|
|
Psychology
Research Methods in Psychology
Research Methods
Psychologists use many different methods for conducting research. Each method
has advantages and disadvantages that make it suitable for certain situations and
unsuitable for others.
Descriptive
or Correlational Research Methods
Case studies, surveys, naturalistic observation, and laboratory
observation are examples of descriptive or correlational
research methods. Using these methods, researchers can describe different
events, experiences, or behaviors and look for links between them. However,
these methods do not enable researchers to determine causes of behavior.
Remember: correlation is
not
the same
as causation. Two factors may be related without one causing the other to occur. Often, a third factor explains the
correlation.
Example:
A psychologist uses the survey method to study the
relationship between balding and length of marriage. He finds
that length of marriage correlates with baldness. However, he
can’t infer from this that being bald causes people to stay
married longer. Instead, a third factor explains the
correlation: both balding and long marriages are associated with
old age.
Measuring Correlation
A correlation coefficient measures the strength of the
relationship between two variables. A correlation coefficient is always a
number between –1 and +1. The sign (+ or –) of a correlation coefficient
indicates the nature of the relationship between the variables.
A positive correlation (+) means that as one variable
increases, the other does too.
Example:
The more years of education a person receives, the
higher his or her yearly income is.
A negative correlation (–) means that when one variable
increases, the other one decreases.
Example:
The more hours a high school student works during
the week, the fewer A’s he or she gets in
class.
The higher the correlation coefficient, the stronger the correlation.
A +0.9 or a –0.9 indicates a very strong correlation; a +0.1 or a –0.1
indicates a very weak correlation. A correlation of 0 means that no
relationship exists between two variables.
Common correlational research methods include case studies, surveys,
naturalistic observation, and laboratory observation.
Case Studies
In a case study, a researcher studies a subject in depth.
The researcher collects data about the subject through interviews, direct
observation, psychological testing, or examination of documents and records
about the subject.
Surveys
A survey is a way of getting information about a specific
type of behavior, experience, or event. When using this method, researchers
give people questionnaires or interview them to obtain information.
When subjects fill out surveys about themselves, the data is called self-report data. Self-report data can be misleading
because subjects may do any of the following:
- Lie intentionally
- Give answers based on wishful thinking rather than the truth
- Fail to understand the questions the survey asks
- Forget parts of the experience they need to describe
Naturalistic Observation
When using naturalistic observation, researchers collect information
about subjects by observing them unobtrusively, without interfering with
them in any way. Researchers create a record of events and note
relationships among those events. With naturalistic observation, researchers
face the challenge of getting a clear view of events without becoming
noticeable to the subjects.
Laboratory Observation
As the name implies, researchers perform laboratory
observation in a laboratory rather than in a
natural setting. In laboratory observation, researchers can use
sophisticated equipment to measure and record subjects’ behavior. They can
use one-way mirrors or hidden recording devices to observe subjects more
freely while remaining hidden themselves. Unlike observation in a natural
setting, laboratory observation offers researchers some degree of control
over the environment.
Psychological Tests
Researchers use psychological tests to collect information
about personality traits, emotional states, aptitudes, interests, abilities,
values, or behaviors. Researchers usually standardize these tests,
which means they create uniform procedures for giving and scoring them. When
scoring a test, researchers often compare subjects’ scores to norms, which are established standards of performance on a test. A
well-constructed standardized test can evaluate subjects better than self-report
data.
Reliability
A test has good reliability if it produces the same
result when researchers administer it to the same group of people at
different times. Researchers determine a test’s test-retest
reliability by giving the test to a group of people and then
giving the test again to the same group of people at a later time. A
reliable test will produce approximately the same results on both occasions.
Psychologists also use alternate-forms reliability to
determine a test’s reliability. They measure alternate-forms reliability by
giving one version of a test to a group of people and then giving another
version of the same test to the same group of people. A reliable test will
produce roughly the same results no matter which version of the test is
used.
Validity
A test is valid if it actually measures the quality it
claims to measure. There are two types of validity:
-
Content validity is a test’s ability to measure all
the important aspects of the characteristic being measured. An
intelligence test wouldn’t have good content validity if it measured
only verbal intelligence, since nonverbal intelligence is an important
part of overall intelligence.
-
Criterion
validity is fulfilled when a test not only measures a trait but also predicts
another criterion of that trait. For example, one criterion of
scholastic aptitude is academic performance in college. A scholastic
aptitude test would have good criterion validity if it could predict
college grade point averages.
Experiments
Unlike correlational research methods or psychological tests, experiments can provide information about cause-and-effect relationships between variables. In an experiment, a
researcher manipulates or changes a particular variable under controlled
conditions while observing resulting changes in another variable or
variables. The researcher manipulates the independent
variable and observes the dependent variable. The
dependent variable may be affected by changes in the independent variable. In
other words, the dependent variable depends (or is thought to depend) on the
independent variable.
Experimental and Control Groups
Typically, a researcher conducting an experiment divides subjects into
an experimental group and a control group. The subjects in both groups
receive the same treatment, with one important difference: the researcher
manipulates one part of the treatment in the experimental group but does not manipulate it in the control group. The variable
that is manipulated is the independent variable. The researcher can then
compare the experimental group to the control group to find out whether the
manipulation of the independent variable affected the dependent variable.
Often, subjects in the control group receive a placebo drug or
treatment, while subjects in the experimental group receive the real drug or
treatment. This helps researchers to figure out what causes the observed
effect: the real drug or treatment, or the subjects’ expectation that they
will be affected.
Example:
Suppose a researcher wants to study the effect of
drug A on subjects’ alertness. He divides 100 subjects into
two groups of 50, an experimental group and a control group.
He dissolves drug A in saline solution and injects it into
all the subjects in the experimental group. He then gives
all the control group subjects an injection of only saline
solution. The independent variable in this case is drug A,
which he administers only to the experimental group. The
control group receives a placebo: the injection of saline
solution. The dependent variable is alertness, as measured
by performance on a timed test. Any effect on alertness that
appears only in the experimental group is caused by the
drug. Any effect on alertness that appears in both the
experimental and control groups could be due to the
subjects’ expectations or to extraneous variables, such as
pain from the injection.
Extraneous Variables
Ideally, subjects in the experimental and control
groups would be identical in every way except for the
variables being studied. In practice, however, this
would be possible only if researchers could clone
people. So researchers try to make groups with subjects
that are similar in all respects that could potentially
influence the dependent variable. Variables other than
the independent variable that could affect the dependent
variable are called extraneous variables.
One way to control extraneous variables is to use random assignment.
When researchers use random assignment, they create
experimental and control groups in a way that gives subjects an equal chance
of being placed in either group. This guarantees the two groups’ similarity.
Disadvantages of Experiments
The main disadvantage of experiments is that they usually don’t fully
reflect the real world. In an experiment, researchers try to control
variables in order to show clear causal links. However, to exert control in
this way, researchers must simplify an event or a situation, which often
makes the situation artificial.
Another disadvantage of experiments is that they can’t be used to
study everything. Sometimes researchers can’t control variables enough to
use an experiment, or they find that doing an experiment would be
unethical—that is, it would be painful or harmful in some way to the
subjects being studied.
Bias in Research
Bias is the distortion of results by a variable. Common types
of bias include sampling bias, subject bias, and experimenter bias.
Sampling Bias
Sampling bias occurs when the sample studied in an
experiment does not correctly represent the population the researcher wants
to draw conclusions about.
Example:
A psychologist wants to study the eating habits of
a population of New Yorkers who have freckles and
are between the ages of eighteen and forty-five. She
can’t possibly study all people with freckles in that
age group, so she must study a sample of people with
freckles. However, she can generalize her results to the
whole population of people with freckles only if her
sample is representative of the population. If her
sample includes only white, dark-haired males who are
college juniors, her results won’t generalize well to
the entire population she’s studying. Her sample will
reflect sampling bias.
Subject Bias
Research subjects’ expectations can affect and change the subjects’
behavior, resulting in subject bias. Such a bias can manifest
itself in two ways:
- Aplacebo effect is the effect on a subject receiving a
fake drug or treatment. Placebo effects occur when subjects believe
they are getting a real drug or treatment even though they are not.
A single-blind experiment is an experiment in
which the subjects don’t know whether they are receiving a real
or fake drug or treatment. Single-blind experiments help to
reduce placebo effects.
- The social desirability bias is the tendency of some
research subjects to describe themselves in socially approved ways. It
can affect self-report data or information people give about themselves
in surveys.
Experimenter Bias
Experimenter bias occurs when researchers’ preferences
or expectations influence the outcome of their research. In these cases,
researchers see what they want to see rather than what is actually there.
A method called the double-blind procedure can help
experimenters prevent this bias from occurring. In a double-blind procedure,
neither the experimenter nor the subject knows which subjects come from the
experimental group and which come from the control group.
|
|
