Sie sind auf Seite 1von 7

Al-Rashed 1

Zaher Al-Rashed

Prof. Kotob

Research Methodology

21 June 2010

Experimental Research

An experimental is a research situation in which at least one


independent variable, called the experimental variable, is
deliberately manipulated or varied by the researcher.

- The meaning of Experimental Design

An experimental design is the structure by which variables


are positioned, arranged, or built into the experiment. The design
includes the independent variable (s), which must include the
experimental variable (s) and possibly other variables, such as
organismic variables. Experimental designs are often diagramed
with symbols to indicate the arrangements of the variables and
conditions. Why bother with experimental design? It enables the
researcher to interpret and understand the data of an
experiment.
Al-Rashed 2

- Experimental variables

Independent variables can take different forms, such as


organismic variables that indicate different treatments or
procedures to be administered to the participants in the research.
An experimental variable is an independent variable, but not all
independent variables are experimental variables. Any number of
independent variable can serve as experimental variables. The
levels of experimental variables are sometimes called the
experimental treatments.

- Use of the Term Subject

In research generally, but especially in experiment, the term


‘subject’ is used to mean someone who participates in an
experiment. Subjects are the participants in the experiment, those
who receive the experimental treatment. The symbol ‘S’ is used to
designate a subject.

- Criteria for a Well-designed Experiment

1- Adequate experiment control—the experimental design is


so structured that if the experimental variable has any
effect, it can be detected.
2- Lack of artificiality—this means that we need a real
classroom for an educational experiment.
3- Basis of comparison—a control group is included.
4- Adequate information from the data—be succinct.
5- Uncontaminated data—errors in applying an experiment
lead to corrupt data.
Al-Rashed 3

6- No confounding of relevant variables—no effects of any


outer dependent variables are taken in consideration.
7- Representativeness— the keyword here is randomness.
8- Parsimony—let your design be as simple as possible.

- Experimental Validity

The criteria of a well-designed experiment can be


summarized as the characteristics that enhance experimental
validity.

Experimental validity is of two types, internal and external.


Internal validity is minimum control necessary to interpret the
result. External validity deals with the extent of generalizability of
the result.

- Threats to experimental Validity

Experimental validity must be considered in the context of


each specific experiment. Attaining validity is not an all-or-nothing
outcome. Possible limits to validity should be recognized and
countered through the design and the way the experiment is
conducted. We have got two types of validity, internal and
external.

Internal validity

1- History—unanticipated events occurring while the


experiment is in progress that affect dependent variable. For
example, knocking on the door of the class during an
experiment.
Al-Rashed 4

2- Maturation—processes operating within the subjects as a


function of time. For instance you reach a period of time
that you cannot proceed due to students’ fatigue.
3- Testing— tests should not affect each other.
4- Instrumentation—the misuse of instruments causes
deformation.
5- Statistical regression—an effect caused by a tendency for
subjects selected on the basis of extreme scores to regress
toward an average performance on subsequent tests.
6- Differential selection of subjects—subjects are randomly
assigned.
7- Experimental morality or differential loss of subjects—an
effect due to subjects dropping out of the experiment on a
non-random basis.
8- Selection—subjects should not be of more than one level,
for example.

External validity

1- Interaction effect of testing—pretesting sometimes affects


the results of the experiment.
2- Interaction of effects of selection biases and the
experimental treatment—subjects are randomly assigned.
3- Reactive effects of experimental arrangements—subjects
should not know that they are participating in the
experiment.
4- Multiple-treatment interference—do not conduct an
experiment on subjects that have been subjected to any
other experiment.
Al-Rashed 5

- Posttest-Only Control Group Design

Pretest refers to a measure or test given to the subjects prior


to the experimental treatment.

Posttest is a measure taken after the experimental treatment


has been applied.

The posttest-only control group in its simplest form involves


just two groups, the group that receives the experimental
treatment and the control group. A control group is a group that
does not receive an experimental treatment.

The posttest-only control group design contains as many groups


as there are experimental treatments, plus a control or
comparison group. Subjects are measured only after the
experimental treatments have been applied.

- Pretest-posttest Control Group Design

The subjects are randomly assigned to the two or more


groups and tested just prior to the experiment on a supposedly
relevant antecedent variable, possibly a second form of the test
that measures the dependent variable. What is gained by
pretesting? It may be that the pretest score can be used as a
statistical control in the analysis.

The pretest-posttest control group design contains as many


groups as there are in experimental treatment, plus a control
group. Subjects are measures before as well as after receiving the
experimental treatment.
Al-Rashed 6

- Solomon Four-Group Design

The Solomon design is called the 2 X 2 factorial because it has


two independent variables. The advantages of factorial design
over simpler designs are generally twofold: factorial design
provides the economy of a single design rather than separate
design for each of the independent variable, and it allows the
researcher to investigate the interactions between the variables.

Interaction in an experiment is an effect on the dependent


variable such that the effect of one independent variable changes
over the level of another independent variable. The simplest type
of interaction is that of two variables interacting. This is
sometimes called a first-order interaction. The Solomon four-
group design is a combination of the posttest-only control design
group design and the pretest-posttest control group design.

- Factorial Design

Two or more independent variables are included in the


experiment. The basic construction of a factorial design is
combination of with the levels of the other independent variables.

- Repeated measure Design

Repeated measure designs are designs in which the same


subject is measured more than once on the dependent variable.

The simplest form of a repeated measure design would consist of


administering all the experimental treatments to all subjects.
Al-Rashed 7

- Designs Extended in Time

Designs can be extended by taking additional observation on


the groups. Such observations provide information about possible
delayed effects of the experimental variables and about the
duration of an effect.

- Time Series Design

Time series designs involve repeated measurement with an


experimental treatment inserted between two of the
measurements.

- Summary

The distinguishing characteristic of experimental research is


the manipulation of variables. The experimental design provides
the structure for the experiment in which the variables are
deliberately manipulated and controlled by the researcher. A truly
sophisticated experimenter need only come up with an
experimental design that will do the job – meet the objectives of
the researcher and adequate for testing the hypotheses.

Das könnte Ihnen auch gefallen