CRITICAL THINKING LESSON 2B:
The Need to Control for Expectations
My name is Taylor and I suffered from terrible lower back pain for
almost 10 years. About a year ago, I bought a Somnocare Sleep System,
which contains a revolutionary new mattress that is like nothing else
on the market today. Within a month after I began sleeping on my Somnocare
mattress, my back pain had decreased to a point where I was able to
go back to the gym and do step aerobics, which I love. Today, I'm a
new woman, with no back pain. I can pick up my young daughter without
any problem, and my husband tells me that I look years younger. All
thanks to the Somnocare Sleep System!
How often have you heard stories like this one on television commercials
or in magazine advertisements? Such stories are called "testimonials."
A testimonial is an anecdote in which the merits of a product
or service are attested to. Although testimonials seem to many people
to be very compelling evidence for the value of a product or service,
they actually are virtually worthless in this regard. This is because
they are anecdotes and, thus, do not control for the effects of important
extraneous variables.
One very important extraneous variable that testimonial evidence does
not control for is the set of expectations that the testimonial-giver
has for the product or service. These expectations may result in a "placebo
effect." A placebo is an inert substance or an irrelevant
activity that is presented as a treatment for a physical or psychological
disorder. Because no active treatment is being given, a placebo should
have no effect on the symptoms of a disorder. Nevertheless, it often has
been found that, in the case of at least some disorders, placebos do
cause improvement. As you learned in Chapter 2, the placebo effect
is the reduction of a disorder's symptoms caused by the belief that
one is receiving a treatment. The variable that seems to underlie
placebo effects is the expectation that one will begin to feel
better when receiving what one believes to be a treatment. This expectation
does not only cause a person to feel subjectively as if they are getting
better, it also has been shown to cause real objectively measured bodily
changes in some cases (see, for example, Russo,
2002). Thus, expectations can have powerful influences on our minds
and bodies. This is why most scientific research on treatments for various
disorders include placebo controls. It is important to determine how much
effect the participants' expectations are having on the symptoms of a
disorder.
But it is not only the participants' expectations that must be controlled
for when performing research. Researchers also have expectations that
may affect the results of a study. In Chapter 2, the effects of researchers'
expectations on a study's results were referred to as experimenter
bias. A well known example of experimenter bias is found in a research
program in physics that took place at the beginning of the twentieth century.
In 1903, soon after the discovery of X-rays, a French physicist by the
name of René Blondlot discovered what he called N-rays (named after
the University of Nancy, which was where he did his work). Just like X-rays,
Blondlot discovered that N-rays are not visible. But he claimed to be
able to detect their presence through their ability to increase the brightness
of luminous objects (such as electric sparks). Blondlot and other researchers
found that N-rays are emitted by the sun, flames, and many natural objects.
And Blondlot stated that some substances, such as the fluid within the
eye, absorb N-rays and then emit them later, which allows people to see
better in darkened rooms. Most of these findings were replicated by a
number of researchers in different laboratories. The evidence seemed conclusive
and many physicists were certain that Blondlot had discovered something
very important.
But others doubted the existence of N-rays because they were unable to
replicate Blondlot's findings. They pointed to a basic flaw in Blondlot's
method of detecting N-rays: the observer judged by sight whether
or not an object bombarded with N-rays had increased in brightness. In
other words, the observations used to detect N-rays were highly subjective.
When observations are subjective, the effects of expectations can have
a large effect on them. In this case, if a researcher believed that N-rays
were present, and expected that the N-rays would increase the brightness
of objects, he or she might perceive an increase in brightness even when
no change actually had occurred. It is noteworthy in this respect that
most of the studies that found evidence for the existence of N-rays were
performed in France, where Blondlot was a preeminent scientific authority.
When he stated that N-rays existed, this was likely to have a strong influence
on the expectations of other French researchers.
How could one test the hypothesis (a hypothesis is a speculation
about the relationship between two or more variables) that researchers'
expectations affected the perceived brightness of objects in N-ray studies?
There are two things that one could do: (a) tell observers that
N-rays were present when they actually were absent, and (b) tell observers
that N-rays were absent when they actually were present. If observers
saw an increase in the brightness of objects only when they believed that
N-rays were present, regardless of whether or not they actually were,
then it could be concluded that their expectations caused their observations
to be mistaken. In 1904, an American physicist by the name of Robert Wood
performed this experiment in Blondlot's laboratory:
N-ray experiments had to be carried out in a darkened laboratory....
This gave Wood an opportunity to make several observations that proved
Blondlot's judgements of brightness changes were a function of his beliefs,
and not of the presence or absence of N-rays. In one experiment, Wood
was to block an N-ray source by inserting a sheet of lead between the
source and a card with luminous paint on it [Blondlot had found that
N-rays could not penetrate lead].... Without telling Blondlot, Wood
changed the experiment in one slight but vitally important way. He would
indicate to Blondlot that the lead sheet was blocking the N-ray source
when it really wasn't, or vice versa. If N-rays really existed, Blondlot's
judgements of the brightness of the luminous paint should be a function
of whether the lead screen really was between the card and the N-ray
source and should have no relationship to whether or not he believed
the sheet was blocking the source.... [Wood found that if Blondlot]
believed the screen was present (blocking N-rays), but it wasn't, he
reported the paint to be less luminous. If he was told the screen was
not present (allowing N-rays to pass), but it really was, he reported
the paint to be more luminous. (Hines, 1988, p. 10).
This study showed that observers' expectations determined their judgements
of brightness in N-ray studies. Thus, Wood concluded that the results
of previous studies supporting the existence of N-rays were contaminated
by experimenter bias. Blondlot, however, never gave up his belief in N-rays:
"convinced until the end that N-rays were real, [Blondlot] pursued
his research on the topic until his death in 1930" (Hines, p. 11).
This shows that researchers' expectations can have such a large influence
on their observations that they will reject even undeniable evidence that
a particular observation is mistaken. In other words, once we have
developed a belief based on a compelling personal experience, we often
are very reluctant to give up this belief. This also is a problem
in the case of placebo effects: a sick person who feels better
after taking what amounts to a placebo often will swear that the treatment
was effective, even when told that it had no active ingredients in it.
In general, the expectations that we bring into a situation can have
powerful effects on what we perceive, how we behave, and what we remember
later on about the situation. In Critical Thinking
Lesson 4, we will return to the topic of personal experience and its
limitations.
CRITICAL THINKING QUESTIONS FOR LESSON 2B
Question 2B-1
Professor Harrington decided to perform a study of maze learning in rats
in order to learn more about a particular brain structure--the "corpus
substantia"--that he thinks may be important for learning. He found
two strains of rat that differ in the size of this brain structure:
the corpus substantia is larger, on average, in rats from Strain A
than it is in rats from Strain B. He developed the following hypothesis:
rats from Strain A, because of the larger size of their corpus substantia,
will learn the maze more quickly than will rats from Strain B.
Because he didn't have time to perform the experiment himself, Professor
Harrington hired student volunteers to test the two strains of rats in
the maze. He described in much detail the rationale of the experiment
to the students, and informed them of his hypothesis. He then gave each
student a cage labelled Strain A that contained several rats, and
a second cage labelled Strain B that contained several rats. The
students found that rats from Strain A learned the maze faster,
on average, than did the rats from Strain B.
Based on what you have learned in this lesson, can we conclude that rats
from Strain A, perhaps because of their larger corpus substantia,
learn the maze more quickly than do rats from Strain B?
Suggested Answer
Question 2B-2
A group of researchers put advertisements in local newspapers asking
for people to participate in a study investigating the effectiveness of
a new anxiety-reducing (anxiolytic) medication. The advertisement stated
that potential participants should be suffering from severe anxiety that
is not being treated with any medication or psychotherapy. Sixty people
meeting the criteria contacted the researchers. The 60 participants received
the anxiolytic medication for a total of eight weeks. Their degree of
anxiety was measured at the beginning of the study and then again at the
end of the study. The researchers found a large average decrease in the
degree of anxiety over the course of the eight-week treatment.
Is this good evidence for the effectiveness of the anxiolytic medication?
Why or why not? If you stated that it was not good evidence, what would
you do to improve the study?
Suggested Answer
Bibliography and References
Hines, T. (1988). Pseudoscience and the paranormal: A critical examination
of the evidence. Amherst, NY: Prometheus.
Ricker, J. P. (2002). An introduction to the science of psychology.
Boston: Pearson Custom Publishing.
Rosenthal, R., & Fode, K. L. (1963). The effect of experimenter bias
on the performance of the albino rat. Behavioral Science, 8, 183-189.
Russo, E. (2002). The biological basis of the placebo effect. The
Scientist, 16(24). Retrieved March 17, 2003, from http://www.the-scientist.com/yr2002/dec/research_021209.html
|