In a study, Merzenich examined whether healthy participants aged 60 years or older would show
improved mental abilities following brain plasticity training that involved working out
performing a prescribed 5-day-a-week regimen or cognitive exercises on their home computers.
What is the scientific and most rigorous way to test whether brain exercises actually work? Can a
causal relationship be demonstrated between brain plasticity training and improved memory?
For the Merzenich study, the question posed is one of cause and effect, and as such, it dictates an
experimental design known simply as a true experiment. Recall, an experiment is a controlled
investigation in which one or more variables are manipulated. And as you know, variables differ
in the degree to which they can be controlled or manipulated. In a true experiment, a researcher
has complete control over the manipulation of the independent variable. Such experimental
control allows for testing whether systematically varying the independent variable causes
changes in the dependent variable. For the Merzenich study, this meant investigating whether
varying training could lead to improved memory.
The defining feature of an experiment is the degree of control a researcher exercises over all
aspects of the study. In an experiment, the researcher controls with whom, with what, and how
the study is conducted. The researcher also directly controls what levels of the independent
variables are to be manipulatedfor example, Merzenich produced 3 levels of values of the
independent variable of treatment or cognitive training. Third, the researcher controls how the
experiment is conducted by holding constant certain aspects, such as the time, place, and setting
of the study. For example, all participants in the Merzenich study are measured on the same
outcome measures of cognition, and participants in either brain plasticity or active computer
training received an equal number of sessions.
A researcher exercises experimental control in a study by holding constant as many extraneous
variables as possible. If a variable is held constant, it cannot account for the results of the
experiment. For example, participants in all groups in the Merzenich experiment should be
treated identically, with the sole exception that some participants receive the treatment and others
do not. Thus, the only difference between the groups is the manipulated variable, which in this
case refers to the group who received the treatment versus the groups who did not receive the
Another technique to control for the unwanted effects of a variable is known as randomization.
Think about a common memory experiment that studies show how many trials are needed for
participants to learn a list of words that vary in terms of visual imagery. Here our independent
variable is visual imagery of words. However, we know that irrespective of imagery, words
presented at the beginning and the end of a list tend to be remembered better than words in the
middle of a list (primacy effect and recency effect). To control for this position effect, an
experimenter could randomize the order of presentation of words so that each list would present
the same words in a different order.
Randomization is used in assigning participants to groups or to different levels of an independent
variable. If we have 2 groups or 2 conditions, we could simply flip a coin with heads being
assigned to one group and tails to the other. **Pg. 175 for diagram** This process of random assignment ensures that any extraneous influence is just as likely to
affect one group as the other group.
In many psychology studies, however, an experimenter cannot use random assignment to create
experimental and control groups. In these instances, the levels of the independent variable are
selected rather than directly manipulated by the experimenter; that is, participants are selected
and grouped after the fact or, on the basis of a particular characteristic, such as gender, race, age
or diagnoses. Under these circumstances, the researcher conducts a quasi-experiment.
For example, suppose Merzenich chose to examine whether brain plasticity training improves
memory in both young adult and elderly groups. He would form their groups not by random
assignment but by selecting, on the basis of age, young and elderly participants. They would
therefore use a quasi-experiment because they could not randomly assign participants to different
levels of the independent variables of age. (Quasi-experiments to be discussed in Chapter 8)
We have learned that an independent variable is something that a researcher manipulates in some
way in order to determine its effect on a dependent variable. We have also learned that an
independent variable must have at least 2 values that are referred to as levels. For example, if
you wanted only to compare participants who received brain plasticity training with those who
received no training, you would need a control group (who receives no training) along with a
group who receives training.
An experiment must have at least one independent variable that is manipulated by the
experimenter; otherwise it would not be an experiment. An experiment with one independent
variable is called a single-factor experiment, and experiments with more than one independent
variable are often referred to as multifactorial designs
Independent variables are the conditions that are directly and systematically manipulated by an
experiment. The term condition simply means a particular way in which participants are treated.
For example, the experimenter may compare 2 groups of participants-those who receive
treatment versus those who receive no treatment. Here conditionlabelled groupand
treatment can be used interchangeably to refer to the independent variable.
In the Merzenich study, treatment or cognitive training is the independent variable, with 3 levels.
The group of participants who received the level of independent variable that is the focus of the
study (computer-based brain plasticity training) is called the experimental group or treatment
group. The group of participants who received the computer-based active training is called the
comparison group. The group that received no training is called the control group. Do not
confuse a control group with a control variable, which is defined as a variable that is held
constant. The control group provides a means to determine changes that might occur naturally in
the absence of treatment. The control group thus provides a baseline against which the treatment
or experimental group can be compared.
All true experiments have as the dependent variable a post-testthat is, measurement of the
outcome in both groups after the experimental group has received the treatment. Manyrandomized experiments also have pre-tests that measure the dependent variable prior to the
A pre-test can be exactly the same as a post-test, just administered at a different time.
Oftentimes, however, using identical pre- and post-tests is to be avoided, especially if
performance is likely to be influenced by practice in taking the same measure twice.
A control variable is a potential independent variable that is held constant throughout the
experiment because its influence is thought to be extraneous, or unimportant to the goal of the
One of the most important decisions pertains to the assignment of participants to the various
levels of the independent variables. The 2 main alternatives to consider are to assign only some
participants to each level of the independent variable or to assign each participant to each level of
the independent variables.
These 2 alternatives constitute 2 types of experimental designs: between-subjects and within-
subjects. With a between-subjects design, independent groups of participants receive the
different levels of the independent variable. By contrast, a within-subjects design is one in
which all research participants receives all levels of the independent variable
Between-Subjects Design For a between-subjects design with 2 levels of an independent
variable, some participants receive one level and others receive the other level of the independent
variable; the same participants never receive all levels of the independent variable. For this
reason, this type of study requires a large number of people to be enrolled. Such a design is
often used in treatm