STATISTICAL TOOLS FOR DATA Student #1: 5’ 5” character or kind but not in amount, and non

ANALYSIS Student #2: 5’ 7” numeric value.

Student #3: 5’ 2” Example: sex, birthplace, geographic,
Statistics - Refers to the techniques by location
which quantitative data are collected, Types of Variables 2. Quantitative Variable - Numerical in
presented, organized, analyzed and 1. Independent Variable nature and can be ordered or ranked,
interpreted. -The independent variable is therefore a numeric value.
the one condition that you Example: weight, height, age
Nature of Statistics change in an experiment.
1. Descriptive Statistics - Includes Example: In an experiment Caution!!!
techniques which are concerned with measuring the effect of Most quantitative variables can be asked in
summarizing quantitative data. This method temperature on such a way as to make them a categorical
make use of graphical or computation of solubility, the independent variable.
data variable is temperature. Example: Age - Please choose the category
2. Inferential Statistics - Considered 2. Dependent Variable - that describes how old you are: 0-19, 20-39,
higher form of statistics. This method make The dependent variable is the 40-59, 60-79, and OVER 80
use of logical reasoning in order to find or variable that you measure or
establish cause and effect relationship. observe. The dependent Classification of Quantitative Variables
variable gets its name because 1. Discrete - Values can be counted using
Uses of Statistics: Sports, Education, it is the factor that integral values.
Medicine, Government Policy, Social is dependent on the state of the Examples: number of enrollees, salary of
Problems and Issues, Business and, independent variable. teachers, length of service
Economics, and Home Example: In the experiment 2. Continuous - Can assume a numerical
measuring the effect of value over an interval.
Review of Terms: Population, Sample, temperature on solubility, Examples: height, weight, speed
Individual, and Variable solubility would be the
dependent variable. 4 TYPES of MEASUREMENT SCALES
Population - The entire group of 3. Controlled Variable 1. Nominal Scale - An unordered set of
individuals is called the population. -A controlled variable or categories identified only by name.
For example, a researcher may be interested constant variable is a variable Nominal measurements only permit you to
in the relation between class size (variable that does not change during an determine whether two individuals are the
1) and academic performance (variable 2) experiment. same or different.
for the population of third-grade children. Example: In the experiment Example Preference for President, Race,
Sample - Usually populations are so large measuring the effect of Gender, Religious affiliation, Major,
that a researcher cannot examine the entire temperature on solubility, Opinion items (favor vs. oppose, yes vs. no)
group. Therefore, a sample is selected to controlled variable could 2. Ordinal Scale - An ordered set of
represent the population in a research study. include the source of water categories. Ordinal measurements tell you
The goal is to use the results obtained from used in the experiment, the the direction of difference between two
the sample to help answer questions about size and type of containers individuals.
the population. used to mix chemicals, and the Example: Political ideology (very liberal,
amount of mixing time liberal, moderate, conservative, very
allowed for each solution. conservative) Anxiety, stress, self esteem
4. Extraneous Variables - are (high, medium, low) Mental impairment
"extra" variables that may (none, mild, moderate, severe) Government
influence the outcome of an spending on environment (up, same, down)
experiment, but aren't taken 3. Interval Scale - An ordered series of
into account during equal-sized categories. Allow us not only to
measurement. These variables rank order the items that are measured, but
won't impact the final also to quantify and compare the sizes of
conclusion drawn by the differences between them. The zero point is
experiment, but they located arbitrarily on an interval scale.
may introduce error into Example: temperature, as measured in
scientific results. degrees Fahrenheit or Celsius, constitutes an
Example: You are conducting interval scale.
Individual - The objects described by a set an experiment to see which 4. Ratio Scale - An interval scale where a
of data, individuals may be people, animals paper airplane design flies value of zero indicates none of the variable.
or things. longest. You may consider the They are very similar to interval variables;
For example, students in a Math class. color of the paper to be an in addition to all the properties of interval
Variable - any characteristic of an extraneous variable. variables, they feature an identifiable
individual. A variable can take on different absolute zero point, thus they allow for
values for different individuals. Classification of Variables statements such as x is two times more than
Characteristics that makes two or more 1. Qualitative/Categorical y.
values which varies across the source of Variable -Represent Example: measures of time or space.
data. differences in the quality,
Example: A Frequency Distribution is a grouping of
Heights of individuals: data into mutually exclusive categories
showing the number of observations in each C. Related or Independent Pearson’s Product Moment Of
class. The table shows a frequency Samples Correlation (r) - It is the common measure
distribution for a set of quantitative data.  Independent - of the correlation between two variable X
Probability of selection and Y. It is designated by the letter (r) and
A Frequency Table is a grouping of of samples in one group called as “Pearson’s (r)”. It reflects the
qualitative data into mutually exclusive is not affected by the degree of linear relationship between two
classes showing the number of observations selection in the other variables. Correlation indicates the strength
in each class. group. and direction of linear relationship between
Example:  Related - Sample in one two random variables.
group is dependent/
affected by the other Degree of Relationship between Two
group (paired, matched). Variables using Pearson’s (r)
D. Assumptions About the  The degree of relationship using
Test Pearson’s r ranges from +1 to -1.
Constructing Frequency Distribution  Parametric Test - Can  A correlation for +1 indicates a
Problem be used if the positive correlation between paired
Prepare a frequency distribution for the assumptions about the variables.
following array of data. parameter like normality,  Positive correlation indicates that as
Step 1: Compute for the range, by independence and the value of one variable increases the
subtracting the lowest score from the homogeneity of variance value of the other variable also
highest score (H-L) hold true. increases.
Step 2: Decide on the number of classes.  Non-Parametric Test -  A positive correlation indicates a
Step 3: Determine the class interval or When the assumptions direct relation between two paired
width. for the use of parametric variables.
The formula is: i = Range/k where i is the tests are questionable in  A correlation of -1 indicates a perfect
class interval, H is the highest observed the data. negative correlation.
value, L is the lowest observed value, and k  A negative correlation indicates that an
is the number of classes. increase in the value of one variable
Step 4: Set the individual class limits. creates a corresponding decrease in the
Step 5: Tally the scores into the classes. Selecting a Statistical Test other variable.
Step 6: Count the number of items in each Type of Data
class. Data Analysis Quantitativ
The data will be Quantitativ e Data from
Categorical
e Data Non-
There are Two Reasons to Construct collected through the Goal
from Normal
Data
(Nominal or
Frequency Distributions: questionnaire and will be Normal Population
Ordinal)
Population or Ordinal
1. Assists with choosing the appropriate test encoded and statistically Data
statistic (Parametric vs. Non-parametric). processed using statistical Proportion
1. Describ Median,
(Frequencies
2. Assists with identification of outliers. software called PASW e One Mean, SD Interquartile
and
Statistics version 24 (SPSS for Group Range
Percentage)
Windows). 1. Compar Fisher’s Test
The data gathered e Two Unpaired T- Mann- (Chi-square
Unpaired Test whitney Test for Large
will be analyzed with the use Groups Samples)
of descriptive statistical tools 2. Compar McNemar’s
such as _________. e Two Paired T- Wilcoxon
Test;
Paired Test Test
Cochran’s Q
Groups
CHOOSING APPROPRIATE 3. Compar
STATISTICAL ANALYSIS e Three or One-way Kruskal- Chi-square
TECHNIQUE More
Unmatched
ANOVA wallis Test Test

Groups
What question type do you want to answer? 4. Compar
Descriptive - What is the typical take home e Three or Repeated-
Friedman
More Measures Cochrane Q
pay of High School teachers at MNHS? Matched ANOVA
Test

Correlational/Predictive - Is there a Groups

relationship between take home pay and 5. Quantif Contingency
performance of High School teachers at y Pearson Spearman Coefficients
Correlation Correlation Correlation (Phi,
MNHS? / Cramer’s V)
Group Differences / Cause and Effect - Is Association
Simple
the performance of Group A better than that 6. Predict Linear
Non- Simple
of Group B in a certain span of time? Do Regression
PEARSON’S PRODUCT Value from
or
parametric Logistic
male teachers have higher incidence of Another Regression Regression
depression versus female teachers?
MOMENT OF Variable Nonlinear
Regression
CORRELATION (R) Multiple
Considerations in Choosing the AND 7. Predict Linear
Non- Multiple
Value from Regression
Appropriate Test Statistic PARAMETRIC TEST Several or Multiple
parametric
Regression
Logistic
Regression
Nonlinear
A. Objective of the study Variables
Regression
B. Variable type & level of measurement
 A correlation of zero (0) means there A study was
is no linear relationship between the conducted to find out if the
two variables. number of hours spent by
students studying their lesson
Correlation Value is correlated to their average
grade. Ten students were Interpretation:
Descriptive
Pearson’s r randomly selected to form part - The computed value of Pearson’s r is 0.71
Equivalent
of the study and the results which indicate a very high positive
Very Small
0.01 – 0.24 were shown on table. correlation between number of hours spent
Correlation
by the students in studying and
Moderately the student’s average grade.
0.25 – 0.49 Small - The computed value of 0.71 is
Correlation greater than the tabulated value of
High Pearson’s r which is 0.6319 at
0.5 – 0.70
Correlation 0.05 level, therefore the null
Very High hypothesis is rejected.
0.71 – 0.99
Correlation - For the students to succeed in
Perfect their studies they need to spend
1.0
Correlation longer hours of studying. The
Formula for Computing Pearson’s positive correlation indicates that
r the longer the time spent in
studying their lesson the higher the grades
they will receive.
N ( ƩXY ) − ( ƩX ) ( ƩY )
r= [N ( Ʃ Y 2 ) − ( ƩY )2 ]¿ Example No. 2:
√¿ ¿
A study was conducted to establish
significant relation ship in the spelling skills
Where: between male and female students. Result
N = sample population of the study was shown as follows:
ƩXY = the sum of the product of X and Y.
ƩX = the sum of all entries representing
variable X.
ƩY = the sum of all entries for variable Y. Solution:
(ƩX²) = the sum of squared entries in X. 1. State your hypothesis. Solution:
(ƩY²) = the sum of squared entries in X. 2. Prepare the table which will 1. State your hypothesis.
show following column: X; Y; 2. Prepare the table which will show
Steps in the Computation of Pearson’s r XY; X²; Y² and get the sum of following column: X; Y; XY; X²; Y² and
1. Use the step by step testing of hypothesis. each column. get the sum of each column.
2. Compute for the value of Pearson’s r by 3. Substitute the formula. 3. Substitute the formula.
preparing a table with the following 4. Set the level of significance 4. Set the level of significance at 0.05 level.
column: X; Y; XY; X²; Y². at 0.05 level. 5. Determine the degree of freedom and
3. Get the sum of each column and 5. Determine the degree of compare the computed value to the
substitute the formula. freedom and compare the tabulated value.
4. Interpret the result by using the computed value to the
descriptive equivalent of Pearson’s r. tabulated value. Hypothesis:
Compare the computed value of r to the Null: There is no significant relationship
tabulated value and interpret the result. Hypothesis:
between the spelling skills of male and
Null: The number of hours
female students.
Tabulated Value spent by the student in
studying has no relationship to Alternative: There is a significant
The tabulated value of Pearson’s r is
student’s average grade. relationship between the spelling skills of
determined by the degree of freedom with:
df = N – 2. Alternative: The number of male and female students.
To identify whether the computed hours spent by the student in
Pearson’s r is significant or the null studying has a relationship to
hypothesis is rejected or accepted at the student’s average
desired significance level, compare the
computed and tabulated value to guide your
decision.
Interpretation of the Pearson’s r
If the computed value of Pearson’s r is
greater than the tabulated value, the null
hypothesis is rejected in favor of the
alternative hypothesis.

Example No. 1:
Interpretation:
- The computed value of Pearson’s r of 0.45
indicates that there is moderately small
correlation between the spelling skills of
male and female students.
- The computed value of 0.45 is less than
the tabulated value of Pearson’s r which is
0.5760 at 0.05 level, therefore the null
hypothesis is accepted.
- Result of the study reveals that the
moderately small relationship between the
spelling skills of male and female students
has no significant relationship.

PARAMETRIC
TEST
- Statistic test is called parametric test if the
information about the population is
completely known by means of its
parameters.
- Specific assumption are made regarding
the population
- If the population is normally distributed.
- Null hypothesis is made on parameters of
the population distribution.
- Applicable for variable
- Are also applied to both interval - scale
and ratio - scale data.
Interval Scale – measures equal intervals,
no true zero ratio relationship.
- May be use in interval between
observation on terms of fixed unit of
measurement, example: Measure of the
Temperature.
Ratio Scale – measures equal interval, true
zero ratio relationship, examples are age
and income.
Examples: t-test, ANOVA, ANCOVA,
Pearson’s r