Chapter 2: Scientific Methods in Psychology
Chapter 2: Scientific Methods in Psychology
Kalat
Chapter 2: Scientific Methods in Psychology
Chapter 2
Scientific Methods in Psychology
Module 2.1
Scientific Theories
What is a theory?
A theory is a comprehensive explanation of observable
events and conditions.
A good theory makes precise and consistent predictions
while relying on a small number of underlying
assumptions.
Scientific Theories
Scientific Theories
Concept Check
Concept Check
Concept Check
Concept Check
Concept Check
Scientific open-mindedness:
Psychology as a Science
Module 2.2
Concept Check
Operational definitions
Which of the following might be used as an operational
definition of attraction?
A feeling of affection when two people are together. (1)
The number of minutes during which two people are
touching each other over a four-hour period. (2)
(2)
Concept Check
Population Samples
Usually in research we are asking questions that are
pertinent to a large population of interest such as:
Seven to ten-year-old children
People diagnosed with depression
Population Samples
But it is not practical to study all the individuals in the
population.
We take a relatively small number of observations or
individuals from the population, and we generalize from
that small number.
The small number of individuals or observations is called
a sample.
Population Samples
There are several types of samples and sampling
procedures:
A convenience sample is a group chosen because of
its ease of availability and study.
A representative sample closely resembles the
population in its percentage of males and females,
ethnic or racial groups, age levels, or whatever other
characteristics might have some relevance to the
results.
Population Samples
A random sample is one in which every individual in the
population has an equal chance of being selected.
A cross-cultural sample is one that contains groups of
people from at least two distinct cultures.
Population Samples
How we go about obtaining a sample has to be carefully
assessed in terms of our resources and goals.
Sometimes it is acceptable and appropriate to rely on a
convenience sample. Other times this strategy will
produce results that are useless in helping us
understand and interpret the real world.
Concept Check
Experimenter Bias
Because (fallible) humans do the research, we need to
keep in check the various tendencies that can work to
create erroneous research findings or erroneous
interpretations of findings.
Experimenter bias is the tendency of an experimenter to
unintentionally distort the procedures or results of an
experiment based on the expected or desired outcome
of the research.
Experimenter Bias
For example, if you were a researcher testing the
hypothesis that children who have been diagnosed with
learning disabilities are on average more creative than
children who have no diagnosis, you may find it hard to
ignore your hypothesis as you observe the children
with an LD diagnosis going about whatever tasks you
have devised to operationalize creativity.
Experimenter Bias
Methods have been devised to help counteract these
normal human tendencies that create bias:
Using blind observers who record data without
knowing what the researcher is studying.
Using a placebo control. A placebo is a pill or other
sham treatment that makes it very difficult for the
subjects (single-blind) or the subjects and
experimenter (double-blind) to know who has
received the treatment and who has not.
Research Design
There are many methods used to study psychological
concepts and phenomena.
We start by asking ourselves What happens, and under
what circumstances does it seem to occur?
We try to choose the best procedure. Each method has
advantages and disadvantages.
Correlational Studies
Correlation
Correlation is a measure of the relationship between
two variables which are both outside of the
investigators control.
Examples of variables include aspects such as:
height, weight, socio-economic level, number of years
of education.
The mathematical estimate of the strength and
direction of a correlation is the correlation coefficient.
Correlational Studies
The value of the correlation coefficient can range from
1.00 to +1.00.
The higher the absolute value, the stronger the
relationship is, regardless of the direction.
A negative correlation (-) means that as one variable
increases, the other decreases. An example of a
negative correlation is: The more absences a student
has, the lower his or her grade in psychology is (more
absences accompanied by fewer points on tests.)
Figure 2.8abc In a scatterplot each dot represents data for one person; for example,
each point in the center graph tells us one persons weight and that persons grade on
the psychology final exam, in this case using hypothetical data. A positive correlation
indicates that, as one variable increases, the other generally does also. A negative
correlation indicates that, as one variable increases, the other generally decreases.
The closer a correlation coefficient is to +1 or -1, the stronger the relationship.
Figure 2.8a In a scatterplot each dot represents data for one person; for example,
each point in the center graph tells us one persons weight and that persons grade on
the psychology final exam, in this case using hypothetical data. A positive correlation
indicates that, as one variable increases, the other generally does also. A negative
correlation indicates that, as one variable increases, the other generally decreases.
The closer a correlation coefficient is to +1 or -1, the stronger the relationship.
Figure 2.8b In a scatterplot each dot represents data for one person; for example,
each point in the center graph tells us one persons weight and that persons grade on
the psychology final exam, in this case using hypothetical data. A positive correlation
indicates that, as one variable increases, the other generally does also. A negative
correlation indicates that, as one variable increases, the other generally decreases.
The closer a correlation coefficient is to +1 or -1, the stronger the relationship.
Figure 2.8c In a scatterplot each dot represents data for one person; for example,
each point in the center graph tells us one persons weight and that persons grade on
the psychology final exam, in this case using hypothetical data. A positive correlation
indicates that, as one variable increases, the other generally does also. A negative
correlation indicates that, as one variable increases, the other generally decreases.
The closer a correlation coefficient is to +1 or -1, the stronger the relationship.
Correlational Studies
A positive correlation (+) means that as one variable
increases, so does the other. An example of a positive
correlation would be the higher the annual income, the
greater the amount and number of donations to charity
(more income accompanied by more charitable giving.)
A zero or near zero correlation means that the variables
have no relationship; that changes in one are not related
to any type of change in the other.
Concept Check
Concept Check
Concept Check
Concept Check
+. 30 or -.90
-.90
Correlational Studies
Some Problems with Interpreting Correlational
Research:
Illusory Correlation: An apparent relationship based
on casual observations of unrelated or weakly related
events.
Example: The belief in moon madness.
Correlational Studies
Some Problems with Interpreting Correlational
Research:
Correlation Causation: Correlational research only
tells us if two variables are related and how strongly.
It does not tell us why two conditions can appear
together and yet not cause each other.
Example: The more someone weighs, the larger
his or her vocabulary is. Do you know why?
Because weight and vocabulary both increase with age.
Figure 2.9 A strong correlation between depression and impaired sleep does not tell us
whether depression interferes with sleep, poor sleep leads to depression, or whether
another problem leads to both depression and sleep problems.
Concept Check
Experiments
Experiment:
A study in which the investigator manipulates at least
one variable (independent) while measuring at least
one other variable (dependent).
Figure 2.11 An experimenter manipulates the independent variable (in this case the programs people
watch) so that two or more groups experience different treatments. Then the experimenter measures
the dependent variable (in this case pulse rate) to see how the independent variable affected it.
Experiments
Example: To test whether the hormone adrenaline
enhances memory in mammals, a researcher teaches
rats to run a maze. She gives a randomly selected
portion of the rats a drug to block production of
adrenaline. She then times all the rats on the maze.
Experiments
Remember: In order for a study to be a true
EXPERIMENT, one of the variables must be directly
under the researchers control, and the other must be
measurable in some scientific way.
Figure 2.12 Once researchers decide on the hypothesis they want to test, they
must design the experiment. These procedures test the effects of watching
televised violence. An appropriate, accurate method of measurement is essential.
Concept Check
Concept Check
Experiments
Other important terminology:
Experimental group: The set of individuals who
receive the treatment that the experiment is designed
to test.
Control group: The set of individuals who are treated
in the same way as the experimental group except for
the procedure that the experiment is designed to test.
Random assignment: A selection method in which the
experimenter assigns subjects to either the
experimental or control group using a procedure
based on chance.
Experiments
Possible problems in carrying out and interpreting the
results of experiments:
Demand Characteristics: Cues that tell a subject what
is expected of him or her, and what the researcher
hopes to find.
Example: If the subject knows that the drug being
tested is supposed to improve mood, he or she
may feel better.
Experiments
Possible problems in carrying out and interpreting the
results of experiments:
Ethical Considerations: In doing research with
humans or animals, researchers must weigh possible
harm that may be inflicted against the usefulness and
other benefits that may be gained.
Psychological Research
Because of the challenges involved in studying intangible
mental processes and human behavior that is the product of
diverse influences, psychologists have developed procedures
that are rigorous and inventive and very frequently do
increase our understanding of the phenomena in this complex
and fascinating science!
Module 2.3
Measuring and Analyzing Results
Figure 2.14 Why statistics can be misleading: Both of these graphs present the same
data, an increase from 20 to 22 over 1 years time. But by ranging only from 20 to 22
(rather than from 0 to 22), graph (b) makes that increase look much more dramatic.
(After Huff, 1954)
Descriptive Statistics
Figure 2.15 Results of an imaginary survey of study habits at one college. This college
apparently has two groups of studentsthose who study as hard as they can and those who
find other things to do. In this case both the mean and the median are misleading. This
distribution is bimodal; its two modes are 0 and 8.
Descriptive Statistics
Descriptive Statistics
Descriptive Statistics
Figure 2.16 The monthly salaries of the 25 employees of company X, showing the mean, median,
and mode. (After Huff, 1954)
Concept Check:
Calculate the mean, median and mode for this distribution of
scores:
2, 3, 4, 4, 7, 10
Mean = 5
Median = 4
Mode = 4
Concept Check
Concept Check
Descriptive Statistics
Measurements of Variation
The range is a statement of the highest and lowest
scores
If our distribution has the following scores: 1, 2, 3, 5, 7,
9, 9, 10, the range is from 1 to 10.
Descriptive Statistics
Measurements of Variation
The standard deviation (SD) is a measurement of the
amount of variation among scores in a normal
distribution.
The more closely the scores are clustered around the
mean, the smaller the standard deviation is.
Standard deviations are used to make meaningful
comparisons on different tests or on different versions of
the same kind of test.
Figure 2.17 These two distributions of test scores have the same mean but different variances and different
standard deviations.
Concept Check:
On your first statistics exam of the semester, you get a score
of 90, the mean for the class is 70 and the standard deviation
is 20. On the second exam of the semester, you get an 80.
The mean for the class is 65 and the standard deviation is 5.
Did you do better, worse, or the same on the second test?
You did much, much better on exam 2!
Figure 2.18 In a normal distribution of scores, the amount of variation from the mean can be
measured in standard deviations. In this example scores between 400 and 600 are said to be within
1 standard deviation from the mean; scores between 300 and 700 are within 2 standard deviations.
Confidence Intervals
Confidence intervals are typically reported at the 90%,
95% or 99% levels of certainty.
The higher the confidence level, the broader the range
that is given by the researcher.
If the standard deviation is small and the sample on
which the confidence interval is based is large, we can
increase our certainty without necessarily broadening
the range.
FIGURE 2.19 The vertical lines indicate 95% confidence intervals. The pair of graphs in part a
indicate that the true mean has a 95% chance of falling within a very narrow range. The graphs
in b indicate a wider range and therefore suggest less certainty that reward is a more effective
therapy than punishment.
Probability Values
A probability value is a way to estimate if a score would
be extremely rare given what we know about the likely
range in which the population mean falls.
If the researcher says that there is a 95% certainty that
the population mean falls between 5.0 and 7.0, and a
particular score falls at 8.2, then that score has a
probability value of less than 5% (p < .05) and it is an
exceptional score in some way.
Figure 2.21 Researchers say that results are statistically significant if they calculate that
chance variations in data would be unlikely to produce a difference between groups as
large as the one that the researchers actually observed.
Concept Check:
Which is a more significant result:
One that is obtained with a p-value of .10 OR
One obtained with a p-value of .001?
.001