States of America: Springer
Science + Business

Gravetter, Frederick J., Larry B. Wallnau (2009). Statistics for the Behavioural Sciences. New
York: Wadsworth Cengage Learning

Hastie, Trevor, Robert Tibshirani, Jerome Friedman (2013).The Elements of Statistical Learning:
Data Mining, Inference, and Prediction. New York: Springer Science

Ramos, Frederick O. and Adela Cristeta Jamorabo Ruiz. (2009). Statistics for Allied Health
Sciences. Manila: Meriam & Webster Bookstore, Inc

Sternstein, Martin (2010). Barron's AP Statistics. New York: Barron’s Educational Series, Inc

Wasserman, Larry (2004). All of Statistics: A Concise Course in Statistical Inference. United
States of America: Springer Science + Business