Methods of Research

1
Methods of Research (Part III)

METHODS OF RESEARCH (PART III)

This learning material is the last part of the module entitled, Methods
of Research wherein the students are given opportunities to define and
characterize a descriptive research and be oriented with the things related to
descriptive research.
At the end of the week, college students should be able:
1. Have a background on the important concepts and principles
in doing experimental method of research
2. Characterize an experimental designs explain how to do the
different matching methods
3. Pin point the difference among the different experimental
designs

EXPERIMENTAL METHOD OF RESEARCH

WHAT IS EXPERIMENTAL RESEARCH?
It is a quantitative research in which one or more variables are
manipulated controlled and measured in terms of change in other variables
by the researcher systematically. It is a cause and effect research and
sophisticate statistical techniques are used in this type of quantitative
research.
CHARACTERISTICS OF EXPERIMENTAL METHOD
1. “Controlled” and the “Experimental” variables are the two
important and necessary actors .
2. The experimental (independent) variable is manipulated and
applied to a dependent variable while the control variable is held
constant.`
3. The effects of independent variable on the dependent is
measured.
4. They are usually numerical and caters statistical design which is
sophisticated.
5. Limited to small samples
6. Labor intensive

Experimental designs involves control and experimental groups

except for quasi experimental single group. The experimental group receives
the treatment which is under investigation and something is expected to gain
while the control group receives the ordinary technique.

Course Module
 DIFFERENT MATCHING METHODS
1. Person-to-person matching – on the basis of similar or identical
personal characteristics, people are selected . Intelligence is one of
the most important variables to be controlled. In this method,
after looking for people with similar or the same on the variable,
they are randomly assign to groups. There should be a
counterpart for the individual in the control group in the other
group – experimental group.

2. Matching groups – It is a less precise grouping method wherein

groups are paired on a variable. Mean and the standard deviation
of the scores I measuring the variables is the basis of similarity of
two groups.

3. Ranking method – Some selected variables like achievements,

grades, socio-economic variables and others depending on the
research problems are utilized to rank the subjects of the study.
There is a continuous ranking until all the subjects have been
assigned.

4. Homogenous group – Groups that are homogenous are compared

in this matching method. For example, if the researcher think that
age can affect the dependent variable, he may use only one age
group. And should be considered constant. The disadvantage of
this method is that the findings of study are limited only to the
involved subjects and less result generalization because there is
only one variable level is employed.

5. Using subjects as their own controls - Same subjects are used for
the treatment. For instance, there are 5 methods of instruction to
try out, all the subjects will receive the 5 treatments, one after the
other which is done in different order.

DIFFERENT EXPERIMENTAL DESIGNS

A. PRE-EXPERIMENTAL DESIGNS
1. Design #1 – The One-Shot Case Study
There is only one group which is subjected to exposure to the
experimental variable. Treatment and posttest are given to the same
group.
a. X, b. O, where: X=treatment and O=post test
Ex. Giving of exam to know the effectiveness of a seminar.

2. Design #2 – The Single Group Pre and Post Test Design

It is sometimes viewed as a quasi-experimental single group
design with pre and posttest variables are more visible.

O = Pretest (1); X = Experimental variable and O = Post test (2)

Here, the contents of Pretest and posttest are the same. An
increase of score in the posttest signifies that the experimental
 Methods of Research
3
Methods of Research (Part III)

variable is effective. Dependent t-test or paired observation is used

to determine the effectiveness of the experimental variable.
Example:
Measuring the effectiveness of a teacher-made module
using a single group.

3. Design #3 – The Static Group Comparison

A group which experienced a treatment (X) is compared with
the other group which has not. The result of the treatment will be the
observed differences between the two groups.

4. Design #4 – The Parallel-Equated Two Group Design

Two group are involved here: The experimental and the
control groups. They should have similarities, for instance the gender,
age, mental ability and the two groups’ pretest results. Pretest and
posttest contents are the same. Dependent and independent t-test is
used in this design.
EXPERIMENTAL GROUP CONTROL GROUP
PRETEST ̸ ̸
PROCESS METHOD ̸
TRADITIONAL ̸
METHOD
POSTTEST ̸ ̸

Example:
Finding the effectiveness of an instructional material.

B. THE TRUE EXPERIMENTAL DESIGNS

These are designs that control nearly all sources of external and
internal validity.

1. Design #5 - The Pre and Post-Test Control Group

Two groups are involved, both of them are randomly selected.
Both groups are given pretest and post test. The experimental group
is given the treatment.

EXPERIMENTAL GROUP CONTROL GROUP

(RANDOMLY (RANDOMLY
SELECTED) SELECTED)
PRETEST ̸ ̸
TREATMENT ̸ x

POSTTEST ̸ ̸

Course Module
 2. Design #6 – The Solomon Four-Group Design
It has the greatest validity and so it has the highest prestige. Here,
two groups are given pre-test and post-test; two groups are given
post-test and two groups do have the experimental treatments.
GROUP #1 GROUP #2 GROUP #3 GROUP #4

Pre-test / /
Post-test / / / /
Treatment / /

FOUR COMPARISONS:
 Pretest and Post test results of Group 1
 Posttest results of Group 1 and Group 2
 Post test results of Group 3 and 4
 Pre test result of Group 2 and Post test result of Group 4

The generated data by this design are computed using the

ANNOVA or the two-way Analysis of Variance on post test scores.

3. Design #7 – The Post-Test-Only Control Group Design

It is a combination of random assignment and the presence of
a control group sources of internal validity. There are two groups and
only one group is receiving the experimental treatment but both of
them are given posttest.

GROUP 1 GROUP 2

Pretest x x
Experimental Treatment / x
Posttest / /

A t-test for two independent random sample is used test the

significant difference between the post test (data analysis)

C. QUASI-EXPERIMENTAL DESIGN
1. Design #8 – The Time-Series Experiment
It is an elaboration of the one-group pre-test-post-test design
involving just one group; It is pretested and post tested four times. It
can be more than 4 depending on the researcher.
1st TIME 2ND TIME 3RD TIME 4TH TIME
Pre-test / / / /
Treatment /
Post test / / / /
(table signifies there is only one group)

Variation can be done by adding additional control group to

the existing one.
1st TIME 2ND TIME 3RD TIME 4TH TIME
Pre-test
1st Group / / / /
2nd Group / / / /
 Methods of Research
5
Methods of Research (Part III)

Treatment
1st Group /
2nd Group /
Post-test
1st Group / / / /
2nd Group / / / /

Generated data analysis in this design involves the study of

test score patterns.

2. Design #9 – Counterbalance Design

In this design, the experimental treatments are arranged in a
different sequence and in a rotation method (in a counterbalanced
method) and ends in Latin square arrangement.

This design only uses post tests since pre tests are not
applicable to the problem.

Let X – treatment ; O= post test

TIME 1 TIME 2 TIME 3 TIME 4

Group A X1; O X2; O X3; O X4; O
Group B X2; O X4; O X1; O X3; O
Group C X3; O X1; O X4; O X2; O
Group D X4; O X3; O X2; O X1; O

3. Design #10- The Non-equivalent control group design

GROUP 1 ( NOT GROUP 2

RAMDOMLY (NOT RANDOMLY
SELECTED, ½ OF THE SELECTED, ½ OF THE
CLASS) CLASS)
Pretest / /
Treatment / /
Posttest / /

It is one of the widely use in educational research. It consists of

two groups, pretest and posttest are both given to the two groups. Only
one group is given the treatment. The subjects are not randomly
chosen. It is commonly used with participants naturally assembled, a
good example is what we have in a classroom. Students in one
classroom are used as the subject. They are divided into two and the
researcher assumes that both groups are equal but the analysis of
Course Module
 covariance may be used in case there are identified effects of
extraneous variables .

D. FACTORIAL DESIGNS
These are extensions of the true experimental designs. These allow
the researcher to measure both the main effect of each independent
variable on the dependent variable and the interaction effects of the said
independent variables. To make the experiment realistic, the researcher
is advised to two or three variables.

In this design, two independent variables are being investigated, each

variable has two levels. For instance, factorial design to employ is 2 x 2
grouping where there are four groups in all.

Example:
 Two kinds of treatment are tried on SHS students in two
mental ability groups.

To evaluate whether the effects of an experimental variable are

generalizable across all levels of a control variable or whether or not the
effects are specific levels of the control variable is the purpose of a factorial
design. Generated scores in the factorial designs of experiments are
analyzed using analysis of variance.

References
Adanza, Estela G., et al., (2009), Methods of Research: A Primer, Manila: Rex
Bookstore
De Ocampo-Acero, Victorina, Leuterio, Florida C., (2006), Methods of
Research, Philippines: Rex Bookstore, Manila
Child Care and Early Education Research Connections. Retrieved from:
https://www.researchconnections.org/childcare/datamethods/preexperime
ntal.jsp