1. Course Supervisor : Dr. H. Nursamsu. M.Pd
Research Methodology in ELT
EXPERIMENTAL RESEARCH
Afida Munawati
Esti Lailatul F
Misroh Anik
Post Graduate Students of TBI B
IAIN TULUNGAGUNG
2. Defining Experimental Research
Experimental research is a powerful research
method to establish cause and effect relationship
(Borg, W.R., Gall, M.D. 1989 : 639)
An Experimental is a scientific in which the
researcher one or more independent variables,
controls any other relevant variables, and
observes the effect of the manipulations on the
dependent variable (Ary, Jacob L.C.,Sorensen.,C :
2010 : 265)
3. Experimental
research
Single or more
Independent variable
True experimental research
Quasi experimental
Pre- experimental
Factorial design(at least 2
independent variable)
TYPES OF EXPERIMENTAL RESEARCH
4. 1. True Experimental Research
The researcher has access to random selection of the
samples and random assignment of the samples into
experimental and control groups to ensure the equivalence of
groups and to control for many interfering variables that might
otherwise contaminate the result of the investigation (Charles,
C.M. 1995:247)
True experimental designs (also called randomized designs)
use randomization and provide maximum control of
extraneous variables.
5. Design of True experimental :
True Experimental Research
The pretest -posttest control
group
The posttest only control
group
The Solomon three group
6. • Group pretest independent V post test
• E Y1 x Y2
• C Y1 - Y2
Pretest-posttest control group
• Group independent V posttest
• E X Y2
• C - Y2
Posttest only control group
• Grouppretest Independent V posttest
• E Y1 X Y2
• C1 Y1 - Y2
• C2 - X Y2
Solomon Three Group
7. 2. Quasi Experimental Research
Quasi-experimental designs lack randomization but employ
other strategies to provide some control over extraneous
variables.(Ary Jacobs L,C , Sorensen C. 2010:302)
When the researcher can only assign randomly different
treatments to two different classes, the researcher uses
quasi- experimental research design (Charles, C.M. 1995 :
247)
8. Types of quasi experimental
Non randomized/non
equivalent control group
Exp group –pre– treatment – post
Control group –pre------------ pottest
Time series Design
Exp group –01--02—03—X-01—02--03
9. The advantages :
1. More frequently used
2. More suitable for real world natural setting than true
experimental
Disadvantages :
1. No control over extraneous variable
2. This design weak for the establish of causal
relationship between independent and dependent
variable.
10. 3. Pre Experimental Research
This design very weak, because the researcher has
very little control over the treatment.
Sometimes without control group.
The group give a pre-test before the experimental
treatment.
After the treatment is finished, the post test is
administrated to see the achievement.
Compute the different score.
11. Types of pre experimental :
Pre experimental
One group
pretest-posttest
Exp group – pretest-
treatment – post test
Static group
Exp group- treatment – post
test
12. Advantages :
Very simple to conduct.
Suitable for beginner.
Disadvantages :
very weak experimental design
little control over the researcher
13. 4. FACTORIAL DESIGN
Researcher manipulating an experimental variable measure the
effect of the independent variables across different sub groups in
the samples like the effect on different age groups of the
samples, different intelligence groups of the samples, different
sex, etc.
This design is useful when there are more than two independent
variables, called factors to be tested.
14. The procedure of Factorial Research
1. Students are randomly Selected as a sample from population.
2. Members of the sample are selected randomly and assigned into two
groups.
3. One of the groups is randomly assigned into experimental group and
other group into an control group.
4. Give experimental treatment for experimental group, and control
treatment for control group.
5. Measure the achievement of two group.
6. The average score Compered score statistically.
7. If the turns out that the experimental group achieve significantly
higher than the control group, then the average scores of the male
experimental group, female experimental group, high intelligence
experimental group, moderate intelligence experimental group are
compared with the groups of the same characteristics with two
group.
15. Variable 2 Variable 1
(X2) (X1)
Treatment A
Treatment B
Level 1 cell 1 cell
3
Level 2 cell 2 cell 4
To illustrate : an experimenter is interested in comparing the effectiveness
of two types of teaching methods A and B on the achievement of ninth grade
students.
16. Threats to Validity in Experimental Research
The interfering variable that may threaten to experimental validity are :
1. History
2. Maturation
3. Testing
4. Instrumentation
5. Statistical regression
6. Experimental mortality
7. Selection Maturation interaction
8. Experimental treatment diffusion
9. Compensatory rivalry by the controlling group
10. Compensatory equalization of treatments
11. Resentful demoralization of the control group
17. 1. History
History is a specific events or condition occurring at the
same time that treatment being applied.
E.g : the different development of motivation and attitude of
the students resulted from process of experimental
research
18. 2. Maturation
Maturation is related to biological and psychological
development of the subject of research.
Maturation interferes the result of experiments if it affects
student’s cognitive ability in learning in experimental
group faster than in the control group.
The different biological and psychological may happen
because of socioeconomic level, socioeconomic
background
19. 3. Testing
Using pretest and posttest
The pretest gives interfering variables which is very
difficult to control
The interfering variables can be controlled by using
equivalent tests for the posttests.
20. 4. Instrumentation
Instrument used to measure pre treatment and post treatment.
The instruments should not any changes during a study to avoid
difficulties in controlling the effect.
E g : a teacher should not use a multiple choice pretest and an
essay posttest.
21. 5. Statistical Regression
Statistical regression refers to the well known tendency
for subjects score in the pretest to score closer to the
mean on a posttest.
If in the regression dealing with extreme scores, the best
solution is by randomly assign to two different treatment.
22. 6. Experimental Mortality (Attrition)
Attrition happen when the number of the members
of two group are not equal
This differential loss may result in differences on
the outcome measure even in the absence of
treatment (Ary. 2010: 279)
23. 7. Selection Maturation interaction
The different age of the subject of
research affect the result of experiment.
Eg : the experiment treatment and the
control treatment are conducted at
different school which have different
admission policy for average age for their
students.
24. 8. Experimental treatment diffusion
Diffusion occur when participants of the control group
borrow and use the instruction material in the experimental
group.
Some members of the control group do the assignments
done the members of experimental group.
The teacher of the control group also uses instructional
material in the experimental group.
25. 9. Compensatory rivalry by the controlling group
The members of control group have highly motivated to
compete with experimental group.
The lack of difference can be attributed to the control
group’s rather than to treatment effects
No differences the average posttest score between control
group and experimental group.
26. 10. Compensatory equalization of treatments
Compensatory equalization of treatments refers to a situation
where one group receives more “desirable goods and
services” .
27. 11. Resentful demoralization of the control group
Members in the control group perceive that the members of
experimental group more desirable treatment.
Members in the control group demoralize themselves so in the
posttest perform lower than normal.
28. Ex post Facto research
1. Ex post facto research measures the relationship between
two variables without manipulating the X variables.
2. Ex post facto research start from identifying the existing
conditions in X-variable (e.g. sex) and preceeds to finding
the differences in Y-variable.
3. e.g : we conduct an Ex – post Facto research to find out
whether sex differences relate to different result in
achievement
29. CONCLUDING REMARK
1. Experimental research used to measure cause – effect
relationship of variable.
2. The problem with the experimental research is in controlling
interfering variables that may lower the validity of the research
findings.
3. Ex- Post Facto Research or Causal Comparative research show
certain relationship between variables but does not proof cause
and effect between relationship.
30. REFERENCES
Ary, Donald; Jacobs,Lucy chaesar & Rezaviah, Ashgar. 2002. Introduction to
research Education. Belmont,CA:Wadsworth.
Ary,Donald; Jacobs,Lucy Cheser &Sorensen,Christine K 2010. Introduction to
Research in Education Eighth Edition. Wadsworth.Belmont,CA:USA
Latief, Muhammad Adnan. 2012. Research Methods on Language Learning
An Introduction: An Introduction. Malang:Universitas Negeri Malang Press.
Wiersma, William. 1991. Research Methods in Education (fifth edition).
Boston : Allyn and Bacon.