18
C H A P T E R 2
Conducting Health Research
Lecture Outline
I. The Placebo in Treatment and Research
A placebo is an inactive substance or condition that can cause people to improve or change
their behavior. Both expectancy and learning contribute to this effect.
A. Treatment and the Placebo
The placebo effect is an advantage for treatment, boosting its effectiveness. This effect
may be responsible for about 35% of improvements, but some treatments (such as
antidepressant drugs) have higher placebo effects than others (such as treatment for broken
bones). Placebos can also produce negative effects, called the nocebo effect. Both placebo and
nocebo effects are real and produce symptoms and relief that are indistinguishable from the
effects obtained from drug and other physical treatments.
B. Research and the Placebo
II. Research Methods in Psychology
For research to contribute to knowledge about health, scientists should be familiar with one
another’s work, use controlled methods, try to keep personal biases from contaminating results,
make claims cautiously, and replicate their studies. To understand behaviors related to health and
disease, researchers use a variety of methods, including correlational studies, cross-sectional and
longitudinal studies, experimental designs, and ex post facto designs.
A. Correlational Studies
Correlational studies indicate the degree of relationship between two variables, such as
the number of stressful life events and the risk of heart attack. This approach is one type of
descriptive research. In correlational studies, the relationship between two variables is
expressed in terms of correlation coefficients. Coefficients range from -1.00 to +1.00, with
numbers closer to ±1.00 indicating stronger relationships. Attributions of cause and effect are
not possible from correlational studies
B. Cross-Sectional and Longitudinal Studies
Cross-sectional studies are conducted at one point in time and compare people of
Conducting Health Research
19
cardiovascular disease (the dependent variable) in the two groups. Well-controlled
experimental designs give scientists their best method to determine causation.
D. Ex Post Facto Designs
Ex post facto designs are similar to experimental designs in that both use contrasting
groups, but these designs do not include manipulation of independent variables. Instead,
III. Research Methods in Epidemiology
Epidemiology is a branch of medicine that investigates factors contributing to the
occurrence of a disease in a particular population. Epidemiology evolved into a scientific
discipline during the 19th century and played an important role in the fight against infectious
diseases. Today, epidemiologists also study factors associated with chronic illness, including its
prevalence and incidence. Prevalence is the proportion of the population affected by a particular
disease at a particular time, whereas incidence is the number of new cases of a disease during a
particular time, usually one year. Research methods used in epidemiology are similar to those
employed by psychologists. Epidemiology research falls into three broad areas: observational
methods, natural experiments, and experimental investigations.
A. Observational Methods
Observational methods parallel correlation studies in psychology. The two types of
observational methods are retrospective and prospective. Prospective studies are
B. Randomized, Controlled Trials
Randomized, controlled trials are equivalent to experiments in psychology. Clinical trials
are randomized, controlled trials designed to test the effectiveness of a drug or treatment.
Researchers assign participants to groups randomly to control for self-selection. Randomized,
placebo-controlled double-blind designs are often considered the zenith of research design.
C. Natural Experiments
Natural experiments are similar to ex post facto studies in psychology; both involve the
selection rather than the manipulation of a variable. Natural experiments can be conducted
when two similar groups of people naturally divide themselves into those exposed and those
not exposed to a pathogen.
D. Meta-analysis
The statistical technique of meta-analysis allows researchers to evaluate many research
studies on the same topic, even if the research methods differed. In addition, meta-analysis
provides an index of the size of the effect, which allows researchers to gauge the importance
of the effect.
E. An Example of Epidemiological Research: The Alameda County Study
Conducting Health Research
20
Epidemiology provides techniques for taking a first look at a health-related problem. An
important example is the Alameda County Study, which began in 1965 in an effort to identify
IV. Determining Causation
Most epidemiological studies do not prove causation, rather, they point to specific risk factors
that are associated with a particular disease or disorder.
A. The Risk Factor Approach
A risk factor is any characteristic or condition that occurs with greater frequency in
people with a disease than it does in people free from that disease. Although they do not
determine causation, risk factors can be used to estimate the probability of developing a
disease. Scientists using the risk factor approach must distinguish relative risk from absolute
risk. Relative risk is the ratio between the number of people in an exposed group with a
disease and the number in an unexposed group with the disease. Absolute risk refers to a
person’s chance of developing a disease independent of the risk for other people.
B. Cigarettes and Disease: Is There a Causal Relationship?
Epidemiologists have gradually uncovered evidence linking cigarette smoking to lung
cancer and heart disease. Although scientists have no evidence from experimental designs
proving that cigarette smoking causes heart disease and lung cancer in humans, the
accumulation and consistency of other evidence has led researchers to conclude that cigarette
V. Research Tools
Theories and psychometric methods are tools that researchers use to determine areas of
inquiry and to measure variables essential to their research.
A. The Role of Theory in Research
A theory can be defined as a set of related assumptions that allow scientists to use logical
deductive reasoning to formulate testable hypotheses. Theories are not idle speculations but
Conducting Health Research
21
B. The Role of Psychometrics in Research
Health psychologists, like other scientists, use measuring instruments to test their
hypotheses and to build their theories. To be useful, measuring devices must be both reliable
(consistent) and valid (accurate).
1. Establishing Reliability
The reliability of any measuring tool is the extent to which that tool measures in a
consistent fashion. Reliability can be established through test-retest and interrater
methods. Reliability is usually expressed in terms of correlation coefficients.
Exploring Health on the Web
The Internet offers exciting possibilities for students to access recent information and
research on health.
Suggested Activities
Personal Health Profile — Reading the Research Literature
The growth of health psychology has resulted in a growing number of psychology studies
on topics related to physical health. In Chapter 1 of this manual, we suggested a course
assignment in which students compile a Personal Health Profile. To complete this series of
assignments, they must become familiar with the current literature on health-related behaviors.
However, much of this research is reported in journals in the fields of medicine, nursing,
rehabilitation, and epidemiology. Psychology students may have little experience reading
journals from these other disciplines (and many psychology students have difficulties in reading
research in psychology journals).
Conducting Health Research
22
one in a psychology journal, one a medical journal, and one an epidemiology journal. What are
the differences and similarities in the format among the three? What are the sources of these
differences? For example, does the psychology journal concentrate on behavioral manipulation
and measures whereas the medical journal emphasizes physiology, or has the interdisciplinary
collaboration of researchers from various areas led to a blurring of the differences among content
areas in health research?
Debate: The Ethics of Placebo Treatment
Urge your students to explore the ethics controversy surrounding placebo treatments. As
the text explains in the “Would You Believe. . . ?” box for this chapter, placebo treatments—
even placebo surgery—are considered ethically acceptable according to current medical and
psychological ethical standards. However, some researchers (and many members of the general
public) find this practice unacceptable.
The controversy has made its way into many journals, so your students have published
sources from which to draw. However, people often have strong opinions. Arrange a debate on
this issue, considering the experts’ opinions and personal opinions of your students.
Medical Research and Media
Assign your students the project of finding media errors or distortions in coverage of
medical research, which they can find by consulting one of the online news sites, such as CBS’s
Health section or the comparable section of many news sites. It is not even necessary to use the
Internet; television, newspapers, and magazines carry health-related stories. The emphasis for the
media is to grab attention and present a concise summary of an interesting research finding. The
result is often to sensationalize and simplify research findings.
Direct your students to search for the original report of the misrepresented findings and
contrast the technical report with the popular coverage, making exhibits of the contrasting
Understanding the Numbers Behind the News
Most people (including students) have trouble understanding the results of health
research and how those results may apply to them. Part of the difficulty stems from the way that
results are reported and the inclusion of numbers that are essential to research and statistical
analysis. For example, one confusion involves the terms absolute risk and relative risk, both of
Conducting Health Research
23
which are terms from epidemiology research. If a person’s risk for a disease is expressed in
terms of absolute risk, the risk seems very small, but if expressed in terms of relative risk, the
risk seems much larger.
Computing a Correlation
Instructors may wish to have each student contribute data to a correlational study on
weight and hypertension. Each student could collect information on height, weight, frame size,
and blood pressure level from four or five acquaintances. The study will be more representative
and a better illustration of correlation if middle-aged or older participants are included. Blood
pressure can be reported in terms of systolic level, diastolic level, or both. When the data are
gathered, a Pearson product-moment correlation can be calculated. Any standard statistics text
will provide the formula, and access to a computer program (such as Excel) will make the
calculation easy.
Whether the resulting correlation coefficient is positive or negative, large or
nonsignificant, students may offer explanations for the reason the correlation turned out the way
it did. They should remember that a positive correlation between weight and blood pressure does
not indicate that weight causes hypertension, but this type of research is the basis for
establishing risk factors for many diseases.
Correlations Do Not Equal Causation
Designing a Research Study
One study described in this chapter is an observational study examining whether posted
calorie counts of food influences food choice (Ebel, Gyamfi, & Kersh, 2011). Although this
study design was observational, similar questions about food choices and calorie counts could be
conducted using a variety of methodologies. In small groups, each group could design a follow-
up study using a different research design. Students should be able to answer specific questions
Conducting Health Research
24
about their study, including: What is the research design of your study? What is your hypothesis?
What are the observational definitions of your variables? What results would support your
hypothesis?
Then by sharing these studies during class, other groups may also be able to identify
possible confounds as well.
Video Recommendations
From Films for the Humanities & Sciences:
Placebo: Mind over Medicine? (2002) explores the placebo effect in treatment and research,
presenting case studies of its effectiveness and interviews with health care professionals who
comment on the placebo effect.
From Filmmakers Library:
Plagued: A Series on Disease and Society (1994) is a 4-part series on epidemics that attempts to
put the AIDS epidemic into global and historical perspective. In this attempt, parts of this series
provide information about epidemics and epidemiology. Part 1: The Origin of Disease
investigates how epidemics break out. Part 2: Epidemics concentrates on the historical
perspective of the spread of disease, including bubonic plague and cholera. Part 3: Invisible
Armies explores the relationship between the immune system and history. Part 4: Will We Ever
Learn? explores the relationship between HIV and syphilis.
Videos from the Web:
Multiple Choice Questions
1. An inactive substance or condition that has the appearance of the independent variable
and that may cause participants in an experiment to improve or change behavior due to
their belief is called
a. a nocebo.
b. a placebo.
Conducting Health Research
c. a dependent variable.
d. an experimental design.
2. The effect of a placebo varies according to people’s expectancies and previous learning,
but in general, that response rate is about
a. 5%.
b. 10%.
c. 35%.
d. 60%.
3. Which of these conditions is likely to produce the highest positive placebo effect?
a. a physician dressed casually in blue jeans and sneakers
b. a physician who is enthusiastic in describing the treatment
c. a physician with a reputation for medical errors
d. a physician who, when prescribing medication says, “This may not help, but it
won’t hurt.”
4. The placebo effect is
a. most prominent in well-designed experiments.
b. an imaginary effect applicable to everyone.
c. an imaginary effect that occurs almost exclusively in hypochondriacs.
d. physiologically real and capable of improving organic and psychological
symptoms.
5. Which of these “sugar pills” is likely to have the greatest positive effect?
a. white pills rather than colored pills
b. very small pills rather than middle size pills
c. capsules rather than tablets
d. generic pills rather than brand-name drugs
6. Placebos have been known to help
a. reduce insomnia.
b. decrease low back pain.
c. lower high blood pressure.
d. bowel movements.
Conducting Health Research
26
7. To determine whether Drug Z lowers blood pressure, it is necessary to demonstrate that
an experimental group, which has been given Drug Z, will have lower blood pressure
than a comparison group, which has been given
a. a higher dose of Drug Z.
b. a lower dose of Drug Z.
c. a placebo.
d. no treatment at all.
8. Which of these statements is true?
a. Placebo effects can influence both psychological and physical disorders.
b. Valuable research is done by people outside the scientific community, but
scientists try to discount the importance of this research.
c. Scientific breakthroughs happen every day.
d. Experimental rather than observational research is required to learn about patterns
of disease.
9. The placebo effect
a. makes it easier to determine the effectiveness of a therapeutic intervention.
b. does not produce any type of unfavorable effect.
c. is easy to control in psychotherapeutic treatment.
d. hampers the evaluation of the effectiveness of treatment programs.
10. When neither the participants nor the experimenters know which group has received the
treatment and which has received a placebo, the design is called
a. confounding.
b. double-blind.
c. correlational.
d. naturalistic.
11. The nocebo effect occurs when
a. participants in a placebo study experience a negative effect.
b. participants in a placebo study experience a positive effect.
c. experimenters use a double-blind study.
d. experimenters use the case-control method.
12. Placebos can be beneficial in treating many conditions EXCEPT:
a. depression
b. hypertension
c. insomnia
d. broken bones
13. The _______ a placebo resembles an effective treatment, the _____ the placebo effect.
Conducting Health Research
a. more; stronger
b. more; weaker
c. less; stronger
d. less; weaker
14. Dr. Smith, a clinical health psychologist, is conducting research on whether relaxation
training before a swim meet improves swimmer’s race times. It is most likely Dr. Smith
is using a __________ design.
a. correlational
b. single-blind
c. double-blind
d. retrospective
15. Most health-related evidence
a. is the result of a variety of research methods.
b. has been discovered accidentally.
c. is withheld from the general public to avoid widespread panic.
d. comes from the results of experimental designs.
16. Dr. Rich is conducting research that examines whether 20-year olds eat more low-fat
foods than 70-year olds. This research is using a ________________ design.
a. cross-sectional
b. experimental
c. ex post facto
d. retrospective
17. Dr. Sweedey is conducting research that examines 5 year olds and their aggressive
behaviors. She follows up with these same participants 10 years later to measure their
reactivity to stress. She is using a ___________ design.
a. cross-sectional
b. experimental
c. longitudinal
d. ex post facto
18. There is a ____________ correlation between physical age and physical health, such that
as age increases, physical health decreases.
a. positive
b. negative
c. weak
d. non-existent
Conducting Health Research
19. A correlation of .80 would indicate a _______ and __________ relationship between two
variables.
a. strong; positive
b. strong; negative
c. weak; positive
d. weak; negative
20. A positive correlation between physical health and education would indicate that as
education __________, physical health ______________.
a. decreases, decreases
b. decreases, increases
c. increases, decreases
d. increases, increases
21. The major difference between longitudinal studies and cross-sectional studies is that
cross-sectional studies occur ___________ whereas longitudinal studies occur
_______________.
a. once; over time
b. over time; once
c. with the same participants; with different participants
d. with the same participants; over time
22. One of the advantages of conducting a longitudinal study is that longitudinal studies can
_____________________.
a. be completed quickly.
b. help identify developmental trends and patterns.
c. determine causality.
d. be completed with relatively few researchers.
23. A recent research study examined whether sending text messages to overweight children
would help enable them to maintain a weight-loss program. In this experiment, some
children were randomly assigned to receive text messages and other children were
assigned to the control condition and did not receive text messages. The dependent
variable in this experiment was
a. continued enrollment in a weight-loss program.
b. text messages.
c. weight loss.
d. not receiving text messages.
Conducting Health Research
29
24. Researchers want to examine whether writing about the experience of breast cancer
increases feelings of hope and optimism. Breast-cancer patients were randomly assigned
to one of two essay conditions: writing about their breast cancer or writing about their
everyday tasks. The independent variable in this experiment would be:
a. hope and optimism
b. breast cancer diagnosis
c. writing about everyday tasks
d. the essay conditions
25. One of the challenges of conducting health research is that it is impossible, and unethical,
to manipulate certain variables, such as sexual behaviors or smoking. The research design
that examines differences between groups without manipulating these groups is:
a. experimental
b. ex post facto
c. random assignment
d. longitudinal
26. If two variables increase or decrease together, they are
a. positively correlated.
b. negatively correlated.
c. positively skewed.
d. negatively skewed.
27. Correlational studies
a. cannot indicate cause and effect.
b. are a type of experimental design.
c. cannot be used in psychological research.
d. cannot be used in epidemiological research.
28. Correlational studies are a type of
a. experimental study.
b. double-blind study.
c. ex post facto design.
d. descriptive research.
29. Small correlations, for example 0.08 or 0.10,
a. may be statistically significant.
b. cannot be statistically significant.
c. indicate cause and effect between the two correlated variables.
Conducting Health Research
30
d. both a and c
30. Cross-sectional studies
a. follow disease-free participants over a long time period.
b. follow participants with a disease over a long time period.
c. are also called longitudinal.
d. compare participants of at least two different age groups or developmental
periods.
31. A study that compares cholesterol levels of 10-year-old children and 30-year-old adults
would most likely be a(n) _____ study.
a. experimental
b. cross-sectional
c. longitudinal
d. ex post facto
32. A study that follows the history of overweight male participants over a 30-year period
would be
a. an experimental study.
b. a longitudinal study.
c. a cross-sectional study.
d. a case-control study.
33. Which of these is NOT a characteristic of longitudinal studies?
a. Longitudinal studies are a comparison of two separate groups.
b. Longitudinal studies tend to be time consuming.
c. Longitudinal studies frequently require a team of researchers
d. Longitudinal studies are prospective designs
34. An investigator measures blood pressure in a group of college students and then repeats
these measurements every year for 20 years. This is an example of
a. a longitudinal study.
b. a cross-sectional study.
c. an experimental study.
d. a clinical trial.
35. Although causality is difficult to determine, which scientific method most strongly yields
evidence for cause and effect relationships?
a. correlational study
b. experimental design