Experimental

design

Between- group Within-group

design design

Randomized- Matched- groups

Factorial design
groups design design

More-than two-
Two-randomized
randomized
groups design
groups design

Research design is the blueprint of the procedures that enables the experimenter to test
various hypotheses.
The selection of design is partly dependent upon the number of group of subjects who will be
tested under different conditions of the independent variable.
I. Independent Groups Design
● An independent groups design, also known as between-subjects design, is a
research design where each participant is only exposed to one level of the independent
variable or assigned to different experimental conditions or groups. Subsequently
their responses are compared across these groups.
● This design is called "independent groups" because the groups are independent of
each other; participants in one group do not influence or interact with participants in
other groups.
● Random assignment to conditions is used to form comparable groups by balancing or
averaging subject characteristics (individual differences) across the conditions of the
independent variable manipulation.
● When random assignment to conditions is used, the independent groups design is
called a random groups design.

1. Randomized group design

Subjects are randomly assigned to different groups meant for the different conditions of
the independent variable. The goal of random assignment is to create groups that are
statistically equivalent, meaning they are balanced in characteristics like age, ability,
attitude, or other relevant traits. This helps ensure that any differences observed in the
results are due to the experimental manipulation (independent variable) and not other
factors. The randomized group designs are of two types:
a. Two-randomized groups design: here the subjects are randomly assigned to two groups
only.
• Treatment Group / Condition 1: This group receives the experimental
condition or "treatment”. Alternatively, they receive Condition 1 as defined by
the researcher.
• Control Group / Condition 2: This group does not receive the experimental
treatment or receives a standard condition for comparison or an alternative
Condition 2 that differs from Condition 1 for comparison purposes.
Several factors that could have affected performance on the memory test are held constant in
the experiment except the manipulation of IV for the experimental group. Holding
conditions constant is a control technique that researchers use to avoid confounding. In such
conditions if there are any significant differences between the experimental and control
groups, then we conclude that the IV has significant effect on DV.
Example: A researcher wants to investigate the effects of a new memory training program on
memory performance on graduate students.
● Experimental Group (50 participants): Participants in this group receive the
new memory training program, which consists of memory exercises and strategies
aimed at improving memory performance.
● Control Group (50 participants): Participants in this group do not receive any
memory training and continue with their regular activities.

b. More-than-two-randomized groups design/ Multi- group design: In this design,

participants are randomly assigned to three or more groups, with each group exposed to a
different condition of the independent variable (IV). The aim is to study and compare the
effects of multiple conditions on the dependent variable (DV). Random assignment ensures
that all groups are unbiased and statistically equivalent.
This method is an extension of the two-randomized groups design and is often considered
superior, as it allows researchers to examine multiple treatments or conditions in a single
study. It also reduces the risk of experimenter fatigue when conducting sequential tests, as all
groups are assigned simultaneously.
Suppose a researcher wants to evaluate the effectiveness of four teaching methods on student
performance:
a. Traditional Lecture
b. Group Discussions
c. Online Learning
d. Project-Based Learning
The researcher randomly assigns 400 students into four groups of 100 students each. Each
group is exposed to one teaching method, and their performance (measured through a
standardized test) is compared to determine which method is most effective.

2. Matched groups design

• Also known as Randomized block design.
• All the subjects are first tested on a common task or a pretest measure (also known as
matching variable) and then they are formed into groups (as many needed for the
experiment) on the basis of the performance of the common task or the matching
variable.
• The groups thus formed are said to be equivalent groups.
• Subsequently, the different conditions of the IV are introduced to each group.
• The researcher might also check for equivalent means on the DV before the
experimental treatment. After the treatment is given, if there is a significant difference
in the DV, and other relevant variables are controlled, the difference can be attributed
to the treatment.
• A matched group design ensures that participants with similar characteristics are
grouped into blocks before being randomly assigned to experimental conditions.
This design reduces variability caused by individual differences, making it easier
to detect the effects of the experimental treatment.
• The principle is simple: when subjects are matched based on a characteristic related to
the dependent variable (DV), each block becomes more uniform. Without the
 experimental treatment, we expect the DV to show similar results within a block. Any
significant differences after the treatment can then be confidently attributed to the
treatment itself, not random differences.
Process of Matched groups design:
1. Selection of the matching variable: the matching variable is a characteristic that is
believed to influence the dependent variable (DV) and is used to pair or group
participants. This relationship must be established in previous research or
investigation. The goal is to ensure that the groups are similar on this characteristic,
reducing variability and isolating the effect of the independent variable (IV). The
matching variable should be something that can be assessed accurately and
consistently, such as age, socioeconomic status, or baseline scores.
2. Methods of matching
• Precision match: Participants are matched exactly on the chosen variable(s). E.g.: If
age is the matching variable, a 25-year-old in the experimental group is paired with
another 25-year-old in the control group. Limitation: This method is stringent and
may exclude participants who cannot be perfectly matched.
• Frequency Matching: Ensures that the overall distribution of the matching variable
is the same in both groups. E.g.: If the experimental group has 40% males and 60%
females, the control group is matched to have the same ratio, but individuals are not
paired directly. This method is less restrictive and allows for more participants to be
included.
• Range Matching: Participants are matched within a range of values on the variable.
E.g.: In a study on physical fitness, participants aged 20–25 could be matched as a
group with those aged 20–25 in the other group. It is more flexible than precision
matching.
• Pair Matching: Participants are matched in pairs based on the closest possible scores
on the matching variable. E.g.: If two participants score 75 and 76 on a pretest, they
are paired together and assigned to different groups.
• Propensity Score Matching (Modern Approach): Statistical techniques calculate the
probability (propensity score) of each participant being in the experimental group
based on multiple variables. Participants with similar scores are matched. This is
particularly useful in non-randomized studies.
3. Factorial design
• This method is used when the researcher should manipulate two or more than two
independent variables simultaneously.
• Factorial design may be defined as “Design in which the selected values of two or
more IVs are manipulated in all possible combinations so that their independent
as well as interactive effects upon the DV may be studied”.
• Factorial design has the following three main characteristics:
a. Two or more IVs are manipulated in all possible combinations.
 b. Different subgroups or subjects must serve under every possible combination
of the IVs. As far as possible an equal number of subjects in all subgroups is
proffered, although this is not an essential condition.
c. Enables the experimenter to study the independent effect as well as the
interactive effect of the two or more IVs.
Examples-
2X2 Factorial Design
In this design, there are two independent variables, one with two levels and the other with
two levels. Example a researcher examine the effects of "Noise" (High and Low) and "Level
of illumination" (Low and High) on "Rate of learning".
Let us denote the first independent variable as A and the second independent variable as B.
The dependent variable is the rate of learning.
Participants are randomly assigned to one of four conditions created by the combination of
these levels.
High-noise as well as high-illumination level (A1B1)
Low-noise as well as high-illumination level (A2B1)
High-noise as well as low- illumination level (A1B2)
Low-noise as well as low-illumination level (A2B2)
Main Effects: Researchers analyze whether noise or level of illumination independently
affect rate of learning. For example, they might find that participants who are exposed to low
noise generally have higher rate of learning.
Interaction Effects: Researchers investigate whether the effect of noise on rate of learning
varies based on level of illumination. For instance, they might find that exposure to low noise
enhance rate of learning more for participants with high level of illumination compared to
those with low level of illumination.
2X3 Factorial Design
In this design, there are two independent variables, one with two levels and the other with
three levels. Example a researcher examine the effects of "Coffee Consumption" (No Coffee,
One Cup) and "Hours of Sleep" (6 hours, 8 hours, 10 hours) on "Attention Span". Participants
are randomly assigned to one of six conditions created by the combination of these levels.
No Coffee + 6 hours of sleep
No Coffee + 8 hours of sleep
No Coffee + 10 hours of sleep
One Cup of Coffee + 6 hours of sleep
One Cup of Coffee + 8 hours of sleep
One Cup of Coffee + 10 hours of sleep
Main Effects: Researchers analyze whether coffee consumption or hours of sleep
independently affect attention span. For example, they might find that participants who
consume more coffee generally have better attention span.
Interaction Effects: Researchers investigate whether the effect of coffee on attention span
varies based on hours of sleep. For instance, they might find that two cups of coffee enhance
attention span more for participants with fewer hours of sleep compared to those with more
sleep.
II. Repeated measures design
• Repeated measures design is also called as within groups design or within subjects
design or repeated-treatments design.
• The same group of subjects is treated differently in different experimental conditions,
and finally their DV scores are compared.
• Example: A study evaluating the effect of a drug on blood sugar levels in diabetic
patients. Each participant’s blood sugar levels are measured under two conditions:
when the drug is taken (Treatment condition) and when the drug is not taken
(control condition). This design ensures that the same participants are involved in
both conditions, reducing variability caused by individual differences.
• This approach highlights the advantage of repeated measures designs, as it controls
for inter-individual variability and focuses on changes within the same individuals
across conditions.
• The repeated measures design can be divided into two parts based on convenience.
• A design having two conditions and many subjects: same group is tested under two
conditions as mentioned in the example of this slide.
• A design having more than two conditions and many subjects: same groups is tested
under several conditions of the test. Example: When studying the effect of a drug on
blood sugar levels across more than two conditions, such as no drug (control), low
dose, and high dose, each participant's blood sugar levels are measured under all
three conditions. This repeated measures design ensures that comparisons are made
within the same group, reducing variability caused by individual differences.
• Since the participants are exposed to all conditions, there must be enough time
between each condition to avoid carryover effects from one condition to another. A
1-2 week washout period between conditions is typically recommended.

Order effects
Order effects mean that tasks completed in one condition may affect task performance in
another. For example, participants may perform better in the second condition either because
of the practice effect or worse because of boredom or fatigue.
Counterbalancing is an experimental technique used to overcome order effects.
Counterbalancing ensures each condition is tested equally first or second. For example,
participants are divided in half, with one half completing the two conditions in one order and
the other half completing the conditions in reverse order. In this way, a researcher can control
the order of the conditions and ensure better validity.

Comparison between independent and repeated measures design:

1. Advantage of economy of subjects in a within-groups design.

2. Since the experimenter repeats the measures on the same group of subjects in a
within-groups design, the factor of individual difference is automatically controlled.
This increases the correlation among measures which, in turn, reduces the error variance.
3. If the experimental situation is such in which preparation for the experiment requires
a fair amount of time and patience (mostly neuropsychological and psychophysiological
experiments), the within-groups design is preferred to the between-groups design because
the same group of subjects with all its initial preparation, can more easily be tested under
different conditions of the experiment than different groups of subjects in different
conditions.
4. When the experiment is such that the order of presentation of conditions is likely to have a
profound effect upon the dependent variable and there are obvious practice and fatigue
effects, the within-groups design should not be preferred. In other words, when the
experimenter expects a differential transfer effect from one experimental condition to
another, he should not use the within-groups design.