Controlling

Extraneous Variables

Diene S. Lonquianas
 Extraneous Variables
• Variables that can threaten an experiment
internal validity

• Physical
• Social
• Personality
• Context
 Physical Variable
• Day of the week, testing room, noise,
distractions
• Experimental group and control group were
tested on two separate occasions with grossly
different physical conditions

• Can be avoided by:

• Elimination
• Constancy of condition
• Balancing
 Elimination

• Simply, eliminating variables that can confound

the experiment

• Noise?
• Use a soundproof room., hang a “Don’t
Disturb” sign
 Constancy of Conditions
• If you cannot eliminate extraneous variable, try
to keep all aspects of the treatment conditions as
nearly similar as possible.
• Same...
• Color of the wall, comfort level, lighting,
instructions, time, ventilation, etc.
 Balancing
• When neither elimination nor constancy can be
used.
• Subjects not available at the same time.
• Limited use of rooms.
• Confounded
Experimental Group Control Group

Room 801; 9am, July 15, 2009 Room 804, 3pm; July 17

Participants: Participants:

E1, E2, E3, E4, E5, E6, E7, E8, E9, E10, C1, C2, C3, C4, C5, C6, C7, C8, C9, C10,
E11, E12, E13, E14, E15, E16 C11, C12, C13, C14, C15, C16
 Balancing
• Distributing the effects of an extraneous variables across
the different treatment conditions.
• Balanced (assign subjects randomly)

Experimental Group Control Group

Room 801; 9am, July 15, 2009 Room 804, 3pm; July 17

Participants: Participants:

E1, E4, E6, E7, E9, E10, E11, E13 C1, C2, C4, C6, C12, C13, C14, C16

C3, C5, C7, C8, C9, C10, C11, C15, E2, E3, E5, E8, E12, E14, E15, E16
 Social Variables

• Qualities of the relationship between subjects

and experiments
• Demand characteristic
• Experimenter bias
 Demand characteristics
• Aspects of the experimental situation that demand
the people behave in a particular way.
• Our behavior is shaped by what we think is expected
in a given situation.
• Research subjects want to be good subjects
• They might try to guess the hypothesis; may set out
to prove or disprove it.
• Can be controlled by:
• Single-blind experiment
• Cover story
 Single-blind Experiment

• An experiment in which subjects do not know

which treatment they are getting.
• Some information about the experiment may be
revealed
• Placebo effect
• Even so, there is still a possibility, though slim,
that the subjects will figure out the hypothesis
 Cover Stories
• A plausible, false explanation for the procedures
used in the study, in order to disguise the actual
research hypothesis.
• Deception is a departure form informed consent
• Whenever possible, do not use one
• Debriefing is required
 Experimenter Bias
• An experimenter may unknowingly give subjects
cues that tell them how he would like then to
respond.
• Gestures, tone of voice, behavior of the
experimenter can vary systematically across
treatment conditions; errors in recoding data
• What can be done:
• Follow written direction, time the experiment, be
consistent, minimize personal contact.
• Can be controlled by:
• Double-blind Experiment
 Double-Blind Experiment
• The subjects do not know which treatment they
are receiving, and the experimenter does not
know either.
• Use of an independent rater