Common Misperceptions in Probability among Students
in an Elementary Statistics Class
By Dorothy Anway, Ph.D. and Erik Bennett, Senior
Department of Mathematics and Computer Science
University of Wisconsin-Superior
Abstract
Students come to the study of probability with several common misperceptions. For his
senior project, the senior student worked with the faculty member to design a project
studying some of these misperceptions among students in an elementary statistics class.
The project involved three elements: measuring the misperceptions of the students,
teaching to correct these misperceptions, and measuring the improvement. The student
created a survey instrument used for the pretest and the posttest from elements the
ARTIST database. The three hour teaching unit used interactive activities which
addressed the specific concepts on which the students performed poorly. Concurrently,
the students attended lecture classes taught by the instructor on the theories of
probability. After this unit, the students were tested again and the results compared with
the pretest.
Introduction
This study grew out of a previous study done by the senior student on misperceptions that
students bring to a probability and statistics course. In the prior study, certain
misconceptions in probability were found to be prevalent among a sample of college
students. This study was initiated to specifically target and eliminate two of the common
misperceptions, the equiprobability bias and the representativeness misconception.
The equiprobability bias is the tendency of students to view several outcomes of an
experiment as equally likely. For instance, students who have an equiprobability bias
think that when two dice are rolled, all the sums possible are equally likely. They do not
realize that the sum of 6 for the two dice is more probable than the sum of 2.
The representativeness misconception refers to the tendency of students to erroneously
think that samples which resemble the population distribution are more probable than
samples which do not. Students with this misperception will think that a sequence of
coin tosses that has approximately equal numbers of heads and tails is more probable than
a sequence with many more heads than tails. However, the probability of each sequence
is the same. People with the representative misconception expect that a small sample
2
will have the at least approximately the same distribution of outcomes as the population
distribution.
The group of students used in this study was a class of Math 130, Elementary Statistics,
at the University of Wisconsin-Superior. The class is a requirement for students majoring
in social work and legal studies, and is recommended for biology majors who do not take
calculus. It also meets the liberal education mathematics requirement, and so usually has
some students with other majors also. In general, these students are not strong in
mathematics, but see the course as a requirement for success in their fields of study, and
so they take the course seriously and work hard at it. There were 27 students who took
both the pretest and the posttest. The study was designed to identify the misperceptions
in this class then teach to correct these. The class consists of three 50-minute class
sessions and a laboratory of 75 minutes per week. The class sessions meet in a traditional
classroom, and consist primarily of lectures, question and answer sessions, and activities
sessions. The laboratories meet in a computer classroom and primarily focus on learning
Minitab software and reinforcing the lecture concepts. During this experiment, group
activities were added to the usual computer sessions. These activities were designed to
break down the misconceptions the students held at the beginning of the unit.