The document discusses extraneous variables that can pose problems in experiments if not properly controlled. There are five main types of extraneous variables: 1) physical variables related to the testing environment, 2) response style where subjects answer in predictable ways, 3) response set where subjects respond based on social desirability rather than the question, 4) demand characteristics where subjects behave how they think the experimenter wants, and 5) experimenter bias where subtle cues from the experimenter influence subjects. Controlling these variables involves techniques like eliminating distractions, counterbalancing questions, blinding subjects and experimenters, and ensuring experimenters remain objective.
2. Extraneous Variables
One of the goals of a researcher should be to set up
experiments that do not create confusion. When
experimenting, conditions are created to clearly see
the effects of the independent variables.
Experiments should be internally validated. Ideally,
only the dependent variable should change
systematically from one condition to another. Control
of the subject variables may lead to worsening.
3. Extraneous Variables
There are three types of extraneous variables that
may pose problems in experiments. They are:
1. Physical variables;
2. Response style;
3. Response set;
4. Demand characteristics; and
5. Experimenter bias
4. Physical Variables
The testing setting, noise, and
other distractions are all physical
aspects of the testing conditions
that need to be controlled.
The techniques for controlling
physical variables are through:
1. Elimination;
2. Constancy;
3. Screening; and
4. Counter balancing
5. Response Style
A subject with a response
style has a characteristic way
of answering questions.
They usually respond “Yes”
or “No”; and “True” or “False”
to most inquiries.
6. Response Style
To control this style,
experiments should
formulate questions in such a
way that subjects cannot
answer with a simple yes or
no. Contents of the inquiry
should be built to force the
subject to think about their
responses.
7. Response Set
Subjects with a response set
respond in terms of latent content or
the meaning behind the questions.
Social desirability is the best-known
response set.
Subjects who have the social
desirability response will lie or distort
the truth in order to look better in the
eyes of the experimenter.
Does being gay or
lesbian goes
against the will of
God?
8. Response Set
Social desirability can be
avoided to some extent by
phrasing questions in such a
way that alternative answers
have roughly the same social
value.
Does being gay or
lesbian goes
against the will of
God?
What is your
stand toward
being a gay or
lesbian?
9. Demand Characteristics
The mere fact of being in an
experiment can lead to changes in
subjects’ behavior which have
nothing to do with the experimental
manipulations.
Their expectation of what will happen
to them in the experiment can shape
their responses.
Subjects may want to be “good”
subjects; they may try to provide data
that confirm the hypothesis.
10. Demand Characteristics
A way of controlling demand
characteristics is to run a blind
experiment in which the
experimenter will tell everything
about the experiment except
what the researcher expect to
find.
This is commonly applicable for
experiments involving nutrition
and drugs.
11. Experimenter Bias
The last potential source of error
is the experimenter.
They may provide subtle cues that
tell the subjects how they are
expected to behave during the
experiment.
Example: The experimenter
smiles at every time the subject
give the desired and predicted
response.
12. Experimenter Bias
To control experimenter bias, the
experimenter must not know how
subjects are expected to behave.
This approach enables the
experimenter to measure the
dependent variable more
objectively.