Abnormal Psychology

Researching Abnormality 157

Confounding variables (confounds)

Factors that might inadvertently affect the

variables of interest in an experiment.

A researcher might separately manipulate several independent variables (always

keeping all else constant while a single variable is changed), hoping to discover

which ones cause the greatest effect on the dependent variable.

Of course, researchers could not use an experimental design to investigate the

specifi c question about helplessness during early loss and subsequent depression

after a breakup. A researcher ethically cannot cause a person to have a major loss

during childhood (or to feel helpless at the time). Participants in an experimental

study come as they are—with particular neurological, psychological, and social his-

tories that can’t be changed.

This is why most of the research on causes of psychological disorders does not

use an experimental design. For ethical reasons as well as practical ones, researchers

cannot alter participants’ genes, subject participants to high levels of stress, cause

traumatic experiences to occur in their lives, or create disruptive family events—all

of which would involve intentionally manipulating independent variables. However,

some aspects of psychopathology can be studied with an experimental design. For

instance, an experimental design has been used to study learned helplessness in dogs

(see Chapter 2; Overmier & Seligman, 1967); the independent variable was whether

the dogs could escape an electrical shock, and the dependent variable was the num-

ber of attempts to escape.

Similarly, Watson and Rayner’s experiment with Little Albert (1920; see Chapter 2),

used an experimental design to test a theory about the etiology of a phobia. In that

study, the independent variable was whether or not the conditioned (that is, the

initially neutral) stimulus had been paired with an aversive unconditioned stimulus.

The dependent variable was the presence of fear–related behaviors—as measured

by Albert’s crying and trying to get away from the white rat. When the conditioned

stimulus (CS) and the unconditioned stimulus (UCS) had not yet been paired, Albert

was not afraid of the rodent. His fear–related behaviors (the dependent variable)

depended on his exposure to the pairing of CS and UCS.

Other examples of experimental designs in research on psychopathology include

studies of people who have panic attacks—specifi c periods of intense dread or fear,

accompanied by physical symptoms of fear; the independent variable is the situation

or condition that may induce a panic attack, and the dependent variable is the num-

ber of such attacks. Further examples include studies of people who have substance

abuse problems, where the independent variable is the type of cues, or stimuli, that

trigger cravings to take the drug or to drink alcohol, and the dependent variable may

be intensity of cravings for the drug or alcohol or physiological measures of arousal.

You could use an experimental design to conduct research that relates to the

relationship between childhood loss and mood. You might, for instance, select three

groups of participants: those who had a childhood loss but were able to cope with

it, those who had a loss and experienced helplessness (like Carlos), and those who

did not have a loss (note that you are not manipulating this variable, and thus com-

posing the three groups is not part of the experiment itself). You could then have

the independent variable be the type of movie viewed by participants; the fi lm could

involve either the theme of loss or some other theme that does not involve loss. The

dependent variable might be participants’ ratings of mood after viewing the movie.

Your hypothesis for this study might be that participants who had suffered a loss

and experienced helplessness during childhood would have the most negative mood

after watching a movie with the theme of loss.

If changes in the independent variable do in fact change the measurements of the

dependent variable (that is, they produce an effect), there is a relationbetween the vari-

ables. But if the experiment has not been carefully designed this relation could be more

apparent than real. Factors that might inadvertently affect the variables of interest in

the experiment are called confounding variables, or confounds. For example, suppose

that you conducted an experiment in which the independent variable was the type of

movie viewed by participants, with one movie involving the death of a loved one and

the other not touching on the theme of loss. Then suppose that the movie involving loss

was always shown immediately after a serious drama for the “loss-helplessness” group,

whereas the movie that did not involve loss was always shown immediately after

Some studies investigating psychopathology

use an experimental design. Here the indepen-

dent variable is different types of music and the

dependent variable is depressed participants’

self-reported mood after listening to the music.

blickwinkel/Alamy