Method of Writing
Research Papers 6
Trương Văn Ánh
HUFLIT

VARIABLES IN EXPERIMENTAL STUDIES
All experimental studies have at least one independent
variable and one dependent variable. The purpose of the
experiments is to estimate the extent to which independent
variables cause changes in dependent variables. As indicated
in the previous section, an independent variable is a stimulus
or input variable. Note that in experiments, researchers
physically manipulate independent variables. Examples of
physical manipulation are (1) giving a new drug to some
participants while giving a placebo to others, (2) providing
some students with computers while denying computers to
others, and (3) using group counseling with some clients while
using individual counseling with others.

Thus, studies that claim to physically manipulate have
administered a treatment that creates a tangible difference
between those receiving and not receiving treatment through
some real-world action or environmental change.
In nonexperimental studies, researchers do not physically
manipulate independent variables. Instead, they observe
independent variables as they occur (or have occurred)
naturally.
For instance, researchers observe the health of individuals
who report drinking alcohol in nonexperimental studies—
researchers do not provide participants with alcohol, and
then monitor the results of ingesting it (see Topic 4 to review
other differences between experimental and onexperimental
studies of causation).
In a simple experiment, as in Example 1 below, there is only
one independent variable and only one dependent variable.

Example 1
On alternate weeks, a disruptive first-grade student is
given extra praise for being in her seat when
appropriate. The purpose of the study is to see if the
extra praise will increase the amount of appropriate
in-seat behavior.
In Example 1, the physical manipulation is giving or
not giving extra praise, which is the independent
variable (i.e., the stimulus or input variable). The
dependent variable (i.e., the response or outcome
variable) is the change in the student's in-seat
behavior.
Often, experiments have more than one dependent
variable. For instance, in Example 1, a researcher

could observe not only if the treatment causes (1)
more in-seat behavior, but also(2) improvement in the
student's achievement and (3) improvement in the
student's attitude toward school during the weeks in
which the extra praise is given. If a researcher did this,
the researcher would have three dependent variables.
Many experiments also have more than one
independent variable. Often, these experiments are
more interesting than those with only one
independent
variable
because
they
provide insights into the causal effects of
combinations of variables. Consider Example 2,
which has two independent variables (childcare and
transportation money).

Example 2
Voluntary, free job training was offered to all
mothers on welfare in a small city. Four groups of
the mothers were formed at random to explore the
effects of these two independent variables: (1)
providing or not providing free childcare while in
training and (2) providing or not providing
transportation money to get to the job-training site.
Each group was assigned at random to one of the
four treatment conditions shown here:

GROUP 1
childcare and transportation money
GROUP 2
childcare and no transportation money
GROUP 3
no childcare and transportation money
GROUP 4
no childcare and no transportation money

It was predicted that those in Group 1 (the group that
was given both childcare and transportation money)
would have the highest attendance rates in the jobtraining program, those in Group 2 would have the
next highest, those in Group 3 would have the next
highest, and those in Group 4 would have the lowest.
Notice that in Example 2, a researcher can determine
(1) how effective childcare is, (2) how effective
transportation money is, and (3) the effectiveness of
both childcare and transportation money in
combination. Thus, a researcher would get more
information by looking at two independent variables
in one study than by looking at each independent

(in which case a researcher could determine only
points 1 and 2).

TOPIC REVIEW
1. All experiments have at least how many dependent
variables?
2. In an experiment, what is the name of a stimulus or
input variable?
3. What does physically manipulate mean in an
experimental context?
4. Are dependent variables physically manipulated?
5. Can an experiment have more than one
independent variable?
6. Every other customer entering a shoe store was
given one of two different coupons. One coupon
offered a second pair of shoes for 50% off. The other
coupon offered to reduce the total price by 25% if two

pairs of shoes were purchased. The purpose was to
determine which coupon was more effective in getting
individuals to buy two pairs of shoes. In this
experiment, what is the independent variable?
7. In Question 6, what is the dependent variable?
8. A teacher showed an educational film on daily
nutritional needs to one group of students
and gave a handout on the same material to another
group. The purpose was to determine which method
of instruction was more effective in increasing
students' knowledge of daily nutritional needs. In this
experiment, what is the dependent variable?
9. In Question 8, what is the independent variable?

DISCUSSION QUESTIONS
1. In an experiment involving which books children
select at the library, name a variable that would be
easy for a researcher to physically manipulate in an
experiment. Then, name a variable that might be
affected by the manipulation.
2. Name a variable that you would be unwilling to
physically manipulate due to ethical or legal
concerns.

RESEARCH PLANNING
If you will be conducting an experiment, name the
independent and dependent variables you will be
studying. Is there an opportunity to look at more
than one independent or dependent variable?
Evaluate the reasons to pursue a study with
additional variables, or the reasons to maintain
single variables.

Good luck!