PROBABILITY PREVIEW
FIVE COMMON PROBABILITY MISCONCEPTIONS
Gambler’s Fallacy
Sample Representativeness
Conditional Probability
The Outcome Interpretation of Probabilities
Equiprobability Bias
Gambler’s Fallacy (or the Law of Averages)
The Misconception:
This is the belief that chance processes are self-correcting in the short run. Suppose one
observes 8 heads out of tosses. What is the probability that the next toss is a head? If
one believes in the Gambler’s fallacy, then one would think that the next toss is more
likely to be a head. Similarly, a gambler with this belief who is going through a bad
losing streak at a casino thinks he’s going to win in future tosses.
The Truth:
Actually, outcomes in coin tosses and different gamblers are independent. This means
that the number of heads in early tosses will have no impact on the chance of heads in
later tosses. In the gambling situation, the outcomes of repeated games are independent.
The gambler’s success or lack of success in early games has no influence or effect on the
chance of winning games in the future.
Sample Representativeness (or the Law of Small Numbers)
The Misconception:
46
We’ll learn about the notion of independence in Topic P4 (Conditional Probability).
This is the belief that a small sample will exactly resemble the characteristics of the
larger population. This is a problem that illustrates this misconception.
A certain town is served by two hospitals. In the larger hospital about 45
babies are born each day, and in the smaller hospital about 15 babies are
born each day. As you know, about 50 percent of all babies are boys.
However, the exact percentage varies from day to day. Sometimes it may
be higher than 50 percent, sometimes lower.
For a period of 1 year, each hospital recorded the days on which more than
60 percent of the babies born were boys. Which hospital do you think
recorded more such days?
(a) The larger hospital
(b) The smaller hospital
(c) About the same (that is, within 5 percent of each other)
Most people are insensitive to the sample size and answer (c). They believe that the
pattern of births each day will resemble the population of ``half boys, half girls” and the
pattern for small samples (such as small hospital) will be the same as the pattern for large
samples (such as the large hospital).
The Truth:
Actually, it is common for a sample to not resemble the population. Also, small samples
tend to be more variable than large samples. In the example, one would expect the
number of days with more than 60 percent boys to be higher for the small hospital than
for the large hospital.
Misconception in Coin Tossing
With the representativeness belief, samples that resemble the population distribution are
believed more likely than samples that do not. So, for example, if you flip six coins,
47
then one may believe the sequence HTHTHT is more likely than HHHHHH, since the
first sequence is equally divided between heads and tails.
The Truth
Actually, in the above coin tossing example, all sequences such as HTHTHT or
HHHHHH or TTTHHH or HHTTHH are equally likely.
Conditional probabilities
There is a lot of confusion with conditional probabilities. Here is one example of
conditional thinking where students reason incorrectly.
An urn has two white balls and two black balls in it. Two balls are drawn out without
replacing the first ball.
1) What is the probability that the second ball is white, given that the first ball
was white?
2) What is the probability that the first ball was white given that the second ball
is white?
Most students correctly state that the probability for (1) is 1/3, but state the probability for
(2) as 1/2. Students appear to take a causal approach to thinking about conditional
probabilities. They appear to understand how the second ball has a causal relationship
with the outcome of the first, but find it difficult to entertain a contingent relationship
between an event and an outcome that came after it.
In addition, there is confusion between P(A|B) and P(B|A) and students frequently mix
up the two conditional probabilities.
48
We’ll get some experience with real coin tossing in Topic P6 (Coin Tossing
Distributions)
We’ll explore conditional probability in Topic P4.
The Outcome Interpretation of Probabilities
One faulty view of probabilities is called the “outcome approach.” People will interpret
the weather forecast “there is a 70% chance of rain” as the prediction “it is going to rain
today.” Chance statements are interpreted in absolute terms – either it is going to rain or
it is not going to rain. People who are outcome oriented will tend to interpret probability
statements that are anchored at three reference points: 0% chance, 50% chance, and
100% chance. A statement of 50% chance rain is interpreted to indicate complete
ignorance of outcomes. A statement of 70% chance is closer to a 100% chance and
interpreted as the prediction that it will rain. Similarly, a statement of 30% chance is
interpreted as the prediction that it will not rain. In this view, the forecaster is “right” or
“wrong” depending on the weather result (rain or no rain). There is no appreciation of
the uncertainty that is implied by a forecast of 30% rain or 70% rain.
The Equiprobability Bias
This is people’s tendency to view several outcomes of an experiment as equally likely. If
there are two possible outcomes, each outcome will be assigned a probability of ½, and if
there are three outcomes, each outcome has a probability of 1/3. There are situations
where the equiprobability assumption is reasonable, but the student with this bias will use
this assumption for many situations where it is clearly inappropriate.
We’ll explore different interpretations of probability in Topic P1 (Probability: A
Measure of Uncertanty)