RESEARCH DESIGNS &

STATISTICS
MA. ROWENA M. BAYRANTE
ASST. PROF. I
CBSUA
 RESEARCH DESIGN

 Refers to the categorized procedure for conducting

the research which includes:
+ method of research to be used
+ study population
+ instrument for gathering data & its preparation
+ sampling design
+ statistical treatment of data
+ analysis and presentation of data
 RESEARCH DESIGNS

 May be classified as:

 + Survey + Descriptive
 + Case Study + Experimental
 + Historical + Comparative
 + Experimental + Correlational
 + Ex Ante
 + Ex Post Facto
 SURVEY

 involves determining information about variables rather than about

individuals
 are employed to measure existing phenomena without inquiring into why it
exists
 do not take into account the relationship between variables
 main intention is to use the data for problem solving rather than for
hypothesis testing
 -maybe conducted as a census or a sample survey
Example:
*Farming Practices in the Bicol Region
*Teaching Practices in Mathematics of Secondary
Teachers in Camarines Sur
 CASE STUDY

 is the type which primarily examines a person,

a specific group of persons or other subjects
as a case or one particular situation or object
 involves deeper inquiry and examination of the
behavior, the practice, etc.
Example:
*Efren Penaflorida: The CNN Hero
*Open University: The UP Experience
 DESCRIPTIVE STUDY

 Used to classify a study whose data shall

primarily be used to describe, characterize
or provide a picture of the research subject,
respondents or population
- Determines and reports the way things are
Example:
*Training Needs of CBSUA Graduate Faculty
*Typhoon Preparedness of Households in
Sta Teresita, Baao, Camarines Sur
 COMPARATIVE STUDY

 used when the study aims to compare, find

significant differences between or among
groups of responses or respondents
Example:
*Water Quality Assessment of Different
Drinking Sources in Pili, Camarines
Sur
*Growth and Yield Response of Tomato
to Different Organic Fertilizerz
 CORRELATIONAL STUDY

 used when the study aims to proves or disproves relationship

between or among variables
 used to estimate a relation the form and degree
 ascertain how much variation is caused by one variable
to that of the other variable
Example:
* Teaching Strategies and Students Academic
Performance in Mathematics
* Correlates of Academic Performance in Science
Among Elementary Pupils in Selected
Public Schools in Camarines Sur
 HISTORICAL STUDY

 attempts to record, describe, organize,

analyze, etc. past events or accounts
concerning the life of a person, an institution,
country and other collective entities
Example:
Ninoy Aquino’s Assasination: Its Contribution
to Philippine Democracy
 EXPERIMENTAL

 is a study which is carried out under controlled

condition in order to discover a presumed but
unknown cause-effect relationship between variables
and test or establish the hypothesized relationship
Example:
*Growth and Yield Performance of Rice Using
Organic Fertilization
 EXPERIMENTAL DESIGNS

Single Factor Experiments

-Completely Randomized Design (CRD)
-Randomized Complete Block Design
-Latin Square
Factorial Experiments
-Completely Randomized Design
-Randomized Complete Block Design
-Latin Square
-Split Plot Design
-Lattice Designs
 EX POST FACTO

means “from after the fact”

researcher investigates a problem by studying the
variables in retrospect
systematic empirical inquiry in which the researcher
does not have direct control of the independent variables
because their manifestations have already occurred or
because they are inherently not manipulable
Researcher observes that groups are different on some
variable and then he attempts to identify the major
factor which have caused this difference
 EX POST FACTO

Researcher observes first an effect and then tries to

determine the cause
Example:
Effects of Prolonged Smoking
 STATISTICS

Why do we need it?

- to increase reliability of research findings
- to produce precise and accurate estimates
 of population parameters
- for forecasts or making predictions
- samples is used instead of the population
-needs to make an inference about the population
through sample data
 What Determines the Choice of Statistics?

Objectives of the Study

Level of Measurement
Sample Size
Sampling Techniques Used
 POPULATION DESCRIPTION

 refers to the totality of the persons or objects

or subjects which the study intends to cover
or include as participants or respondents of
the study
 SAMPLE SIZE

 refers to the number or quantity of the persons or

objects that could represent
the entire population
 also include how it will be determined
showing a formula like:
N
n = -------------------
1 + Ne2
 TYPES OF STATISTICS

Parametric Statistics
Non-parametric Statistics

Parametric Statistics
- requires a knowledge about the population
where the samples were/ will be drawn
- use when the sample size is 30 or more
- level of measurement is at least interval scale
- samples should be/ are randomly drawn
- more powerful than non-parametric
Non-parametric Statistics
- less powerful
- used for sample size less than 30
- level of measurement should be at least in
the nominal
- used for rank-ordered data
 MEASUREMENT

Refers to the manner or way of assigning

values to a particular variable, or an
attribute or trait such as gender, marital status,
age, height, weight, etc.
 LEVELS OF MEASUREMENT

Nominal
- lowest level of measurement
- take values that give names or labels to various
categories or classifications with no particular
ordering
- information that can be obtained form processing
data on these variables are limited to frequency
counts and percentages
Examples:
Sex, Marital Status, College Degree, Variety
Ordinal
- same with the nominal with the categories
having an inherent ordering
- differences between categories cannot be measured
and has no meaning
- information that can be obtained from processing data
on these variables is limited to frequency counts with
additional insight on the rank or order of the categories
specified
Example:
Educational Attainment, Year Level,
Degree or Order of Preference, Social Class
Interval Scale
- basically quantitative variables with differences
between two consecutive quantities being constant
- intervals between categories can be quantified and
have meaning
- has no true zero value or point of origin
Examples:
Temperature, Intelligence Quotient (IQ)
Ratio
- highest level of measurement
- has all the characteristics of an interval scale
- has an absolute zero or true zero point
Example:
height, weight, yield, width, no. of bros. & sis.
 SUMMARIZING MEASURES

Measures of Central tendency

- describes the centralness of a given data
Commonly Used Measure:
* For quantitative data
+ Mean – is the average of all the values
+ Median – is the middle value
* For qualitative data
+ Mode – is most frequently occurring value
 FREQUENCY COUNT

 is the method by which the number of

respondents, responses or occurrences
of the subject of study is determined
 is used to describe the respondents or
responses in terms of quantity and is
considered as a descriptive statistical
tool
-applicable for nominal and ordinal measurement
or qualitative data such as sex, educational attainment
 ARITHMETIC MEAN

- enables the researcher to arrive at one value

that will describe the responses or
respondents in totality which is commonly
known as the average
- also regarded as a descriptive statistical tool
 WEIGHTED MEAN

 is used as a summarizing value for responses

that vary in terms of degree of importance
such as responses with rating scale, grades
with different units, etc.
 provides an “overall” value description to a
particular variable
 also considered as a descriptive statistical
tool
 STANDARD DEVIATION

 is the average deviation of the responses

or observations from the mean which
measures the variability of the responses
in terms of how close or far are them from
each other
 PERCENTAGE

 refers to the representation of a response

with respect to a group of 100
 computed by dividing the frequency with
the total frequency multiplied by 100
 MEASURES OF VARIABILITY

 Describe the clustering or spread of the given data

 Determines if data are very close or dispersed from
each other
Commonly Used Measures:
+ Range – determines distance between highest and
lowest value
+ Standard Deviation – average deviation from the
mean
+ Coefficient of Variation – determines degree of
variability
 Test for Significant Differences

Parametric
- Z test
- T test
- ANOVA
Non-parametric
- Sign Test
- Wilcoxon Signed Rank Test
- Kruskal Wallis
- Friedman Test
 Z & T - TEST

 Determines whether the means of the values

of the variable being measured for two or
more groups under study is significant
or not
- used when the mean can be computed
- ideal for interval and ratio scale measurement
- Z test is used for a sample size of 30 or more or
the population standard deviation is known
- T test is used for sample size less than 30 and the sample
standard deviation is used to estimate the population
standard deviation
 ANOVA

 determines whether or not there is significant

differences among variables which is done
through the Analysis of Variance
 Analysis of Variance segregates the total variation
of responses or data taken into appropriate parts
that measures different sources of variation
 uses the F test
 can be performed using one way or two way ANOVA
 ONE WAY ANOVA

 used when the data can be classified or grouped

using one criterion
 equivalent to the analysis for completely randomized
design where the treatments are completely
randomized in the experimental units where the
units are homogeneous
 TWO WAY ANOVA

 Used when the data are grouped using two criterion

as row and column variables
 Equivalent to the analysis performed for
Randomized Complete Block Design where the
experimental units are blocked to control inherent
variation before the treatments are randomized
 ideal for field experiments with one known source of
variation
 Sign Test

 used instead of the Z and T test when the sample size

is less than 30 and the population is non normal
 a statistical tool for comparing means from one
sample data
 values are converted into plus and minus signs
 CHI-SQUARE

 is a tool which tests whether one variable

is dependent or not with another variable
wherein the values of the variables are
obtained through frequency counts
- used for nominal and ordinal measurement
 Wilcoxon Rank Sign Test

 is used instead of the Z and T test for two sample

case when the sample size is less than 30 and the
population is non normal
- Used for comparing means of two sample data
 Kruskal Wallis

 a non-parametric test equivalent to the one way

ANOVA
 used for testing differences in means from three or
more sample data
 data are usually converted into ranks
 Correlation

 used to determine the degree of relationship

between an independent and dependent variable
wherein the values of the variables are numeric
which can be associated into a one to one
correspondence
 REGRESSION

 use to determine the form of relationship

between an independent and dependent
variable which is expressed in an equation
of the form:

Y = a + bX (simple regression)

Y = βo + β1X1 + β2X2 + … + βnXn

(multiple regression)
 Friedman Test

 a non-parametric test equivalent to the two way

ANOVA
 data are usually converted into ranks
 used for small size sample
 Example

Title:
Teaching Strategies in Science and Academic
Performance of Grade V Pupils in Science
Statement of the Problem:
1. What is the profile of Science Teachers?
2. What are the teaching strategies used in Science?
3. Does the profile of the teachers affects the choice
of teaching strategies?
4. Is there significant differences in the teaching
strategies used in Science among teachers?
4. What is the academic performance of Grade V
pupils in Science?
5. Do the teaching strategies affect academic
performance of Grade V pupils in Science?
Thank you!!!