Experimental

Research

Prof. Regie P. De Jesus

 Introduction
 Experimental is most scientifically
sophisticated research method.
 It is defined as ‘observation under controlled conditions’.

 Experimental research design are concerned with

examination of the effect of independent variable on the
dependent variable, where the independent variable is
manipulated through treatment or intervention(s), & the
effect of those interventions is observed on the dependent
variable.
Experimental research
• An inquiry on cause-and effect relationships

• The researcher manipulates independent variables (e.g.,

type of treatment, teaching method, communication
strategy) and measures dependent variables (anxiety level,
English comprehension, satisfaction)

• The independent variable is CONTROLLED or set by

the researcher.

• The dependent variable is MEASURED by the

researcher.
 Concept…
True experimental research designs
are those where researchers have
complete control over the
extraneous variables & can predict
confidently that the observed effect
on the dependable variable is only
due to the manipulation of the
independent variable.
Essential Characteristics…
A true experimental research design must
essentially consist of the following three
characteristics:
 Manipulation

 Control

 Randomization

 Validity
 MANIPULATION
 Manipulation refers to conscious control of the
independent variable by the researcher
through treatment or intervention(s) to observe
its effect on the dependent variable.
 In other words, it is a conscious act by the
researcher, where he or she varies the
independent variable & observes the effect that
manipulation has on the dependent variable of
interest.
 For example, a researcher is conducting a study
on efficacy of cholrhexidine mouthwash on the
prevention of ventilator-associated pneumonia
(VAP) among patients admitted in ICUs.
 In this example, chlorhexidine mouthwash is the
independent variable, which is manipulated by the
researcher, & is used as an intervention for the
experimental group, while the control group is kept
deprived of it to observe its effect on the incidence
of VAP.
 CONTROL
 Control refers to use of control group & controlling
the effects of extraneous variables on the
dependent variable in which researcher is interested.
 The subject in the control & experimental groups are similar
in number & characteristics, but the subject in the control
group receive no experimental treatment or any intervention
at all.
 The experimental group receives the planned treatment or
intervention & a comparison is made with the control group
to observe the effect of this treatment or intervention.
 RANDOMIZATION
 Randomization means that every subject has an equal
chance of being assigned to experimental or of study
subjects on a random basis.
 Chances of systemic bias is eliminated.
 Randomization is used in true experimental research design to
minimize the threat of internal validity of the study & to
eliminate the effect of extraneous variables on dependent
variables.
 On average the characteristics of the subject in
experimental & control groups are similar, thus influence of
extraneous variables on dependent variable is eliminated by
dispersing the variability of the subject characteristics equally in
both the groups.
 METHODS OF RANDOMIZATION:

 Random assignment of subject may done with simple flip of a

coin for each subject; if coin lands on its ‘head’, subjects are
assigned to first group & with ‘tail’ subjects are assigned to
control group.
 Another possible method is to write the names of the subjects
on slips of paper & put the slips into a bowl & then draw lots.
The first designated numbers of subjects are placed in one
group, & rest are assigned under another group.
 Thirdly a random table may be used to facilitate the
randomization process. In this method, blind-folded subjects
choose a number from a table of number horizontally (row) or
vertically (columns), till a requisite number is reached for both
experimental & control groups. Computer-assisted random
sequences also may be used for the random assignment of the
subjects
 Techniques of Research Control
1. Control of External Factors
a. Control of Environment – has to be researcher-
structured
b. Control of Time – constancy of data collection at
the same time of the day
 Techniques of Research Control
1. Control of Intrinsic Factors
a. Homogeneity – only subjects with similar
characteristics are included
e.g. A study on effects of age on physical fitness.
Participation is limited only to those within specified
age group
Validity
The researcher is interested in controlling or
altogether removing extraneous variables that
may create spurious or inaccurate results.
Hence, ensuring validity helps to objectively
test hypothesis.
Threats to experimental research
Internal validity
Internal invalidity asks the question, “Are the
measurements I make on my dependent (i.e., the
variable I measure) variable influenced only by the
treatment (IV), or are there other influences which
change it?”

External validity
Degree to which results are influenced by
external factors or population and settings.
Internal validity
History,
Maturation,
Testing,
Instrumentation,
Statistical regression,
Differential selection,
Experimental Mortality,
Selection-maturation interaction
The John Henry effect
History
History refers to events other than the treatment that occur during
the course of an experiment which may influence the post-
treatment measure of treatment effect. If the explosion of the
nuclear reactor in Chernobyl, Ukraine had occurred in the middle
of a six-month treatment to help people reduce their “anxiety of
nuclear power,”

Must occur during the experiment.

Controlled by control group

Maturation (endogenous change)
Subjects change over the course of an
experiment. These changes can be physical,
mental, emotional, or spiritual. Perspective can
change. The natural process of human growth
can result in changes in post-test scores quite
apart from the treatment.

Question: How would a “control group”

control this source of internal invalidity
Testing (endogenous change)
A common research design is to give a
group a pre-test, a treatment, and then a
post-test . If the same test is used both
times, the group may show an improvement
simply because of their experience with the
test. This is especially true when the
treatment period is short and the tests are
given within a short time.
Instrumentation
If you use different tests for pre- and post-
measurements, then the change in pre- and
post-scores may be due to differences
between the tests rather than the
treatment.
The best remedy is to use randomization
and a post-test only design.
Statistical regression (endogenous change)
Statistical regression refers to the tendency of extreme
scores, whether low or high, to move toward the average
on a second testing. Subjects who score very high or very
low on one test will probably score less high or low when
they take the test again.
That is, they regress toward the mean.

Do not study groups formed from extreme scores. Study the

full range of scores.
Differential selection/ Selection Bias
If we select groups for “treatment” and
“control” differently, then the results may be
due to the differences between groups
before treatment.

Randomization solves this problem by

statistically equating groups.
E xperimental mortality
Experimental mortality, also called “attrition,”
refers to the loss of subjects from the experiment.
If there is a systematic bias in the subjects who
drop out, then posttest scores will be are biased.
For example, if subjects drop out because they are
aware that they’re not improving as they should,
then the post-test scores of all those who complete
the treatment will be positively biased. Your results
will appear more favorable than they really are.
How does use of a control group solve the
problem of attrition?
Selection-Maturation Interaction
of Subjects
Interaction means the mixing or combining of separate elements. If you draw
a group of subjects from one school to serve as the treatment group, and a
second group from a different schools to serve as a control, you could well
find -- beyond the simple problem of selection differences (“Are the two
groups equivalent?”) -- a mixing of selection and maturation factors to
compound the extraneous influence on your measurements. For example, if
the two schools differ in the average age of their members, they may well
respond to the treatment differently due to inherent maturational factors.

Randomly selecting all subjects from a defined population solves

this problem.
The John Henry Effect/ Compensatory
Rivalry
John Henry, the legendary “steel drivin’ man,” set himself to prove he
could drive railroad spikes faster and better than the newly invented
steam-powered machine driver. He exerted himself so much in trying
to outdo the "experimental" condition that he died of a ruptured
heart. If subjects in a control group find out they are in competition
with those in an experimental treatment, they tend to work harder.
When this occurs, differences between control and treatment groups
is decreased, minimizing the perceived treatment effect.
Threats to external
validity
1.Hawthorne effect
2.Halo effect
3.Multiple Treatment Effect
Hawthorne Effect
It occurs when study participants respond in a
particular manner, or there is obvious change of
behavior because they are aware that they are
being observed.
Halo effect
Tendency of the researcher to rate the subject high
or low because of the impression he has on the
latter
Remedy to minimize threats
Double Blind Method – neither the subject nor the
person administering the treatment or placebo
knows what is being administered

Double Observer Method – two observers record

the subjects performance
Multiple Treatment Effect
Normally we find a single treatment in an
experiment. If, however, an experiment exposes
subjects to, say, three treatments (A, B, and C) and
test scores show that treatment C produced the best
results, one cannot declare treatment C the best. It
may have been the combination of the treatments
that led to the results.
Treatment C, given
alone, may produce different results.
Different types of experimental research can be conducted
depending on the nature of subjects and the instruments,
and the way data are collected and analyzed.

- Will there be a control group?

- How many subjects will there be?
- Will the subjects be randomly selected?
- Will each group be pretested?
- How will the obtained data be analyzed?
- What factors may affect the internal validity?
- What factors may affect the external validity
 True Experimental
Group Design
TYPES OF TRUE EXPERIEMNTAL DESIGN
True Experiential Design

Crossover
Post-test
only

Randomized
Pretest block
post-test
only

Solomon 4 Factorial
groups
 POST-TEST-ONLY CONTROL DESIGN:
 Composed of two randomly assigned group, i.e. experimental
& control, but neither of which is pretested before the
implementation of treatment on the experimental group.
 In addition, while treatment is implement on the experimental
group only, post-test observation is carried out on both the
group to assess the effect of manipulation.
 This design can be helpful in situations where it is not
possible to pretest the subjects.
 For example, to study the effect of an educational
intervention related to urinary incontinence on the
subsequent help-seeking behavior of older adults.
 R O1
X
R O2
Exp. group Treatment Post-test

Random
assignment

Control group Post-test

 PRETEST-POST-TEST-ONLY DESIGN
 In this research designs, subjects are randomly assigned to
either the experimental pr the control group.
 The effect of the dependent variable on both the groups is
seen before the treatment (pretest).
 Later, the treatment is carried out on experimental group only,
& after-treatment observation of dependant variable is made
on both the groups to examine the effect of the manipulation
of independent variable on dependant variable.
 For example, such a design could be used for ‘an
experimental study the effectiveness of the use of guava
leaves for external flushing to post partum patients to
prevent infections
 R O1 O2
X O4
R O3

Exp. Post-
pretest Treatment
group test

Random
assignment

Control
pretest Post-test
group
 SOLOMON FOUR-GROUP DESIGN
 There are two experimental groups (experimental group 1 &
experimental group 2) & two control groups (control group 1 &
control group 2).
 Initially, the investigator randomly assigns subjects to the
four groups.
 Out of the four groups, only experimental group 1 & control
group1 receives the pretest, followed by the treatment to the
experimental group 1 & experimental group 2.
 Finally, all the four groups receive post-test, where the effects of the
dependant variables of the study are observed & comparison is
made of the four groups to assess the effect of independent
variable (experimental treatment) on the dependant variable.
 Cont…

Inthis, experimental group 2 was observed at one

occasion, & that score should be similar to average
scores of those in experimental & control groups.
To estimate the amount of change in experimental
& control group 2, the average test scores of
experimental & control groups 1 are used as
baseline
The solomon four-group design is believed to be
most prestigious experimental research
design, because it minimizes the threat to internal
& external validity.
 Exp. Group 1 pretest treatment Post-test

pretest Post-test
Control Group 1

Random
assignment

Exp. Group 2 treatment Post-test

Control Group 2 Post-test

 FACTORIAL DESIGN
 In factorial design, researcher manipulates two or more independent variables
simultaneously to observe their effects on the dependent variables.
 This design is useful when there are more than two independent variables,
called factors to be tested.
 For example, a researcher wants to observe the effect of two different
protocols of mouth care on prevention of VAP when performed at different
frequencies in a day.
 This design also facilitates the testing of several hypothesis at a
single time.
 Typical factorial design incorporates 2X2 or 2X3 factorial, but it can be in any
combination.
 The first number (α) refers to the independent variables or the type of
experimental treatments, & the second number (β) refers to the level or
frequency of the treatment.
Frequency Protocols of the mouth care
of mouth Chlorhexidine Saline (α2)
care (α1)

4 hourly (β1) α1….β1 α2….β1

6 hourly (β2) α1….β2 α2….β2
8 hourly (β3) α1….β3 α2….β3
 RANDOMIZED BLOCK DESIGN
 Control of inherent differences between experimental
subjects & differences in experimental conditions is one of
the difficult problems faced by researcher in biological
sciences.
 When there are a large number of experimental comparison
groups, the randomized block design is used to bring
homogeneity among selected different groups.
 This is simple method to reduce the variability among the
treatment groups by a more homogeneous combination of
the subjects through randomized block design.
 For example, a researcher wants to examine the
effects of three different antihypertensive drugs on
patients with hypertension.
 In this example, to ensure the homogeneity
among the subjects under
treatment, researcher randomly places the subjects
in homogeneous groups (blocks) like patients with
primary hypertension, diabetic patients with
hypertension, & renal patients with hypertension .
 Blocks
Types of
antihypertensive Patients with Diabetic patients Renal patients
drugs primary with hyper with
hypertension (I) tension (II) hypertension
(III)

A A, I A, II A, III
B B, I B, II B, III
C C, I C, II C, III
CROSSOVER/ COUNTERBALANCED DESIGN
 In this design, subjects are exposed to more than one
treatment, where subjects are randomly assigned to different
orders of treatment.
 It is also known as ‘repeat measures design’.
 This design is more efficient in establishing the highest possible
similarity among subjects exposed to different conditions, where
groups compared obviously have equal distribution of
characteristics.
 Though crossover design is considered as an extremely powerful
research design, sometimes it is not effective because when
subjects are exposed to two different conditions, their responses of
the second condition may be influenced by their experience in the
first condition.
 For example, when we are comparing the effectiveness of the
chlorhexidine mouth care protocol on group I & saline mouth
care protocol on the subjects of group II.
 Later, the treatment is swapped, where group I receives the
saline mouth care & group II receives chlorhexidine. In such
studies, subjects serve as their own control.
Groups Protocols of the mouth care
Group I Chlorhexidine (α1) Saline (α2)

Group II Saline (α2) Chlorhexidine (α1)

Group1 X1 O X2 O
Group2 X2 O X1 O
QUASI – EXPERIMENTAL
RESEARCH DESIGN

The term quasi- (pronounced

kwahz-eye) means almost,
near, partial, pseudo, or
somewhat.

“almost true experiments,”

 QUASI – EXPERIMENTAL
RESEARCH DESIGN
 Quasi-experimental research design involves the manipulation of
independent variable to observe to effect on dependent variable,
but it lacks at least one of the two characteristics of the true
experimental design; randomization or a control group.
 Quasi-experimental designs are generally used to establish
the causality (effect of independent variable on dependent
variable) in situations where researchers are not able to randomly
assign the subjects to groups or for various reasons no control
group is available for an experimental study.
TYPES OF QUASI – EXPERIMENTAL
 the time series,

 the nonequivalent control group design,

 the counterbalanced design.

Time Series
O1 O2 O3 O4 X O5 O6 O7 O8
 This design is useful when the experimenter wants to
measure the effects of a treatment over a long period of time.
 The experimenter would continue to administer the treatment
& measure the effects a number of times during the course of
the experiment.
 Generally it is a single-subject research, in which the
researcher carries out an experiment on an individual or on a
small number of individuals, by alternating between
administering & then withdrawing the treatment to determine
the effectiveness of the intervention.
 O1 O2 O3 O4 X O5 O6 O7 O8

Exp. O2 O3
group O1 O2 O3 Treatment O1
A researcher might assess pain levels of a group of
patients with low-back pain. After 3 weeks of pain
assessment, subjects are taught special exercises
to reduce that pain. During the next 3 weeks, pain
levels would again be measured.
 Measuring a child’s school performance on a
weekly basis, & then introducing a new teaching
technique. Then again measuring on a weekly
basis.
NONRANDOMIZED CONTROL GROUP DESIGN
 AKA‘nonequivalent control group design’.
 identical to the pretest-posttest control group design, except
there is no random assignment of subjects in experimental
& control groups.
 In this design, experimental & control groups are selected
without randomization, & dependent variables are observed in
experimental as well as control groups before the intervention.
 Later, the experimental group receives treatment & after that
posttest observation of dependant variables is carried out for
both the groups to assess the effect of treatment on
experiment group.
 O1 X O2
---------------------
O3 O4

Exp. group Pretest Treatment Post-test

Control group Pretest Post-test

SIMILAR TO PRETEST-POSTTEST WITH THE

EXCEPTION OF RANDOMIZATION
 PRE – EXPERIMENTAL
RESEARCH DESIGN
 This research design is considered very
weak, because the researcher has very little
control over the experiment

TYPES OF PRE-EXPERIMENTAL
RESEARCH DESIGN:
o One-shot case design

o One-group pretest-posttest design

o Static-group comparison design:

One-shot case studies:
 In this research design, a single experimental group is
exposed to a treatment & observations are made after the
implementation of that treatment.
 There is no random assignment of subjects to the
experimental group & no control group at all.

Exp. group Treatment Post-test

One-group pretest-posttest design:
 It is the simplest type of pre-experimental design, where
only the experimental group is selected as the study
subjects.
 A pretest observation of the dependent variables is made
before implementation of the treatment to the selected
group, the treatment is administered, & finally a posttest
observation of dependent variables is carried out to assess
the effect of treatment on the group.
 Some researcher also argue this design as sub type of
quasi-experimental research design. However in absence
of both randomization & control group.
Static-group comparison design
At least two groups are involved. After one group
receives the treatment, all groups are posttested.

This design has better control over most of the

variables
Thank Y
ou