Professional Documents
Culture Documents
A research design is a systematic plan to study a scientific problem. The design of a study
The research design refers to the overall strategy that you choose to integrate the different
components of the study in a coherent and logical way, thereby, ensuring you will effectively
address the research problem; it constitutes the blueprint for the collection, measurement, and
analysis of data. Research problem determines the type of design you can use.
Non experimental
Experimental
A design in which the researcher is a passive agent, who observes, measures, and describes a
According to Goodwin (2008) and McBurney (2001) non experimental research design includes:
Observational research
Archival research
Case study
Survey research
Correlational research
Observational, archival, survey research methods simply obtain information about a current
characteristics such as: age, sex, education, occupation, marital status, income etc
A systematic research study in which the investigator directly varies some variables and holds all
other factors constant and observes the result of systematic variation (Goodwin, 2008)
There are four goals of research in psychology: description, prediction, explanation, and
behavior, often in natural settings. Survey research methods allow psychologists to describe
people’s attitudes and opinions. Psychologists are able to make predictions about behavior
and mental processes when they discover measures and observations that covary (correlations).
Description and prediction are essential to the scientific study of behavior, but they are not
sufficient for understanding the causes of behavior. Psychologists also seek explanation—the
phenomenon.
The best available research method for identifying causal relationships is the
experimental method. Researchers conduct experiments to test hypotheses about the causes of
changes behavior. One of the primary reasons that psychologists conduct experiments is to make
empirical tests of hypotheses they derive from psychological theories. The self-correcting
use to understand the causes of the ways we think, feel, and behave.
Well-conducted experiments also help to solve society’s problems by providing vital
information about the effectiveness of treatments in a wide variety of areas. This role of
experiments has a long history in the field of medicine for example, near the beginning of the
19th century, typhoid fever and delirium tremens were often fatal. The standard medical practice
at that time was to treat these two conditions by bleeding, purging, and other similar “therapies.”
In an experiment to test the effectiveness of these treatments, researchers randomly assigned one
group to receive the standard treatment (bleeding, purging, etc.) and a second group to receive
nothing but bed rest, good nutrition, and close observation. The group given the standard medical
treatment of the time did worse than the group left untreated. Treating such conditions using
early-19th-century practices was worse than not treating them at all. Experiments such as these
contributed to the insight that many medical conditions are self-limited: The illness runs its
Professor Fisher (as cited in kathori, 2004) has enumerated three principles of experimental
designs:
(Research Methodology)
According to the Principle of Replication, the experiment should be repeated more than once.
Thus, each treatment is applied in many experimental units instead of one. By doing so the
statistical accuracy of the experiments is increased. For example, suppose we are to examine the
effect of two varieties of rice. For this purpose we may divide the field into two parts and grow
one variety in one part and the other variety in the other part. We can then compare the yield of
the two parts and draw conclusion on that basis. But if we are to apply the principle of
replication to this experiment, then we first divide the field into several parts, grow one variety in
half of these parts and the other variety in the remaining parts. We can then collect the data of
yield of the two varieties and draw conclusion by comparing the same. The result so obtained
will be more reliable in comparison to the conclusion we draw without applying the principle of
replication. The entire experiment can even be repeated several times for better results.
Conceptually replication does not present any difficulty, but computationally it does. For
require a three-way analysis of variance since replication itself may be a source of variation in
the data. However, it should be remembered that replication is introduced in order to increase the
precision of a study; that is to say, to increase the accuracy with which the main effects and
the effect of extraneous factors by randomization. In other words, this principle indicates that we
should design or plan the experiment in such a way that the variations caused by extraneous
factors can all be combined under the general heading of “chance.” For instance, if we grow one
variety of rice, say, in the first half of the parts of a field and the other variety is grown in the
other half, then it is just possible that the soil fertility may be different in the first half in
comparison to the other half. If this is so, our results would not be realistic. In such a situation,
we may assign the variety of rice to be grown in different parts of the field on the basis of some
random sampling technique i.e., we may apply randomization principle and protect ourselves
against the effects of the extraneous factors (soil fertility differences in the given case). As such,
through the application of the principle of randomization, we can have a better estimate of the
experimental error.
The Principle of Local Control is another important principle of experimental designs. Under it
the extraneous factor, the known source of variability, is made to vary deliberately over as wide a
range as necessary and this need to be done in such a way that the variability it causes can be
measured and hence eliminated from the experimental error. This means that we should plan the
experiment in a manner that we can perform a two-way analysis of variance, in which the total
variability of the data is divided into three components attributed to treatments (varieties of rice
in above mentioned case), the extraneous factor (soil fertility) and experimental error.* In other
words, according to the principle of local control, we first divide the field into several
homogeneous parts, known as blocks, and then each such block is divided into parts equal to the
number of treatments. Then the treatments are randomly assigned to these parts of a block.
Dividing the field into several homogenous parts is known as ‘blocking’. In general, blocks are
the levels at which we hold an extraneous factor fixed, so that we can measure its contribution to
the total variability of the data by means of a two-way analysis of variance. In brief, through the
principle of local control we can eliminate the variability due to extraneous factor(s) from the
experimental error.
Experimental control allows researchers to make the causal inference that the
An experiment has internal validity when it fulfills the three conditions required for
alternative causes. When confounding occurs, a plausible alternative explanation for the
observed covariation exists, and therefore, the experiment lacks internal validity.
Plausible alternative explanations are ruled out by holding conditions constant and
balancing.
A true experiment involves the manipulation of one or more factors and the measurement
(observation) of the effects of this manipulation on behavior. The factors the researcher
must have at least two levels (also called conditions). One level may be considered the
“treatment” condition and a second level the control (or comparison) condition. Often,
more than two levels are used for additional comparisons between groups. The measures
used to observe the effect (if any) of the independent variables are called dependent
variables. One way to remember the distinction between these two types of variables is to
understand that the outcome (dependent variable) depends on the independent variable.
Experiments are effective for testing hypotheses because they allow us to exercise a
be able to state with confidence that the independent variable caused the observed
The most effective independent groups design is one that uses random assignment of subjects
variable. When random assignment to conditions is used, the independent groups design is called
a random groups design. The logic of the design is straightforward. The groups are formed so
as to be similar on all important characteristics at the start of the experiment. Next, in the
experiment itself, the groups are treated the same except for the level of the independent
variable. Thus, any difference between the groups on the dependent variable must be caused by
the independent variable. Block randomization can be used for randomly assigning subjects.
With a randomized block design, the experimenter divides subjects into subgroups called blocks,
such that the variability within blocks is less than the variability between blocks. Then, subjects
Gende Treatment
r Placebo Vaccine
Femal
250 250
e
The table to the right shows a block randomization for a hypothetical medical experiment.
Subjects are assigned to blocks, based on gender. Then, within each block, subjects are randomly
assigned to treatments (either a placebo or a cold vaccine). For this design, 250 men get the
placebo, 250 men get the vaccine, 250 women get the placebo, and 250 women get the vaccine.
It is known that men and women are physiologically different and react differently to
medication. This design ensures that each treatment condition has an equal proportion of men
and women. As a result, differences between treatment conditions cannot be attributed to gender.
This randomized block design removes gender as a potential source of variability and as a
A matched groups design may be used to create comparable groups when there are too few
The matched groups design is a good alternative when neither the random groups design nor the
repeated measures design can be used effectively. The logic of the matched groups design is
simple and compelling. Instead of trusting random assignment to form comparable groups, the
researcher makes the groups equivalent by matching subjects. The matched groups design is
useful only when a good matching task is available. The most preferred matching task is one that
uses the same task that will be used in the experiment itself. For example, if the dependent
variable in the experiment is blood pressure, participants should be matched on blood pressure
prior to the start of the experiment. The matched groups design is an alternative to the random
groups design when only a small number of subjects are available, when a good matching task is
available, and when the experiment requires separate groups for each treatment
The biggest problem with the matched groups design is that the groups are equated only on the
Researchers in many areas of psychology are interested in independent variables that are called
assigning people to Catholic, Jewish, Muslim, Protestant, or other groups. Instead, researchers
“control” the religious affiliation variable by systematically selecting individuals who naturally
extraversion, race, or age are important independent variables in many areas of psychology. It is
important to differentiate experiments involving independent variables whose levels are selected
from those involving independent variables whose levels are manipulated. Experiments
involving independent variables whose levels are selected—like individual differences variables
—are called natural groups designs. The natural groups design is frequently used in situations
in which ethical and practical constraints prevent us from directly manipulating independent
variables. For example, no matter how interested we might be in the effects of major surgery on
subsequent depression, we could not ethically perform major surgery on a randomly assigned
group of introductory psychology students and then compare their depression symptoms with
There are several threats to the internal validity of experiments that involve testing independent
groups. Extraneous variables, such as different rooms or different experimenters, must not be
allowed to confound the independent variable of interest. A more serious threat to the internal
validity of the random groups design is involved when subjects fail to complete the experiment
successfully. Selective subject loss occurs when subjects are lost differentially across the
conditions and some characteristic of the subject that is related to the outcome of the experiment
is responsible for the loss. We can help prevent selective loss by restricting subjects to those
dropping comparable subjects from the group that did not experience the loss. Demand
characteristics and experimenter effects can be minimized through the use of proper
experimental procedures, but they can best be controlled by using placebo control and double-
blind procedures.
An investigator wants to evaluate whether a new technique to teach math to elementary school students is more
effective than the standard teaching method. Using an experimental design, the investigator divides the class
randomly (by chance) into two groups and calls them “group A” and “group B.” The students cannot choose their
own group. The random assignment process results in two groups that should share equal characteristics at the
beginning of the experiment. In group A, the teacher uses a new teaching method to teach the math lesson. In group
B, the teacher uses a standard teaching method to teach the math lesson. The investigator compares test scores at the
end of the semester to evaluate the success of the new teaching method compared to the standard teaching method.
At the end of the study, the results indicated that the students in the new teaching method group scored significantly
higher on their final exam than the students in the standard teaching group.
Repeated Measures Designs
There are times when it is more effective to have each subject participate in all the conditions of
an experiment. These designs are called repeated measures designs (or within-subjects
designs). In an independent groups design, a separate group serves as a control for the group
given the experimental treatment. In a repeated measures design, subjects serve as their own
controls because they participate in both the experimental and control conditions individual
differences among participants cannot be eliminated in the random groups design, but they can
be balanced by using random assignment. Similarly, the practice affects that participants
experience due to repeated testing in the repeated measures designs cannot be eliminated. Like
individual differences in the random groups design, however, practice effects can be balanced, or
averaged, across the conditions of a repeated measures design experiment. When balanced across
the conditions, practice effects are not confounded with the independent variable and the results
Researchers who do experiments with children, the elderly or special populations such as
individuals with brain injuries frequently have a small number of participants available.
The sensitivity of an experiment refers to the ability to detect the effect of the independent
variable even if the effect is a small one. There is usually more variation between people than
there is within people. When the research question involves studying changes in participants’
behavior over time, such as in a learning experiment, a repeated measures design is needed.
Advantages
Each subject serves as own control so that the variability between subjects gets isolated.
Repeat-measure designs are more economical since each subject is own control and so
Disadvantages
Carry-over effect occurs when a treatment is administered before the effects of previous
treatment have worn off. It can be avoided by allowing sufficient time between
treatments.
Latent effect occurs when a treatment can activate the dormant effects of a previous
treatment.
Learning effect occurs in situations where response improves each time a person takes a
test.
There are two types of repeated measures designs (complete and incomplete) that differ in the
The absence of the potential for confounding by individual differences variables is a great
advantage of the repeated measures designs. The repeated testing of participants in the repeated
measures design gives them practice with the experimental task. As a result of this practice,
participants may get better and better at doing the task because they learn more about the task, or
they may get worse at the task because of such factors as fatigue and boredom with repeated
testing in the repeated measures designs are called practice effects. In general, practice effects
should be balanced across the conditions in repeated measures designs so that practice effects
“average out” across conditions. The key to conducting interpretable experiments using the
repeated measures designs is learning to use appropriate techniques to balance practice effects.
The two types of repeated measures designs are the complete and the incomplete design. The
specific techniques for balancing practice effects differ for the two repeated measures designs,
but the general term used to refer to these balancing techniques is counterbalancing. In the
complete design, practice effects are balanced for each participant by administering the
conditions to each participant several times, using different orders each time. Each participant
can thus be considered a “complete” experiment. In the incomplete design, each condition is
administered to each participant only once. The order of administering the conditions is varied
across participants rather than for each participant, as is the case in the complete design.
Block randomization
ABBA Counterbalancing
In its simplest form, ABBA counterbalancing can be used to balance practice effects in the
complete design with as few as two administrations of each condition. ABBA counterbalancing
involves presenting the conditions in one sequence (i.e., A then B) followed by the opposite of
that. ABBA counterbalancing is used when the number of conditions and the number of
repetitions of each condition are relatively small. Although ABBA counterbalancing provides a
simple and elegant means to balance practice effects, it is not without limitations. For example,
ABBA counterbalancing is ineffective when practice effects for a task are not linear. ABBA
counterbalancing is also ineffective when anticipation effects can occur. Anticipation effects
occur when a participant develops expectations about which condition should occur next in the
sequence.
The use of all possible orders, Latin Squares, and random starting orders with rotation are
equally effective in balancing practice effects because all three techniques ensure that each
In the incomplete design, each participant is given each treatment only once. The results for any
one participant, therefore, cannot be interpreted because the levels of the independent variable
for each participant are perfectly confounded with the order in which those levels were
presented. For instance, the first participant in an incomplete design experiment might be tested
first in the experimental condition (E) and second in the control condition (C). Any differences in
the participant’s performance between the experimental and control conditions could be due to
the effect of the independent variable or to the practice effects resulting from the EC order. To
break this confounding of the order of conditions and the independent variable, we can
administer different orders of the conditions to different participants. For example, we could
administer the conditions of our incomplete design experiment to a second participant in the CE
order, testing the control condition first and the experimental condition second. In this way, we
could balance the effects of order across the two conditions using two participants instead of one.
The preferred technique for balancing practice effects in the incomplete design is to use all
possible orders of the conditions. Each participant is randomly assigned to one of the orders.
With only two conditions there are only two possible orders (AB and BA) with three conditions
there are six possible orders (ABC, ACB, BAC, BCA, CAB, CBA).
Selected Orders
The preferred method for balancing practice effects in the incomplete design is all possible
orders. There are times, however, when the use of all possible orders is not practical. Practice
effects can be balanced by using just some of all the possible orders. The number of selected
orders will always be equal to some multiple of the number of conditions in the experiment. For
example, to do an experiment with one independent variable with seven levels, we need to select
7, 14, 21, 28, or some other multiple of seven orders to balance practice effects.
Differential transfer occurs when the effects of one condition persist and influence performance
in subsequent conditions.
Consider a problem-solving experiment in which two types of instructions are being compared in
a repeated measures design. One set of instructions (A) is expected to enhance problem solving,
whereas the other set of instructions (B) serves as the neutral control condition. It is reasonable
to expect that participants tested in the order AB will be unable or unwilling to abandon the
approach outlined in the A instructions when they are supposed to be following the B
instructions. Giving up the “good thing” participants had under instruction A would be the
counterpart of successfully following the admonition “Don’t think of pink elephants!” When
participants fail to give up the instruction from the first condition (A) while they are supposed to
be following instruction B, any difference between the two conditions is reduced. For those
participants, after all, condition B was not really tried. The experiment becomes a situation in
Complex designs are those in which two or more independent variables are studied
simultaneously in one experiment. Complex designs can also be called factorial designs because
pairing each level of one independent variable with each level of a second independent variable.
This makes it possible to determine the effect of each independent variable alone (main effect)
1) It is usually more efficient. Instead of doing one study on the effect of the type of
psychotherapy a patient gets and a second study on the length of therapy, a researcher
2) It represents the complexity of the real world more accurately. For example, in the real
3) It allows you to see whether the effect of one independent variable depends on the level
of other independent variables. For example, maybe some types of psychotherapy are
effective in a short form, but others are not. This is called an interaction
A factorial design is the most common way to study the effect of two or more independent
variables. In a factorial design, all levels of each independent variable are combined with all
levels of the other independent variables to produce all possible conditions. For example, a
researcher might be interested in the effect of whether or not a stimulus person (shown in a
photograph) is smiling or not on ratings of the friendliness of that person. The researcher
might also be interested in whether or not the stimulus person is looking directly at the
camera makes a difference. In a factorial design, the two levels of the first independent
variable (smiling and not smiling) would be combined with the two levels of the second
(looking directly or not) to produce four distinct conditions: smiling and looking at the
camera, smiling and not looking at the camera, not smiling and looking at the camera, and
This would be called a 2x2 (two-by-two) factorial design because there are two independent
variables, each of which has two levels. If the first independent variable had three levels (not
smiling, closed-mouth smile, open-mouth smile), then it would be a 3x2 factorial design. Note
that the number of distinct conditions formed by combining the levels of the independent
variables is always just the product of the numbers of levels. In a 2x2 design, there are four
Design Tables
One way to represent a factorial design is with a design table. The table below represents a 2x2
factorial design in which one independent variable is the type of psychotherapy used to treat a
sample of depressed people (behavioral vs. cognitive) and the other is the duration of that
therapy (short vs. long). (The dependent variable--which is not actually shown in this table--is a
measure of improvement.) Each cell in the table, therefore, represents one of the four distinct
conditions: short behavioral therapy, short cognitive therapy, long behavioral therapy, and long
cognitive therapy. Inside the cells, you can put different things. In this example, it is the number
of participants in each condition. (The symbol n generally refers to the number of subjects in a
condition.) You could also put expected results or actual results (e.g., means and standard
Therapy (B)
Behavioral Cognitive
Short
n = 50 n = 50
Long n = 50 n = 50
In a factorial design, each independent variable can be manipulated between subjects or within
subjects--and this decision must be made separately for each one. In the design above, it makes
sense that participants will receive only one kind of psychotherapy. They will receive either
behavioral or cognitive, not both. And they will receive either short or long, not both. That
explains the "B"s in parentheses after each variable name; they stand for "between subjects."
What this implies is that each participant will be in only one of the four distinct conditions.
In the design below, however, both independent variables are manipulated within subjects. That
is, participants are tested under both quiet and noisy conditions and under both cool and warm
conditions. That explains the "W"s; they stand for "within subjects." This implies that each
Quiet Noisy
Cool n = 50 n = 50
Temperature (W)
Warm n = 50 n = 50
In addition to the between-subjects design and the within-subjects design, it is possible to have a
mixed design, in which one independent variable is manipulated between subjects and the other
is manipulated within subjects. For example, participants might be tested under a quiet condition
and under a noisy condition (so that noise level is a within-subjects variable), but they might be
tested in either a cool room or a warm room (so that temperature is a between-subjects variable).
In this case, noise level would be labeled with a "W" and temperature with a "B."
Data analysis
Data analysis and statistics provide researchers with an alternative to replication for determining
whether the results of a single experiment are reliable and can be used to make a claim about the
The choice among techniques hinges on the number of IVs and DVs and whether some variables
are conceptualized as covariates. Further distinctions are made as to whether all DVs are
measured on the same scale and how within-subjects IVs are to be treated.
Data analysis for a complete design begins with computing a summary score (e.g., mean,
median) for each participant. Descriptive statistics are used to summarize performance across
Dependent t-test
The dependent t-test compares two means, when those means have come from the same entities;
for example, if you have used the same participants in each of two experimental conditions.
The independent t-test compares two means, when those means have come from different groups
of entities; for example, if you have used different participants in each of two experimental
conditions.
The one-way repeated-measures ANOVA compares several means, when those means have
come from the same participants for example, if you measured people’s statistical ability each
assumption: sphericity. This assumption needs to be considered only when there are three or
The one-way independent ANOVA compares several means, when those means have come from
different groups of people. For example; if you have several experimental conditions and have
Mixed ANOVA compares several means when there are two or more independent variables, and
at least one of them has been measured using the same participants and at least one other has
Field, A. (2009)
ANCOVA
ANCOVA is used to adjust for differences among groups when groups are naturally occurring
and random assignment to them is not possible. For example, one might ask if attitude toward
abortion (the DV) varies as a function of religious affiliation.
With ANCOVA, selected and/or pooled group means are adjusted for differences on covariates
before differences in means on the DV are assessed. (Tabachnick & Fidell, 2007)
References
Bordens, K.S. & Abbot, B.B. (2005). Research Design and Methods. 6th ed. Tata Mcgraw hill
Publication
Field, A. (2009). Discovering statistics: Using SPSS. 3rd ed. SAGE Publications Ltd
Goodwin,C.J.(2008). Research in Psychology: Methods and design. 5th ed. John Wily and Sons
Kothori, C.R. (2004). Research Methods: methods and techniques. (2nd ed). New Age
Retrieved on05.6.2014
http://faculty.ksu.edu.sa/Hanan_Alkorashy/Nursing%20management%20489NUR/
Non_experimental_Designs.pdf
Shaughnessy, J.J; Zechmeister, E.B. & Zechmeister, J.S. (2012). Research Methods in
Tabachnick, B.G. & Fidell, L.S. (2007). Using Multivariate Statistics. 5th ed. Pearson Education,
Inc.