PSYC 2000 Chapter : 5 Modules 40 41 28
1
Psychology 4030 – Final Study Guide
Helping Doctors and Patients Make Sense of Health Statistics
What is collective statistical illiteracy? (p. 54) What are the 3 main points in this monograph?
Collective statistical illiteracy is the widespread inability to understand the meaning of numbers. The three main
points: 1. It is common to patients, physicians and politicians 2. It is created by nontransparent framing of information
that may be unintentional or intentional 3. It can have serious consequences for health
What is the difference between absolute and relative risk? Which type of risk doe the media tend to report (p.
54-top of p. 55)
Relative risk: 100% higher. The media tends to report this because the numbers seem larger/more dramatic
Absolute risk: 1 in 7,000
What is the difference between conditional probabilities and natural frequencies? Which is easier to
understand? Why?
Conditional probabilities: includes the sensitivity and the false-positive rate (1-specificity).
Natural frequencies: a transparent representation that can achieve the same in comparison to conditional
probabilities. Natural frequencies are easier to understand because they represent the way humans encoded
information before mathematical probabilities were invented. They are simple counts that are not normalized with
respect to base rates. In the article, the 4 natural frequencies added up to the total number of 1000 women.
Define positive predictive value.
The probability that a person has a disease given a positive screening test (the probability that a positive test actually
means that a person has the disease)
What is the key difference between survival and mortality rates? What is the lead time bias?
Survival rates: imagine a group of people who are diagnosed with cancer on the same day. The proportion of these
patients who are still alive after a given amount of years (i.e. 5). To calculate you take the number of patients
diagnosed with cancer still alive X years after diagnosis and divide is by the number of patients diagnosed with
cancer.
Mortality rates: imagine a group of people, not defined by a caner diagnosis. The proportion of people in the group
who are dead after 1 year is the mortality rate. To calculate the annual mortality rate, you take the number of people
who died from cancer over 1 year and divide it by the number of people in the group.
The key difference between these statistics is the word “diagnosed.” It appears in the formula for the survival rate,
but not the mortality rate.
Lead time bias: screening biases survival in how it affects the timing of diagnosis. For example, there is a group of
cancer patients currently diagnosed at age 67, all of whom die at age 70. The 5-year survival of the group would be
0%. If the group was diagnosed with cancer by tests earlier, i.e. at age 60, and still died at 70 – their 5-year survival
rate would be 100%. Survival rates change dramatically, but time of death remains the same.
What is the overdiagnosis bias?
Leads to high survival rates – it is the detection of pseudodisease (i.e. detect what seems like cancer, but will not
progress). Even though the survival rate has changed, the number of people who die has not changed.
What are the 4 questions we should ask about all risks?
1. Risk of what? (understand the outcome); 2. Time frame? (understand the time); 3. How big? (absolute terms – 13
out of 1000 will die by age 50); 4. Does it apply to me (see if the risk info is based on studies on people like you –
gender or age or similar health problems).
What are some of the potential harms of screening tests?
Costs, inconvenience, and false alarms – and in our view, the most important harm of overdiagnosis. Overdiagnosis
leads to hard through overtreatment.
Distinguish between false positive and false negative errors in screening tests. What is specificity and
sensitivity?
False positive: false alarm. The test is positive in people who do not have the disease
False negative: miss. The test is negative in someone who does have the disease
Specificity: the proportion of negative tests among clients without the condition
Sensitivity: the proportion of positive tests among clients with the condition
2
What seems to be a necessary precondition for minimal statistical literacy? Define the illusion of certainty.
Basic numeracy; knowing how to concert 1% to 10 in 1,000, concert 1 in 1000 to 0.1%, and how many heads in
1,000 coin flips?
Illusion of certainty: an emotional need for certainty when none exists. This feeling can be attached to test results that
are taken to be absolutely certain and to treatments that appear to guarantee a cure.
According to Kalet, Roberts, and Fletcher (1994), how many patients discussed risks and benefits with their
doctors during visits? How does this relate to people’s ability to understand basic risks?
Audiotapes of 160 adult patents’ visits to doctors in North Carolina revealed that in only one out of four visits did the
patient and doctor actually discuss risks or benefits. Only few (about one in six) of these discussions were initiated by
the patient, and in the majority of the discussions, the physician stated the risk with certainty). Of the 42 patients who
said that they actually had discussed risks with their doctors, only 3 could recall immediately after the discussion
what was said. Yet almost all (90%) felt that they had their questions answered, had understood all that was said,
and had enough information. This finding reveals that question-asking behavior was generally low and it suggests
that many patients are reluctant to ask questions, which is at odds with the goal of shared decision making.
According to research, people are most likely to have a favorable evaluation of a treatment when benefits are
described in what terms?
When the benefit of the test was presented in the form of relative risk reduction, 80% of participants said they would
likely accept the test (i.e. if you have this test every 2 years, it will reduce your chance of dying from this cancer by
around 1/3 over the next 10 years). Relative risk reduction would lead to more correct answers by patients
Name 3 problems with the quality of press coverage of health statistics.
Journalism schools tend to teach everything except understanding numbers. They fail to report any numbers, frame
numbers in a nontransparent way to attract readers’ attention, and fail to report important caution about study
limitations.
How do advertisements typically discuss the benefits and potential harms of new drugs?
When benefits of medication are quantified they are reported using only a relative risk reduction format without
providing a base rate. This leads to readers to overestimate the magnitude of the benefit. To clarify risk reduction,
present the fundamental information about the absolute risks in each group (i.e. this drug lowered the risk of
*disease* by 10 in 100: from 20 in 100 to 10 in 100 over 10 years). Harms are mentioned in only about 1/3 of reports,
and are rarely if ever quantified. Benefits are displayed in a nontransparent format, while harms are often stated in a
way that minimizes their salience. Harms are listed in long lists in very fine print. On television, more time is allotted
to the benefits than the side effects. Also, they tend to report benefits in relative risks (big numbers) and harms in
absolute risks (small numbers)
Name 2 consequences of misleading advertising.
Emotional manipulation and impediments to informed consent and shared decision-making. Ads are illustrated to try
to raise anxieties or hopes. The ads explicitly or implicitly overstate a risk, benefit, or both. Instead of calling for
universal screening, it is now recommended that patients make an informed decision after learning about the benefits
and harms. Yet, studies show that clinicians rarely communicate the uncertainties about risks and benefits of
treatments to patients.
Give 2 reasons why collective statistical illiteracy persists. Define concealed paternalism.
Statistical illiteracy persists because of the lack of training in statistical thinking in primary education and medical
training. Also, doctors rely on their skills and some see statistics as antimedical, and they must rely on intuition
because is presupposed a level of homogeneity among patients that might be appropriate for physics but were utterly
unrealistic in medicine. Also, the illusion of certainty
Concealed paternalism: extreme form in which physicians do not even inform patients about tests or treatments
performed on them.
Which statistics are used to communicate risk in transparent forms? Which are in non-transparent forms?
What is mismatched framing?
Nontransparent statistics, such as relative risks without the base rate, often appear in leading medical journals. From
these sources the numbers spread to physicians, the media, and the public. Many of these fail to report the absolute,
transparent risks. Statistics presented in natural frequencies considered transparent. Nontransparent were presented
in conditional probabilities.
Mismatched framing: readers can be misled more directly than just via nontransparent framing. In some cases,
benefits and harms of treatments are reported in different currencies (benefits in big numbers and risks in small)
3
What are the author’s four recommendations on probability formatting and frequency formatting?
1. Use frequency statements, not single event probabilities 2. Use absolute risks, not relative risks 3. Use mortality
rates, not survival rates 4. Use natural frequencies, not conditional probabilities
Bink (Gladwell, 2005)
What is the finding of the “Iowa” experiment? What is the adaptive unconscious?
The Iowa experiment is a gambling game involving 4 decks of cards: 2 decks of red and 2 blue. The object of the
game is to turn over the cards in a way to maximize your winnings. The red gave you large rewards, but also large
losses. They found that after about 50 cards are turned, we start to figure out what is going on. After 80 card turns,
we are able to figure out what is going on and explain the types of the decks. In the experiment they hooked up
participants to a machine that measured the activity of sweat glands (i.e. palms). The readings from the sweat activity
suggests that after about 10 card turns, the participants started to produce a stress response to the red decks. Also
around this time, participants began to favor the blue decks and took less from the red decks. There are two
strategies our brain uses to make sense of the situation – one is conscious, slow and requires information, and the
other operates quickly but below the surface of the conscious and sends us messages
The adaptive unconscious is the part of our brain that jumps to conclusions (make quick judgments with little
information). It is thought of as a giant computer that quickly and quietly processes a lot of the data we need in order
to keep functioning as human beings.
What kind of research occurs in the “love lab” and how accurate are their predictions about the success of
relationships? What is “thin-slicing”?
Gottman observes the couples talking and reads/watches their facial expressions. Heart rate, amount of sweat,
temperature, and how much they moved around was measured. Each person had a camera recording their actions
and words. If Gottman observes for an hour, he can predict with 95% accuracy whether that couple will still be
married fifteen years later. If he watches for about 15 minutes, his accuracy is 90%. Gottman and a professor that
works with him figured out that they could predict this with only 3 minutes.
“Thin-slicing” is the ability of our unconscious to find patterns in situations and behavior based on very narrow slices
of experience.
What is Gottman’s marriage survival rule with respect to expressed emotion?
The ratio of positive to negative emotion in a given encounter has to be at least five to one. Also, all marriages have
a distinct pattern, and it shows up in any meaningful interaction.
According to Gottman, what are “the four horsemen”?
Defensiveness, stonewalling, criticism, and contempt. Within the Four Horsemen, there is one emotion that he
considers the most important of all – contempt (a lot of the time it is in the form of an insult).
What were the results of Gosling’s study on personality judgments? Why is it sometimes helpful to have
less information?
Judging people’s personalities is a really good example of how surprisingly effective thin-slicing can be. Used the Big
Five Inventory – our friends can describe us fairly accurately. He had strangers look at dorm rooms and then rated
them on the Big Five Inventory. The dorm room observers weren’t as good in measuring extraversion (same for
agreeableness). The remaining three – they were more accurate at measuring conscientiousness and they were
much more accurate at predicting both the students’ emotional stability and their openness to new experiences. The
strangers did a much better job. You can learn as much, or more, from one glance at a private space as you can
from hours of exposure to a public face.
When you have less information, it can be helpful because you get to avoid confusing and complicated an ultimately
irrelevant pieces of information that can screw up your judgment.
What features predict whether a doctor will be sued for malpractice?
The relationship between the doctor and his patients. Surgeons who had never been sued spent more than three
minutes longer with each patient than those who had been sued. Explaining procedures and leaving time for the
patient ask questions. More likely to engage in active listening. If a surgeon’s voice was judged to sound dominant,
the surgeon tended to be in the sued group. The less dominant and more concerned voice tended to be in the non-
sued group. It all boils down to respect – respect that is communicated through tone of voice.
