Peer instruction: Getting students to think in class

Eric Mazur

Citation: 399, 981 (1997); doi: 10.1063/1.53199

View online: http://dx.doi.org/10.1063/1.53199
View Table of Contents: http://aip.scitation.org/toc/apc/399/1
Published by the American Institute of Physics
 Peer Instruction:
Getting Students to Think in Class
Eric Mazur

Harvard University, Cambridge, MA

The first time I taught introductory physics, I spent much time preparing
lecture notes, which I would then distribute to my students at the end of each
lecture. The notes became popular because they were concise and provided a
good overview of the much more detailed information in the textbook.
Halfway through the semester, a couple of students asked me to distribute the
notes in advance so they would not have to copy down so much and could pay
more attention to my lecture. I gladly obliged, and the next time I was teaching
the same course, I decided to distribute the collected notes all at once at the
beginning of the semester. The unexpected result, however, was that at the end of
the semester a number of students complained on their questionnaires that I was
lecturing straight out of my lecture notes!
Ah, the ingratitude! I was at first disturbed by this lack of appreciation but
have since changed my position. The students had a point: I was indeed lecturing
from my lecture notes. If they had read the textbook, they might also have noticed
that my lecture notes closely followed the material in the book. Later research
showed that my students were deriving little additional benefit from hearing me
lecture if they had read my notes beforehand. Had I lectured not on physics but,
say, on Shakespeare, I would certainly not spend the lectures reading plays to the
students. Instead, I would ask the students to read the plays before coming to
lecture and I would use the lecture periods to discuss the plays and deepen the
students' understanding of and appreciation for Shakespeare.
Year after year, I had written on the blackboard that pressure is defined as
force per unit area--a definition that is printed in the book and in my lecture notes.
Year after year the students copied it from the blackboard into their notebooks.
What a waste of time, both for the students and the teacher! What inefficiency!
And the students and I believed this lecturing constituted 'teaching.' What a
fallacy!
In most introductory science courses we require the students to buy textbooks
of encyclopedic dimensions and then we use lecture time to present what is printed
in the text. At best, the textbook is there to clarify the material introduced in
lecture. Small wonder, then, that the attendance at introductory science lectures is
relatively low compared to lectures in the humanities. And small wonder that
student opinions of introductory science lectures are very poor.
CP399, The Changing Role of Physics Departments in Modern Universities: Proceedings of ICUPE,
edited by E. E Redish and J. S. Rigden
© 1997 The American Institute of Physics 1-56396-698-0/97/$10.00
981
 In these days of overhead projectors, video cassette recorders, multimedia
computers, and the world-wide web, books may strike some as outdated teaching
aids. Yet the truth is that, at least in introductory science, we have never really
used textbooks to their full potential. We write the material on the blackboard and
students copy it into their notebooks. If we are lucky they can follow the first
fifteen minutes of lecture. If they lose the thread somewhere---and this is bound to
happen sooner rather than later--note taking becomes completely blind: "I'11 think
about it later." Unfortunately the thinking is not always happening, and many
students resort to memorization of the equations and algorithms copied in their
notebooks. Many bad study habits are a direct result of the lecture system.
The surprising similarity between lecture and sermon suggests that the lecture
dates back to quite ancient times. There is no doubt that the lecture system
predates the invention of the printing press. After all, before the mechanization of
book printing, lectures were the only efficient method to transmit knowledge. The
ideas of theologians and scholars were dutifully reproduced by scribes. In the 13th
century, as the center of intellectual life moved from courts and monasteries to
universities, professional scribes became the principal creators of books. As it had
been since the ancient Egyptians, the printed word was the only way to accurately
preserve human knowledge. While book priming in Europe dates back to the
middle of the fifteenth century, it was not until the middle of the nineteenth century
that fast, mechanized book printing turned print into a mass medium. So at least
until then, lectures and note-taking were necessary for the transmission of
knowledge.
The main reason we are still using this method is habit: we tend to teach the
way we were taught. Since my teachers lectured to me, I lectured to my students,
and so will they eventually lecture to their students. Yet everyone will agree that
for getting information listening is not as efficient as self-paced reading. While
listening is largely a passive activity, reading more easily engages the mind and it
allows more time for the imagination to explore questions. Besides, an author has
more time than a lecturer to choose the best possible wording to convey an idea.
Am I suggesting that we stop teaching altogether? That we simply ask
students to read books instead of coming to lecture? Certainly not. What I am
suggesting is that in the sciences, as is done in the humanities, the first exposure to
new material comes from reading printed material before the lecture period.
Lectures can then be used to give students a sense of what is most important in the
material they have read, to relate this material to previously studied material, to
check conceptual understanding, to paint a broader picture, to relate theories to
observations, to provide a different perspective, or even to lecture on points not
covered in the reading.
There are a number of problems with this method. First of all, in most large
introductory science classes neither teachers nor students expect any preparation
using printed material. Students have come to expect what teachers are
accustomed to giving: a lecture. It will take a considerable effort to change this

982
deeply ingrained habit. Second, reading a science text book is quite different from
reading a novel. Most students at first tend to read their text books too quickly--
without pausing or pondering the meaning of what they have just read. Perhaps
the method I am advocating will require a change in the way science textbooks are
written. Third, if one doesn't lecture during class time, what does one do?
During the past five years I have tried to address these problems by radically
changing my teaching strategy. First, I assign the students pre-class reading for
each lecture period. To make sure the students carry out this important
assignment, I begin each and every lecture period with a five-minute mini quiz on
the material they have read. I then divide the remainder of the class time into ten-
to fit~een-minute long periods, each devoted to one of the main points of the
reading. I might begin each such period with a very brief lecture on a point I wish
to get across or with a lecture demonstration. This is followed by a conceptual
question, which tests the students' understanding of the idea or point presented. I
project these multiple-choice questions, which I call ConcepTests, onto a screen
and give the students one minute to select an answer. Each student individually
must commit to an answer--I do not allow the students to speak to each other
during this minute. After the students have recorded their answer, I ask them to
try to convince their neighbors of their answer. The ensuing discussions are
surprisingly animated. After a minute or so, I again ask the students to select an
answer (one can use a show of hands, flashcards, scanning forms, or a
computerized voting system). The proportion of students who chose the correct
answer always increases after the discussion, suggesting that the students are
successfully explaining their reasoning, and in the process teaching are each other.
If about half the students select the right answer (with the correct reasoning)
before discussion, a minute or so of discussion is sufficient to dramatically improve
the level of understanding of the class. No lecturer, however engaging and lucid,
can achieve this level of involvement and participation simply by speaking.
I have successfully applied this method to large classes of about 250 students.
The results are very encouraging. Attendance is high. What is more, attention and
student involvement are high. And the answers to the ConcepTestsprovide instant
feedback to the teacher; there is never a gulf between the class' understanding and
the teacher's expectation. But best of all, testing shows this teaching style
engenders a better understanding of the fundamental concepts and discourages a
number of bad study habits such as rote memorization and an exclusive focus on
problem solving. The students' energy and enthusiasm during the discussions are
contagious: once one has experienced it, it is difficult to revert to lecturing to a
passive and mostly silent audience.
I now believe the days of straight lecturing in introductory science courses are
numbered--we can no longer afford to ignore the inefficiency of the traditional
lecture method, regardless of how lucid or inspiring our lectures are. The time has
come to offer our students in introductory science classes more than a mere
regurgitation of printed material.

983
 SAMPLE LECTURE

As an example of Peer Instruction, let's consider a 90-minute lecture on

Newton's laws, the outline of which is:

1. Newton's first law

2. Definitions of force and mass
3. Newton's second law
4. Newton's third law

Before coming to class, students are required to read the lecture notes as well
as corresponding sections in the textbook. At the beginning of class, they
complete the short reading quiz shown in Figure 1. Note that this quiz tests only
whether or not the pre-class reading was done; it does not test understanding of
the material because doing so would penalize (and therefore discourage) the
student who does the reading but is unable to master the concepts from the
reading.

Figure 1. Pre-class reading quiz for lecture on particle dynamics.

1. Which of these laws is not one of Newton's?
1. To every action there is an opposed equal reaction.
2. F=ma.
3. All objects fall with equal acceleration.
4. In the absence of a net external force, objects at rest stay at rest and objects
in uniform motion stay in uniform motion

2. The law of inertia

1. is not covered in the reading assignment.
2. expresses tendency of bodies to maintain their state of motion.
3. is Newton's third law.

3. "Impulse" is
1. not covered in the reading assignment.
2. another name for force.
3. another name for acceleration.
The correct answers are 1-3, 2-2, 3-1. Response statistics: la: 15%, lb: 2%, lc: 83%,
ld: 0%, 2a: 1%, 2b: 98%, 2c: 1%, 3a: 82%, 3b: 16%, 3c: 2%. These and subsequent
statistics are from a representative semester during which Peer Instruction was used.

I use the same lecture notes I used when I taught this material conventionally.
I describe the scope of classical mechanics and introduce Newton's first law by
writing it on the chalkboard. After introducing the first law, I use a computer
animation to show that it is really a statement about reference frames. Next, to
firmly establish the relationship between forces and acceleration, I project the
ConcepTest question shown in Figure 2. The students generally do well on this
question, and its main purpose is to bolster their confidence. In any case, I don't

984
dwell too long on this topic as Newton's other two laws generally cause far
greater difficulties.

Figure 2. ConcepTeston Newton's first law.

A car rounds a curve while maintaining a constant speed.

Is there a net force on the car as it rounds the curve?

1. No, its speed is constant.

2. Yes.
3. It depends on the sharpness of the curve and the speed of the car.
Choice 2 is correct. Response statistics: 1: 3%, 2: 96%, 3: 1%.

Then I define the concepts of force and mass and formulate Newton's second
law. To make sure that the relationship between force, acceleration, and speed is
clear, I use the question shown in Figure 3. The statistics under the Figures show
how the convince-your-neighbors discussion increases the number of correct
responses and bolsters the students' confidence. With nearly 20% of the students

Figure 3. ConcepTeston force.

A constant force is exerted on a cart that is initially at rest on an air track. Friction
between the cart and track is negligible. The force acts for a short time interval and
gives the cart a certain final speed.

force

To reach the same final speed with a force that is only half as big, the force must be
exerted on the cart for a time interval

1. four times as long as

2. twice as long as
3. equal to
4. half as long as
5. a quarter of

that for the stronger force.

Choice 2 is correct. Response statistics before (after) discussion: 1: 16% (5%), 2: 65%
(83%), 3: 19% (12%). Confidence before (after) discussion pretty sure 50% (71%), not
quite sure: 43% (25%), just guessing: 7% (4%).

985
providing wrong answers after the discussion, I would probably spend extra time
discussing the correct answer.
An important point in explaining this question is to avoid (at all cost!) using
equations. My verbal argument goes as follows: force causes acceleration, which
tells how much an object's speed increases in a given interval of time. So if the
force is half as large, the acceleration will be half as large. The force thus needs to
act for a time interval twice as long to give the cart the same increase in speed.
The next ConcepTest(Figure 4) further elaborates on the previous question.
Notice how much better he students do this time before the convince-your-
neighbors discussion. With 90% providing the fight answer before any discussion,
there is little room for improvement. Still, the discussion does increase the
students' confidence. The percentage of correct answers after discussion is a clear
indication that not much further discussion of this question is required.

Figure 4. ConcepTeston force.

A constant force is exerted for a short time interval on a cart that is initially at rest on an
air track. This force gives the cart a certain final speed.

force

The same force is exerted for the same length of time on another cart, also initially at
rest, that has twice the mass of the first one. The final speed of the heavier cart is

1. one-fourth
2. four times
3. half
4. double
5. the same as

that of the lighter cart.

Choice 3 is correct. Response statistics before (after) discussion: 1: 10% (1%), 3: 90%
(99%). Confidence: pretty sure: 64% (95%), not quite sure: 34% (4%), just guessing: 2%
(1%).

I immediately follow this ConcepTestwith the one shown in Figure 5. To save

time, I do not ask the students to discuss their answers.
With all these questions yielding more than 80% correct responses, I move on
to Newton's third law, emphasizing that the two components of a third law force
pair never act on the same object. To make this point clear, I discuss the example
of a person standing in an elevator. While the normal force exerted by the elevator
floor in the person is equal to and opposite the weight of the person when the
elevator is at rest, the two are not an action-reaction pair.

986
Figure 5. ConcepTeston force.
A constant force is exerted for a short time interval on a cart that is initially at rest on an
air track. This force gives the cart a certain final speed.

force

Suppose we repeat the experiment, but instead of starting from rest, the cart is already
moving in the direction of the force at the moment we begin to apply the force. After we
exert the same constant force for the same short time interval, the increase in the cart's
speed

1. is equal to two times its initial speed.

2. is equal to the square of its initial speed.
3. is equal to four times its initial speed.
4. is the same as when it started from rest.
5. cannot be determined from the information provided.
Choice 4 is correct. Response statistics: 1: 10%, 2: 3%, 3: 5%, 4: 82%. Confidence:
pretty sure: 63%, not quite sure: 35%, just guessing: 2%.

When the elevator is accelerating, these two forces are no longer equal--the
difference being responsible for accelerating the person. I make free-body
diagrams for the person and the elevator and indicate which force pairs are third
law pairs. This presentation is followed by a lecture demonstration, immediately
after which I confront the students with the classic question in Figure 6. In spite of

Figure 6. ConcepTest on Newton's third law.

I A locomotive pulls a series of wagons. Which is the correct analysis of the situation?

1. The train moves forward because the locomotive pulls forward slightly harder on the
wagons than the wagons pu I backward on the locomotive.

2. Because action always equals reaction, the locomotive cannot pull the wagons--the
wagons pull backward just as hard as the locomotive pulls forward, so there is no motion. ;

3. The locomotive gets the wagons to move by giving them a tug during which the force
on the wagons is momentarily greater than the force exerted by the wagons on the
locomotive.

4. The locomotive's force on the wagons is as strong as the force of the wagons on the
i locomotive, but the frictional force on the locomotive is forward and large while the
backward frictional force on the wagons is small.

'5. The locomotive can pull the wagons forward only if it weighs more than the wagons.

Choice 4 is correct. Response statistics before (after) discussion: 1: 14% (7%), 2: 2%

(2%), 4: 74% (86%),5: 9% (5%). Confidence before (after) discussion: pretty sure: 59%
(71%), not quite sure: 36% (26%), just guessing: 5% (3%).

987
the conceptual difficulty of this question, a surprisingly large fraction of the class
answer correctly the first time around. This question always raises a large number
of questions--it really gets students thinking--and I usually end up spending time
alter class explaining it a few more times.
Next, returning to the basic purpose of classical mechanics, I show the dual
utility of Newton's second law: given the forces on an object, one can use this law
to determine the motion of that object. As examples, I cite the existence of the
normal force, the forces on celestial bodies, and so forth.
I finally move to the first force law--that of gravitation. I spend some time
making clear the distinction between inertia (an object's tendency to maintain its
state of motion) and gravitation (an object's tendency to attract matter): an
astronaut on the Moon can easily lilt a massive object, but kicking it would hurt as
much as it does on Earth.
The last question I use (Figure 7) involves gravitation, but really tests the
students' understanding of acceleration. This question offers the opportunity to
spiral back and make the connection between the material in previous lectures
(kinematics) and that in this lecture. While two thirds of the students provide the
right answer, only one third are confident of their answer (the most frequent
mistake is to assume that if speed increases, acceleration must increase too).

Figure 7. ConcepTest on gravitation, acceleration, and speed along an incline.

A cart on a roller- coaster rolls down the track shown below. As the cart rolls beyond the
)oint shown, what happens to its speed and acceleration in the direction of motion?

1. Both decrease.
2. The sped decreases, but the acceleration increases.
3. Both remain constant.
4. The speed increases, but acceleration decreases.
5. Both increase.
6. Other.
Choice 4 is correct. Response statistics: 1: 3%, 2: 4%, 3: 8%, 4: 70%, 5: 11%, 6: 4%.
Confidence: pretty sure: 34%, not quite sure: 57%, just guessing: 9%.

988
Adv Physiol Educ 29: 107–111, 2005;
doi:10.1152/advan.00060.2004. How We Learn

Peer instruction enhanced meaningful learning: ability to solve novel problems

Ronald N. Cortright,1 Heidi L. Collins,2 and Stephen E. DiCarlo2
1
Department of Exercise and Sport Science, East Carolina University, Greenville, North Carolina;
and 2Department of Physiology, Wayne State University School of Medicine, Detroit, Michigan
Submitted 2 November 2004; accepted in final form 17 March 2005

Cortright, Ronald N., Heidi L. Collins, and Stephen E. little or no effort to relate new information to existing knowl-
DiCarlo. Peer instruction enhanced meaningful learning: ability to edge or novel situations. In contrast, meaningful learning (5,
solve novel problems. Adv Physiol Educ 29: 107–111, 2005; 13, 14) occurs when the learner interprets, relates, and incor-
doi:10.1152/advan.00060.2004.—Students must be able to interpret, porates new information with existing knowledge and applies
relate, and incorporate new information with existing knowledge and
the new information to solve novel problems. Meaningful
apply the new information to solve novel problems. Peer instruction is
a cooperative learning technique that promotes critical thinking,
learning requires multiple opportunities for the student to be
problem solving, and decision-making skills. Therefore, we tested the actively engaged in the reasoning and application of concepts
hypothesis that peer instruction enhances meaningful learning or (5, 13, 14).
transfer, defined as the student’s ability to solve novel problems or the Lymna’s Think-Pair-Share (11) and Mazur’s Peer Instruc-
ability to extend what has been learned in one context to new contexts. tion (12) provide opportunities for students to be actively
To test this hypothesis, our undergraduate exercise physiology class of engaged in the reasoning and application of concepts (3).
38 students was randomly divided into two groups: group A (n ⫽ 19) Think-Pair-Share occurs two to three times during a lecture
and group B (n ⫽ 19). A randomized crossover design in which when the instructor asks a question or poses a problem.
students either answered questions individually or during peer instruc- Students spend a minute or two alone thinking about an answer
tion was used to control for time and order effects. The first factor that or solution (Think). Subsequently students pair up (Pair) to
influences meaningful learning is the degree of mastery of the original discuss their answers with each other (Share) (3). Mazur used
material. Importantly, peer instruction significantly enhanced mastery
of the original material. Furthermore, the student’s ability to solve
a very similar approach (12). Two to three times during a
novel problems was significantly enhanced following peer instruction. lecture the students solve a physics problem, mark down their
Thus pausing two to three times during a 50-min class to allow peer answer, and rate how confident they feel about the correctness
instruction enhanced the mastery of the original material and en- of their answer. For the pair phase, Mazur allows students 1
hanced meaningful learning, i.e., the student’s ability to solve novel min to convince their neighbor of their answer. After discuss-
problems. ing the problem with classmates, students may revise the
active learning; cooperative learning; problem solving skills; knowl-
answer and again rate their confidence in their second answer.
edge transfer; student attitudes There was a dramatic increase in the confidence level and
percentage of correct answers after students discussed the
concepts (12).
EMPLOYMENT OPPORTUNITIES IN the future will require greater We recently reported that peer instruction increased medical
ability to work together to solve novel problems, because much student performance on quizzes (20). Similarly, we recently
of the knowledge that will be employed in the students’ future reported that collaborative testing, a similar peer instruction
careers is not known today and therefore must be learned after procedure, increased medical student performance on quizzes
graduation (24). Furthermore, not all that is known can be (19) and undergraduate student performance on exams as well
taught in 4 yr and not all that is taught can be learned or as increased student retention of previously learned informa-
remembered. Some of what is taught is erroneous and other tion (6). Specifically, performances on quizzes and retention of
material will soon be obsolete (24). Students must be capable previously learned information were significantly higher when
of working together, gathering evidence, learning from it, and students completed exams in groups than when they completed
applying the information to novel situations. Without the exams individually. These results document that peer instruc-
proper training of the work force for the future, effects on tion enhances exam performance as well as student retention of
professional research and development, economy, society, and previously learned information. Thus peer instruction signifi-
our standard of living will be detrimental. Therefore, it is in the cantly enhanced mastery of the original material. This is
best interest of the nation to raise the level of education of all important because the first factor that influences meaningful
its citizens in an effort to meet the demands of a challenging learning is the degree of mastery of the original material.
society and remain competitive in the scientific and world However, it is unknown if peer instruction enhances meaning-
arenas. However, much of what we do in classes with large ful learning. Therefore, we tested the hypothesis that peer
numbers of students conflicts with these goals. These activities instruction enhances mastery of the original material as well as
do not prepare students for solving novel problems because meaningful learning, i.e., the student’s ability to solve novel
many of these activities encourage memorization of detailed problems.
information. Memorization occurs when the learner makes MATERIALS AND METHODS

Design. All procedures were reviewed and approved by the Intu-

Address for reprint requests and other correspondence: S. E. DiCarlo, Dept. itional Review Board and informed consent was obtained from all
of Physiology, Wayne State Univ. School of Medicine, Detroit, MI 48201 students before beginning the study. We borrowed concepts of the
(e-mail: [email protected]). Lymna and Mazur peer instruction activities to promote student
1043-4046/05 $8.00 Copyright © 2005 The American Physiological Society 107
Downloaded from www.physiology.org/journal/advances (184.149.000.204) on July 9, 2019.
How We Learn
108 LEARNING WITH UNDERSTANDING

involvement in the learning process and test the hypothesis that peer Subsequently students from one group only were allowed 1 min to
instruction enhances mastery of the original material as well as discuss their answers with classmates. Students were then allowed to
meaningful learning (11, 12). Students were divided by the instructor change their first answer if desired, and both answers were recorded.
into two permanent equal groups: group A (n ⫽ 19) and group B (n ⫽ Finally, the instructor and students discussed the answer.
19). Students from the groups were positioned on opposite sides of an To record the student responses, we included five colored sheets of
aisle that physically divided the classroom in half. It was important paper labeled A (red), B (white), C (blue), D (green), and E (yellow)
that the groups be heterogeneous: diverse in gender, ethnic back- in the students’ lecture notes. Students use these colored letters to
ground, and academic ability. Heterogeneous grouping typically en- answer questions during the class (7). For example, students answer
hances the likelihood of success and permits students to work con- the question by holding up the appropriate colored letter. All students
structively with varied individuals who bring different strengths and hold up their choice of colored letter at the same time. These
approaches to tasks. Positive interactions with diverse individuals also procedures increase class participation as well as allow us to imme-
prepare students for the real workplace and for society. When a group diately determine whether students understand a particular concept by
was involved in peer instruction, the students were positioned in observing the sea of colors. If the majority of students answer
clusters of ⬃4 students. We wanted to have 3–4 students per cluster correctly, we feel comfortable moving on. However, if a significant
because this size is large enough to include students of diverse number of students answer incorrectly, the concept is reviewed.
opinions, experiences, and learning styles to assist with the problem Students enjoy this activity and appreciate the opportunity to assess
solving. In addition, if a group member is absent, the cluster can their own understanding as well as participate in class. Furthermore,
continue to function smoothly. A group of four is not so large that a students report that these activities help to hold their attention. Finally,
student can “hide”; thus all students must carry the load. Finally we we get an immediate idea of student learning and record the responses.
used permanent groups so that the students remained together long After determining the effect of peer instruction on the mastery of
enough to establish positive working relationships and to develop original material, we determined the effect of peer instruction on
team building. Classes consisted of two to three short presentations on meaningful learning. A randomized, crossover design was again
key points. Each short presentation was followed by a one-question applied to evaluate performance on problem-solving questions. One
multiple choice quiz. The questions on the quizzes ranged from simple novel problem-solving question was introduced to each class. Stu-
recall to comprehensive questions. The students from both groups dents were allowed ⬃10 min to answer each question. Students
were allowed 1 min to formulate the answer. Subsequently, the alternated from peer instruction groups (group A) to nonpeer instruc-
students from group A only were allowed 1 min to discuss their tion groups (group B) every class period. Each student participated in
answers with their peers. This process encouraged critical thinking, an equal number of questions as group A or group B. Content area was
problem solving, and decision making skill as well as provided a way diverse including metabolism, cardiovascular, and respiratory physi-
to assess the level of understanding. Students from the other group, ology. All students experienced equal numbers of problem-solving
group B, were not allowed to consult with peers. After the first exam, questions from all content areas across the study period. Each prob-
students from group B followed the peer instruction procedures when lem-solving question was worth two points and assessment of mean-
responding to questions, and students from group A were not allowed ingful learning was based on the number of points gained for each
to consult with peers. We analyzed these responses to determine the individual question when operating as group A or group B participants.
effectiveness of peer instruction on the student performance on quiz- Analysis. All results are presented as means ⫾ SE. To determine
zes (mastery of the original material). the effect of peer instruction on student performance on multiple
For the last third of the course, virtually identical procedures were choice questions (mastery of original material), we used a Student’s
followed to determine the effect of peer instruction on meaningful paired t-test to compare raw scores obtained when all students an-
learning; however, all of the questions involved novel situations. The swered the questions as individuals with raw scores obtained when the
students were required to interpret, relate, and incorporate new infor- students answered the questions in collaboration with others (peer
mation with existing knowledge and apply the new information to instruction).
solve novel problems. Students were not exposed to the novel situa- To determine the effect of peer instruction on meaningful learning
tion; however, they received the material required to solve the novel (the ability to solve novel problems) we used a Student’s paired t-test
problem. We analyzed these responses to determine the effectiveness to compare raw scores on novel problem-solving questions obtained
of peer instruction on meaningful learning (the ability to solve novel when all students answered the questions as individuals compared
problems). All problem-solving questions were prepared by an in- with raw scores obtained when the students answered the novel
structor (S. E. DiCarlo) at a different institution who was not formally problem-solving questions in collaboration with others (peer instruc-
active in the class. The questions were developed based on the course tion). Statistical significance was established a priori as P ⬍ 0.05.
outline provided by the class instructor (R. N. Cortright). For exam- A questionnaire (Table 1) was used to evaluate the peer instruction
ples of the problem-solving questions, readers are referred to two procedures. The questionnaire evaluated the goals and objectives,
recent publications (4, 16). Figures and questions related to the specific procedures, students’ attitudes, and personal preferences as
cardiovascular and respiratory sections were taken from these papers. well as summary of recommendations. The students completed the
Procedures. This peer instruction, active learning technique was evaluation at the end of the course. Results from the questionnaire
implemented during the Exercise Physiology class (EXSS 3805) at were analyzed using descriptive statistics and are expressed as
East Carolina University (Greenville, NC). The class consisted of 38 means ⫾ SE.
students. The course, offered through the Department of Exercise and
Sport Science (EXSS) must be completed in the third or fourth year in
RESULTS
order for EXSS majors to meet the graduation requirements in: 1)
Physical Activity and Fitness (BS), 2) Physical Education (BS), 3) Figure 1 presents the effect of peer instruction on student
Exercise and Sport Science (BA), or 4) Exercise Physiology (BA). performance on multiple choice questions (mastery of original
Students from other basic science departments could also enroll. The
material). The daily average number of students who partici-
class was lecture based, with laboratories scheduled throughout the
semester. The peer instruction technique was used for all classes. Each pated in peer instruction (answered questions in clusters) was
class of 50 min was divided into two to three short presentations of 15 16.3 ⫾ 0.2. The daily average number of students who an-
to 20 min each. Each presentation was followed by a one-question swered questions individually was 17.1 ⫾ 0.4. The students
multiple choice quiz on the subject discussed. All students were who interacted with their peers and those who did not answered
allowed 1 min to think and to record their answers (see below). 2–3 questions per class throughout the first two-thirds of the

Advances in Physiology Education • VOL 29 • JUNE 2005

Downloaded from www.physiology.org/journal/advances (184.149.000.204) on July 9, 2019.
 How We Learn
LEARNING WITH UNDERSTANDING 109

Table 1. Student evaluation of peer instruction

Questions Results

1. The purpose of and rationale behind the educational process was fully explained. 4.6 ⫾ 0.09
2. The process was not too lengthy or complex in its format. 4.6 ⫾ 0.11
3. An opportunity to assess an individual’s understanding through questions and answers was provided. 4.6 ⫾ 0.09
4. Peer instruction increased my confidence. 4.4 ⫾ 0.11
5. Peer instruction allowed me to go beyond my previous level of knowledge. 4.3 ⫾ 0.12
6. Peer instruction facilitated my learning of the material. 4.3 ⫾ 0.11
7. Every member “pulled their weight” (contributed to the learning process). 4.1 ⫾ 0.14
8. The level of discussion during peer instruction was high. 4.4 ⫾ 0.12
9. I appreciated the immediate feedback afforded by peer instruction. 4.6 ⫾ 0.10
10. It was difficult to convince students of correct answers. 3.1 ⫾ 0.20
11. Peer instruction enhanced my understanding and ability to synthesize and integrate material. 4.4 ⫾ 0.10
12. Peer instruction provided a more positive relationship among students. 4.6 ⫾ 0.10
13. Peer instruction provided a more positive relationship between students and faculty. 4.6 ⫾ 0.09
14. Peer instruction provided a more constructive classroom learning environment. 4.6 ⫾ 0.09
15. Peer instruction provided the opportunity to discuss incorrect answers and fill in knowledge gaps
and therefore improve understanding of the material. 4.6 ⫾ 0.09
16. My level of involvement during the discussions was high. 4.2 ⫾ 0.13
17. This method of learning was as effective as any other I have encountered. 4.4 ⫾ 0.10
18. This process was educationally attractive due to the novelty of this style and format. 4.5 ⫾ 0.10
19. I would recommend this process for other content areas. 4.6 ⫾ 0.09
20. I enjoyed peer instruction 4.7 ⫾ 0.08

Results are expressed as means ⫾ SE. Students were given the following instructions to respond to the questions. Circle the number that most accurately defines
the way you feel regarding each statement: 1, strongly disagree; 2, tend to disagree; 3, neither agree nor disagree; 4, tend to agree; 5, strongly agree.

course for a total of 34 questions. Without peer instruction, the ing process) (question 7, 4.1 ⫾ 0.14). We observed all students
students answered questions correctly 44 ⫾ 5% of the time. In “pulling their weight” as well as students monitoring group
contrast, when students were allowed to collaborate with fel- behaviors. Students were observed listening, providing con-
low classmates, they answered the questions correctly 59 ⫾ structive feedback, and reflecting on their learning. We ob-
6% of the time. This 27% improvement of the raw scores was served students developing leadership, decision making, com-
statistically significant (P ⫽ 0.02). munication, and conflict resolution skills while gaining mutual
Figure 2 presents the effect of peer instruction on meaning- respect of peers. The peer instruction activities fostered posi-
ful learning (the ability to solve novel problems). The students tive interdependence (question 12, 4.6 ⫾ 0.10) and individual
who interacted with their peers and those who did not answered accountability (question 7, 4.1 ⫾ 0.14), because the activity
one problem-solving question per class throughout the last was carefully structured. Students reported that the level of
one-third of the course for a total of six questions. When discussion during peer instruction was high (question 8, 4.4 ⫾
students were allowed to interact with their peers, the level of 0.12). Indeed, students were observed focused and seriously
performance on novel problem-solving questions was signifi- engaged in the discussion. It was clear that the students learned
cantly greater (47 ⫾ 5 vs. 24 ⫾ 2%, P ⫽ 0.04) than when they more and better by becoming actively involved with the ma-
completed novel tasks in the traditional format. terial. However, activity in and of itself does not result in
The questionnaire used to evaluate the peer instruction higher learning. The students invested in the discussion and we
procedures and the students’ responses are presented in Table observed a high level of involvement from the students (ques-
1. Thirty-four of the 38 students completed the questionnaire; tion 16, 4.2 ⫾ 0.13), which helped the students make what they
this represents an 89% response rate. Among 20 responses, the were learning meaningful. The students reported that peer
students reported that the purpose of and rationale behind the instruction provided a more positive relationship between stu-
educational process was fully explained (question 1, 4.6 ⫾ dents and faculty (question 13, 4.6 ⫾ 0.09) and among students
0.09). It was clear that the students understood the educational (question 12, 4.6 ⫾ 0.10). Interaction between teacher and
goals and figured out ways to achieve the goals. By under- students and students and students is one of the most powerful
standing the goals, the students were observed becoming more factors in promoting learning. The students reported that they
efficient and effective with the discussions. In addition, it was enjoyed peer instruction (question 20, 4.7 ⫾ 0.08). It was clear
clear that the students understood the nature and value of the that the students were motivated, eager to learn, and having
activity, which prevented many of the concerns faculty may fun. This created a wonderful classroom environment where
have about peer activities. Thus when establishing cooperative student’s confidence (question 4, 4.4 ⫾ 0.11), learning (ques-
peer activities, it is important that clear instructions are pre- tions 3, 4.6 ⫾ 0.09 and 5, 4.3 ⫾ 0.12) and involvement
sented. Students reported that peer instruction facilitated their (question 16, 4.2 ⫾ 0.13) were high.
learning of the material (question 6, 4.3 ⫾ 0.11). It was clear
that there were many concepts that could not be learned DISCUSSION
quickly. The time allowed during peer instruction provided the
opportunity for students to master several complex concepts. In this study, we examined the effectiveness of peer instruc-
As the authority, the instructor often summarized the lesson tion, a pedagogical method that promotes student participation
and validated the learning that occurred. Students reported that in class and increases student interaction with each other and
every member “pulled their weight” (contributed to the learn- with the instructor, on student performance on quizzes (mas-

Advances in Physiology Education • VOL 29 • JUNE 2005

Downloaded from www.physiology.org/journal/advances (184.149.000.204) on July 9, 2019.
How We Learn
110 LEARNING WITH UNDERSTANDING

topic. Thus learning cannot be rushed; the complex cognitive

activity of information integration requires time (5). It is
important to recognize that time alone will not result in the
benefit seen with peer instruction. Rather it is the practice and
the feedback provided by peer instruction that enhances the
mastery and transfer of knowledge. As stated more succinctly
by a quote attributed to world-renowned physiologist and
medical historian Horace Davenport, “There is a great differ-
ence between teaching and learning. There is too much teach-
ing and not enough learning.” (9, 23). In this context, I taught
George, my cat, to fetch the remote to my sound system;
however, he has failed to learn this task. Importantly, peer
instruction provides the time for students to test existing
knowledge and apply it to novel situations in a safe, supportive,
environment. This quality learning experience allows students
to evaluate their concepts and experiences while providing
Fig. 1. Effect of peer instruction on student performance on multiple choice feedback about their progress.
questions (mastery of original material). The daily average number of students A questionnaire (Table 1) was used to evaluate the peer
who participated in peer instruction (answered questions in clusters) was
16.3 ⫾ 0.2 (means ⫾ SE). The daily average number of students who answered
instruction procedures. The questionnaire evaluated the goals
questions individually was 17.1 ⫾ 0.4. Without peer instruction (Individual), and objectives, specific procedures, students’ attitudes, and
the students answered questions correctly 44 ⫾ 5% of the time. In contrast, personal preferences as well as summary and recommenda-
when students were allowed to collaborate with fellow classmates (Peer tions. The students completed the evaluation at the end of the
Instruction), they answered the questions correctly 59 ⫾ 6% of the time. This course. The questionnaire documented that the students devel-
27% improvement in raw scores was statistically significantly (*P ⫽ 0.02).
Values are means ⫾ SE. oped a better understanding of the material and in the process
gained more self confidence (questions 4 and 6). Furthermore,
the questionnaire documented that peer instruction resulted in
tery of original material). In addition, we examined the effec- more positive relationships among students (question 12) and
tiveness of peer instruction on meaningful learning, defined as between students and faculty (question 13), more positive
the learner interprets, relates, and incorporates new information psychological well being (questions 4-6), and a more construc-
with existing knowledge and applies the new information to tive classroom-learning environment (question 14).
solve novel problems. Results from this study confirm previous The results of the questionnaire are important because the
reports documenting that peer instruction and collaborative students also reported that they enjoyed the procedures (ques-
group test taking enhanced student performance on quizzes. tion 20) and would recommend this process for other content
Specifically, previous studies (6, 15, 19, 20, 21) have docu- areas (question 19). It is well documented that motivation
mented that student performance on examinations is signifi- affects the quality of the learning experience and that high
cantly higher when students completed the same exam in levels of motivation increase learning with understanding (8).
groups than when they completed the examinations individu- According to the ancient Greek scholar Plato, “Bodily exercise,
ally. The new finding from this study is that peer instruction
enhanced meaningful learning. That is, peer instruction pro-
vided a learning experience that lead to transfer, defined as the
ability to extend what has been learned in one context to new
contexts (2, 22). Quality learning experiences should lead to
transfer. All learning experiences can appear equivalent when
measures of learning are focused on the ability to repeat
previously taught facts. However, quality learning experiences
(learning with understanding) can be identified when tests of
transfer are used. Thus because peer instruction leads to trans-
fer, this study documents that peer instruction provides a
quality learning experience.
Mastery of the subject is essential for meaningful learning
(e.g., for the successful transfer of knowledge to solve novel
problems). Without an adequate level of initial learning, trans-
fer cannot be expected (5). Thus transfer is affected by the
degree to which students learn with understanding rather than
merely memorize sets of facts (5). Learning with understand-
ing requires time to allow for practice. Faculty must be realistic
about the amount of time required to learn complex concepts
and provide the practice time to achieve the goal. Students Fig. 2. Effect of peer instruction on meaningful learning (the ability to solve
novel problems). When students were allowed to interact with their peers (Peer
need to take time to explore underlying concepts and to Instruction), the level of performance on novel problem-solving questions was
generate connections to other information. Students must have significantly greater (47 ⫾ 5 vs. 24 ⫾ 2%, *P ⫽ 0.04) than when they completed
time to “grapple” with specific information relevant to the novel tasks in the traditional format (Individual). Values are means ⫾ SE.

Advances in Physiology Education • VOL 29 • JUNE 2005

Downloaded from www.physiology.org/journal/advances (184.149.000.204) on July 9, 2019.
 How We Learn
LEARNING WITH UNDERSTANDING 111

when compulsory, does no harm to the body; but knowledge DC: The George Washington University, School of Education and Human
which is acquired under compulsion obtains no hold on the Development, 1991.
2. Byrnes JP. Cognitive Development and Learning in Instructional Con-
mind.” (17). Peer instruction motivates students! Furthermore, texts. Boston, MA: Allyn and Bacon, 1996.
learning with understanding and transfer requires the student to 3. Carbone E. Teaching in Large Classes. Thousand Oaks, CA: SAGE
actively choose and evaluate strategies, consider resources, and Publications, 1998.
receive feedback. 4. Collins HL, Rodenbaugh DW, Murphy TP, Kulics JM, Bailey CM,
By actively involving students in peer instruction activities, and DiCarlo SE. An inquiry-based teaching tool for understanding
the student’s attention span may be greatly increased. It has arterial blood pressure regulation and cardiovascular function. Adv Physiol
Educ 277: 15–28, 1999.
been reported that students in a lecture-based college class- 5. Committee on Developments in the Science of Learning, Committee
room are not attentive ⬃40% of the time (18). During sus- on Learning Research, and Educational Practice, and the National
tained lectures, student attention decreases with each passing Research Council. How People Learn: Brain, Mind, Experience, and
minute. Furthermore, sustained lectures appeal only to auditory School. Washington, DC: National Academy Press, 2000.
learners and tend to promote lower level learning of factual 6. Cortright RN, Collins HL, Rodenbaugh DW, and DiCarlo SE. Student
retention of course content is improved by collaborative-group testing.
information. Finally, sustained lecturing assumes that all stu- Adv Physiol Educ 27: 102–108, 2003.
dents learn the same information at the same pace (10). Thus 7. DiCarlo SE and Collins HL. Colored letters: a tool to increase class
by incorporating peer instruction, active learning activities may participation in a large classroom. Adv Physiol Educ 25: 71, 2001.
increase student’s attention. 8. Dweck CS. Motivation. Hillsdale, NJ: Erlbaun, 1989.
Faculty are often reluctant to incorporate active learning 9. Haramati A. “Teaching physiology: filling a bucket or lighting a fire?”
activities in the class. The reasons most often advanced for not Physiologist 43: 117–121, 2000.
10. Johnson DW, Johnson RT, and Smith KA. Active learning: cooperation
including these active learning activities include not being able in the college classroom. Edina, MN: Interaction Book. 1991.
to cover as much content in the time available and the exces- 11. Lymna F. The responsive classroom discussion. In: Mainstreaming Di-
sive preparation time required for devising strategies promot- gest, edited by Anderson AS. College Park, MD: University of Maryland
ing active learning (1). In terms of preparation time, minimal College of Education, 1981.
extra time is required for this active approach. Furthermore, the 12. Mazur E. Peer instruction: getting students to think in class. In: The
minimal extra time pays dividends in understanding and reten- Changing Role of Physics Departments in Modern Universities, edited by
Redish EF and Rigden JS. The American Institute of Physics: Proceedings
tion of material (6). As stated by Mazur (12), using time for of ICUPE 399: 981–988, 1997.
peer instruction greatly improves student’s level of understand- 13. Michael J. In pursuit of meaningful learning. Adv Physiol Educ 25:
ing with relatively little effort and no capital investment. 145–158, 2001.
Although little additional time is required in preparation for 14. Mintzes JJ and Wandersee JH. Reform and Innovation in Science
the peer instruction activities, the instructor has several impor- Teaching: a Human Constructivist View. Teaching Science for Under-
standing. San Diego, CA: Academic, 1997, p. 29–58.
tant roles during the process. For example, the instructor must 15. Muir SP and Tracy DM. Collaborative essay testing: just try it. Coll
model appropriate social skills, including listening and provid- Teach 47: 33–35, 1999.
ing constructive feedback or eliciting more in-depth responses 16. Patil RD, Karve SV, and DiCarlo SE. Integrated cardiovascular physi-
through probing questions. The instructor must also reinforce ology: a laboratory exercise. Adv Physiol Educ 265: 20–31, 1993.
these positive behaviors by publicly commenting on the ways 17. Plato. The Republic. Greek Author and Philosopher in Athens. (427 BC–347
students use them effectively. BC), Book VII [online]. The Literature Page. http://www.literaturepage.com/
read/therepublic-284.html. [2005].
In conclusion, pausing two to three times during a 50-min 18. Pollio HR. What Students Think About and Do in College Lecture Classes.
class to allow peer instruction of concepts enhanced the stu- Teaching-Learning Issues. Knoxville, TN: University of Tennessee Press,
dent’s level of understanding and ability to synthesize and Learning Research Center. 1984.
integrate material. Specifically, peer instruction enhanced the 19. Rao SP, Collins HL, and DiCarlo SE. Collaborative testing enhances
mastery of original material and meaningful learning. It is student learning. Adv Physiol Educ 26: 37–41, 2002.
20. Rao SP and DiCarlo SE. Peer instruction improves performance on
possible to create an effective active learning environment with
quizzes. Adv Physiol Educ 24: 51–55, 2000.
relatively little effort by implementing peer instruction tech- 21. Russo A and Warren SH. Collaborative test taking. Coll Teach 47:
nique. We would be wise to heed the words of British math- 18–20, 1999.
ematician, logician, and philosopher Alfred North Whitehead 22. Thorndike EL and Woodworth RS. The influence of improvement in
who said, “So far as the mere imparting of information is con- one mental function upon the efficiency of other functions. Psychol Rev 8:
cerned, no university has had any justification for existence since 247–261, 1901.
23. Vander AJ. The Claude Bernard Distinguished Lecture. The excitement
the popularization of printing in the fifteenth century.” (25). and challenge of teaching physiology: shaping ourselves and the future.
Adv Physiol Educ 267: 3–16, 1994.
REFERENCES
24. West KM. The case against teaching. J Med Educ 41: 766–771, 1966.
1. Bonwell CC and Eison JA. Active Learning: Creating Excitement in the 25. Whitehead A. The Aims of Education and Other Essays. New York:
Classroom. ASHE-ERIC Higher Education Report No. 1. Washington Macmillan, 1929.

Advances in Physiology Education • VOL 29 • JUNE 2005

Downloaded from www.physiology.org/journal/advances (184.149.000.204) on July 9, 2019.
 Adv Physiol Educ 30: 168 –173, 2006;
How We Teach doi:10.1152/advan.00013.2006.

Peer instruction enhanced student performance on qualitative

problem-solving questions
Mauricio J. Giuliodori,1 Heidi L. Lujan,2 and Stephen E. DiCarlo2
1
Cátedra de Fisiologı́a, Facultad de Ciencias Veterinarias, Universidad Nacional de La Plata, La Plata, Argentina;
and 2Department of Physiology, Wayne State University School of Medicine, Detroit, Michigan
Submitted 13 March 2006; accepted in final form 12 July 2006

Giuliodori, Mauricio J., Heidi L. Lujan, and Stephen E. Di- sponse of a physiological system to a perturbation. Qualitative
Carlo. Peer instruction enhanced student performance on qualitative problems require integration of multiple concepts; however,
problem-solving questions. Adv Physiol Educ 30: 168 –173, 2006; the problems can be answered quickly with single, best mul-
doi:10.1152/advan.00013.2006.—We tested the hypothesis that peer tiple-choice questions. These are important considerations be-
instruction enhances student performance on qualitative problem-
cause one of the most important factors influencing learning is
solving questions. To test this hypothesis, qualitative problems were
included in a peer instruction format during our Physiology course. what the student already knows. The students must link new
Each class of 90 min was divided into four to six short segments of 15 information to concepts they already possess (8). This process
to 20 min each. Each short segment was followed by a qualitative is critical for solving novel problems. Peer instruction is a
problem-solving scenario that could be answered with a multiple- cooperative learning technique that may promote this process.
choice quiz. All students were allowed 1 min to think and to record Therefore, we tested the hypothesis that peer instruction en-
their answers. Subsequently, students were allowed 1 min to discuss hances the students’ performance on qualitative problem-solv-
their answers with classmates. Students were then allowed to change ing questions. To test this hypothesis, qualitative problems
their first answer if desired, and both answers were recorded. Finally, were included in a peer instruction format during our Veteri-
the instructor and students discussed the answer. Peer instruction nary Physiology course.
significantly improved student performance on qualitative problem-
solving questions (59.3 ⫾ 0.5% vs. 80.3 ⫾ 0.4%). Furthermore, after
METHODS
peer instruction, only 6.5% of the students changed their correct
response to an incorrect response; however, 56.8% of students Design. We borrowed concepts from Lymna (11) and
changed their incorrect response to a correct response. Therefore,
students with incorrect responses changed their answers more often
Mazur (13) peer instruction activities to promote student in-
than students with correct responses. In conclusion, pausing four to volvement in the learning process and test the hypothesis that
six times during a 90-min class to allow peer instruction enhanced peer instruction enhances student performance on qualitative
student performance on qualitative problem-solving questions. problem-solving questions.
Procedures. This peer instruction, active-learning tech-
collaboration; meaningful learning; transfer nique was implemented during the Physiology class (Fisiologı́a
No. 423) at the Facultad de Ciencias Veterinarias, Universidad
THERE HAS BEEN REMARKABLE PROGRESS in our understanding on Nacional de La Plata, La Plata, Argentina. The class consisted
how people learn. It is now clear that concept construction of 114 veterinary medical students. The class was lecture
requires active processing of information. That is, we under- based, and the peer instruction technique was used for 10
stand and remember the information we think about (8)! classes involving cardiovascular, respiratory, and renal physi-
However, processing information requires time. Faculty mem- ology. Each class of 90 min was divided into four to six short
bers must be realistic about the amount of time required to presentations of 15–20 min each. Each short presentation was
learn complex concepts and provide the time needed to achieve followed by a qualitative problem-solving scenario that could
the goal (5). Students need time to explore the underlying be answered with a one-question, multiple-choice quiz. All
concepts and to generate connections to other information. students were allowed 1 min to think and to record their
Students must have time to “grapple” with specific information answers. Subsequently, students were allowed 1 min to discuss
relevant to the topic. Thus, learning cannot be rushed; the their answers with classmates (2–3 students/group). Students
complex cognitive activity of information integration requires were then allowed to change their first answer if desired, and
time (4). both answers were recorded. Students were instructed to pro-
However, students attending our Physiology course are in vide reasons for their answers and to convince their peers that
class 5 h/wk. This computes to ⬃2.4% of the total hours in a their answers were correct. In this format, the students had two
student’s week. Herein lies the problem: How do we provide roles: as a teacher, explaining the rationale for their answer;
time for information processing during this limited class time and as a student, listening to the reasoning for their peers’
(8)? One way is to use qualitative problems (14, 15) in a peer answer. Finally, the instructor and students discussed the
instruction format (12, 13). These problems require a qualita- answer.
tive prediction (increase/decrease/no change) about the re- The questions were qualitative problem-solving scenarios
generated by M. J. Giuliodori using the format provided by
Michael and co-workers (see Appendixes A–C) (14, 15). The
Address for reprint requests and other correspondence: S. E. DiCarlo, qualitative problem-solving scenarios asked for a qualitative
Wayne State Univ. School of Medicine, 540 E. Canfield Ave., Detroit, MI prediction (increase/decrease/no change) about the response of
48201 (e-mail: [email protected]). a system to a perturbation; for example, If the heart is dener-
168 1043-4046/06 $8.00 Copyright © 2006 The American Physiological Society
Downloaded from www.physiology.org/journal/advances (184.149.000.204) on July 9, 2019.
 How We Teach
COLLABORATIVE PROBLEM SOLVING 169

vated, what change, if any, will occur to heart rate (will it A

increase, decrease, or stay the same)? (14). Specifically, the
questions posed conceptual problem-solving scenarios that
required the integration of multiple concepts but were an-
swered with single, best multiple-choice questions. In addition,
tables were used, instead of multiple-choice questions, when
more than one prediction was required (see Appendixes A–C);
for example, Predict how cutaneous blood flow, shivering, and
sweating will be affected at the onset of a fever (15).
Statistical analysis. All results are presented as means ⫾
SE, and significance was set at the P ⬍ 0.05 level. To
determine the effect of peer instruction on student performance
on qualitative problem-solving questions (see Fig. 1), we used
a Student’s paired t-test to compare responses obtained when
the students solved problems as individuals with responses
obtained when the students solved problems in collaboration
with peers (peer instruction). B
To determine which students changed their individual re-
sponse (see Fig. 2), we used a Kruskal-Wallis nonparametric,
one-way ANOVA. Once statistical significance was estab-
lished, post hoc analysis was performed with a Student-New-
man-Keuls test. Finally, to compare the positive effects (indi-
vidual incorrect responses changed to peer-instructed correct
responses) with negative effects (individual correct responses
changed to peer-instructed incorrect responses), we used a
Mann-Whitney rank-sum test (see Fig. 3). Significance was set
at the P ⬍ 0.05 level.
RESULTS

Figure 1 presents the percentage of correct responses when

the students solved problems as individuals and when the
students solved the same problems in collaboration with peers. Fig. 2. A: percentages of correct individual responses that did not change
As individuals, the students responded correctly 59.3 ⫾ 0.5% (correct to same correct) or changed to incorrect responses (correct to incor-
of the time. In sharp contrast, in collaboration with peers, the rect) after peer instruction. In the group of students having correct individual
students solved the same problem correctly 80.3 ⫾ 0.4% of the responses (⬃60% of the total population), 93.5% did not change their response
after peer instruction (no effect). However, 6.5% changed their responses to
time. This 21% increase was statistically significant (P ⬍ incorrect responses (negative effect). B: percentages of incorrect individual
0.001). responses that changed to correct responses (incorrect to correct), changed to
Figure 2A presents the percentage of correct individual a different incorrect response (incorrect to incorrect), or did not change
responses that did not change (correct to same correct) or (incorrect to same incorrect) after peer instruction. In the group of students
having individual incorrect answers (⬃40% of student population), 56.8%
changed their initial incorrect response to a correct response after peer
instruction (positive effect), 35.9% did not change (incorrect to same incorrect,
no effect), and 7.3% changed to another incorrect response (no effect).
One-way nonparametric ANOVA revealed significant group effects (P ⬍
0.001). Post hoc analysis revealed that fewer students change their responses
from correct to incorrect (6.5%) than from incorrect to correct (56.8%, P ⬍
0.05). Thus, most correct students (93.5%) did not change their individual
responses; however, many incorrect students did (64.1%). Taken together,
these data document that it is easier to convince someone who is incorrect than
someone who is correct.

changed to incorrect responses (correct to incorrect) after peer

instruction. In the group of students having correct individual
responses (⬃60% of the total population), 93.5% did not
change their response after peer instruction (no effect). How-
ever, 6.5% changed their responses to incorrect responses
(negative effect).
Fig. 1. Effect of peer instruction on students’ qualitative problem-solving Figure 2B presents the percent of incorrect individual re-
skills. Without peer instruction (individual), the students solved problems sponses that changed to correct responses (incorrect to correct),
correctly 59.3 ⫾ 0.5% of the time (mean ⫾ SE). In contrast, when students
were allowed to collaborate with classmates (peer instruction), the students
changed to a different incorrect response (incorrect to incor-
solved the problems correctly 80.3 ⫾ 0.4% of the time. *P ⬍ 0.001, individual rect), or did not change (incorrect to same incorrect) after peer
vs. peer instruction. instruction. In the group of students having individual incorrect

Advances in Physiology Education • VOL 30 • DECEMBER 2006

Downloaded from www.physiology.org/journal/advances (184.149.000.204) on July 9, 2019.
How We Teach
170 COLLABORATIVE PROBLEM SOLVING

the topics. The students reported that the level of discussion

was high and the immediate feedback was helpful. Finally, the
students enjoyed the methodology and recommended it for
other courses.

DISCUSSION

In this work, we examined the effect of peer instruction, a

pedagogical tool that increases student interactions with each
other and with the instructor, on student performance on
qualitative problem-solving questions. The main finding was
that peer instruction increased student performance on quali-
tative problem-solving questions. Specifically, there was a 35%
improvement in correct responses to qualitative problems after
discussions with peers (absolute gain: 21 percentage units, P ⬍
Fig. 3. Effect of peer instruction on the way students changed their responses 0.001; Fig. 1). Similar results have been reported by other
(positive change vs. negative change). Of the student population, only 4% investigators. For example, Crouch and Mazur (7) observed
showed a negative change (individual correct answers to peer-instructed significant increases in conceptual problem-solving skills in-
incorrect answers). However, 22.4% showed a positive change (individual
incorrect answers to peer-instructed correct answers, P ⬍ 0.001). volving physics scenarios over a 10-yr period of peer instruc-
tion experience. Similarly, we (17) recently reported that peer
instruction increased medical student performance on quizzes.
answers (⬃40% of student population), 56.8% changed their Furthermore, collaborative testing, a similar peer instruction
initial incorrect response to a correct response after peer procedure, also increased medical student performance on
instruction (positive effect), 35.9% did not change (incorrect to quizzes (16, 18). Similarly, peer instruction as well as collab-
same incorrect, no effect), and 7.3% changed to another incor- orative testing increased undergraduate student performance on
rect response (no effect). One-way nonparametric ANOVA exams and increased student retention of previously learned
revealed significant group effects (P ⬍ 0.001). Post hoc anal- information (5, 6). Specifically, performances on quizzes and
ysis revealed that fewer students changed their response from retention of previously learned information were significantly
correct to incorrect (6.5%) than from incorrect to correct higher when students completed exams in groups rather than
(56.8%, P ⬍ 0.05). Thus, most correct students (93.5%) did not when they completed exams individually (5, 6). Importantly,
change their individual response; however, many incorrect we (5) also documented that peer instruction enhanced mean-
students did (64.1%). Taken together, these data document that ingful learning (the students’ ability to solve novel problems).
it is easier to convince someone who is incorrect than someone The new finding from this study is that peer instruction
who is correct. enhanced student performance on qualitative problem-solving
Figure 3 compares the positive effects (individual incorrect questions. Qualitative problems require the integration of mul-
responses changed to peer-instructed correct responses) with tiple concepts. Thus, peer instruction provides a learning ex-
negative effects (individual correct responses changed to peer- perience that leads to the ability to extend what has been
instructed incorrect responses). Only 4% of the students had a learned in one context to new contexts (2, 20). All learning
negative change (individual correct response to peer-instructed experiences can appear equivalent when measures of learning
incorrect responses). In contrast, 22.4% of the students had a are focused on the ability to repeat previously taught facts.
positive change (individual incorrect response to peer-in- However, quality learning experiences require the ability to
structed correct responses). Post hoc analysis revealed that the extend what has been learned in one context to new contexts.
positive effects were higher than the negative effects (22.4% Quality learning experiences require time to allow for practice.
vs. 4.0%, P ⬍ 0.001). Faculty members must be realistic about the amount of time
Students’ perceptions regarding the peer instruction activity required to learn complex concepts and provide the practice
are presented in Table 1. One hundred four students (of 114 time to achieve the goal. Students need time to explore under-
students total) returned the completed questionnaire (91.2% lying concepts and to generate connections to other informa-
response rate). The students reported that the peer instruction tion. Students must have time to “grapple” with specific infor-
methodology was simple and helped them to better understand mation relevant to the topic. Thus, learning cannot be rushed;

Table 1. Student perceptions regarding the PI activities

Questionnaire Items Responses

1. The methodology of PI was clear and easy to follow. 3.80⫾0.09

2. The methodology of PI was interesting and enjoyable. 3.84⫾0.08
3. The methodology of PI helped me to better understand the topics. 3.91⫾0.09
4. The levels of discussion with peers and instructor were high. 3.75⫾0.09
5. The immediate feedback given by discussions with peers and instructor was positive. 3.92⫾0.08
6. Other faculty members should include PI in their courses. 3.72⫾0.10

Data are shown as means ⫾ SE; n ⫽ 104 students who returned the completed questionnaire (91.2% response rate) regarding the peer instruction (PI) activities.
Responses were on a 5-point scale, where 1 was completely disagree, 2 was disagree, 3 was neither agree nor disagree, 4 was agree, and 5 was completely agree.

Advances in Physiology Education • VOL 30 • DECEMBER 2006

Downloaded from www.physiology.org/journal/advances (184.149.000.204) on July 9, 2019.
 How We Teach
COLLABORATIVE PROBLEM SOLVING 171

the complex cognitive activity of information integration re- answer for the right reasons (13). Thus, there is always an
quires time (4). Importantly, peer instruction provides the time increase and never a decrease in the number of correct answers
for students to test existing knowledge and apply it to novel (solutions) after discussion with classmates.
situations in a safe, supportive environment. This quality learn- Faculty members are often reluctant to incorporate active learn-
ing experience allows students to evaluate their concepts and ing activities in class. The reasons most often advanced for not
experiences while providing feedback about their progress. including these active learning activities include not being able to
A previous study (13) has shown that students obtain opti- cover as much content and the excessive preparation time required
mal benefits of peer instruction when the percentage of correct for devising strategies promoting active learning (1). However, as
individual responses is between 35% and 70%. Specifically, stated by Mazur (13), using time for peer instruction greatly
when the percentage of correct individual responses is too low improves the student’s level of understanding with relatively little
(⬍35%) or too high (⬎70%), there is little improvement. For effort and no capital investment. In addition, the instructor has
example, when the percentage of correct individual responses several important roles during the process. For example, the
is too low (⬍35%), most of the students have not obtained instructor must model appropriate social skills, including listening
sufficient understanding of the concept to have meaningful and providing constructive feedback or eliciting more indepth
discussions. In contrast, when the percentage of correct indi- responses through probing questions. The instructor must also
vidual responses is too high (⬎70%), there is less room for reinforce these positive behaviors by publicly commenting on the
improvement (13). In this study, the percentage of correct ways students use them effectively (5).
individual responses (59.3 ⫾ 0.4%) was within the range for Student perceptions regarding the peer instruction activities
optimal improvement (13). are in agreement with previous work (5, 6, 16, 17). Students
The beneficial effects of peer instruction are due, in part, to appreciated the interactions with peers and with the instructor.
two major factors. First, student attention decreases with each This interaction provided immediate feedback, which is not
passing minute during sustained lectures. Importantly, peer possible during the traditional lecture format. The students
instruction activities increase attention by actively involving reported that peer instruction facilitated their learning of the
students in problem-solving activities. Furthermore, sustained topics. In this content, students were seen to be enthusiastically
lectures appeal only to auditory learners and tend to promote engaged in content-based discussions, giving support to their
lower-level learning of factual information (10, 18). Finally, choices. Finally, students enjoyed this learning experience and
sustained lecturing assumes that all students learn the same recommended it for other courses.
information at the same pace (9). Rowe (19) reported that In conclusion, pausing four to six times during a 90-min
pausing every 15 min during a lecture increased students’ class to allow peer instruction of qualitative problems en-
attention and retention. hanced the students’ performance on qualitative problem-solv-
Second, the value of peer instruction derives from the ing questions.
student generating explanations for their answers. Students
obtain benefits when they generate their own explanations
Appendix A: Samples of Assessed Cardiovascular Questions
(“self-explanation”) for their new knowledge (3) and when
they explain their reasoning to classmates, that is, when the 1. Predict (increase/decrease/no change) what would happen
learner acts as a teacher. Thus, “the best way to learn some- to the velocity of blood flow through systemic vessels if you
thing is to teach it,” because teaching requires the generations provide a medication causing smooth muscle contraction:
of explanations, both for oneself and for the learner (14). All of A. Increase (correct)
us who teach have experienced and understand the true mean- B. Decrease
ing of this concept. C. No change
In this study, only 6.5% of the students with individual 2. Predict (increase/decrease/no change) what would happen
correct responses changed their answers after peer discussion to the resting membrane potential in cardiac muscle cells if the
to incorrect responses. In sharp contrast, 64.1% of the students extracellular K⫹ concentration increases:
with individual incorrect responses changed their answers after A. Increase
peer discussion. Thus, most of the students who changed their B. Decrease (correct)
responses changed to a correct answer to correct responses C. No change
(56.8%), whereas a small portion changed to an incorrect 3. Predict (increase/decrease/no change) what would happen
answer (7.3%, P ⬍ 0.05; Fig. 2). Specifically, 22.4% of the to cardiac output if you provide a medication causing smooth
student population changed their responses in a positive way muscle relaxation:
(from incorrect to correct answers), whereas only 4.0% of the A. Increase
student population changed their responses in a negative way B. Decrease (correct)
(from correct to incorrect answers, P ⬍ 0.001; Fig. 3). Taken C. No change
together, the magnitude of the peer instruction positive effect 4. Predict (increase/decrease/no change) what would happen
was 5.6 times higher than the magnitude of the peer instruction to stroke volume, end-diastolic volume, and cardiac output if
negative effect. Therefore, the beneficial effects of peer in- afterload increased:
struction on students’ performance were observed in the group
of students having individual incorrect answers (“weaker stu-
dents”) (Fig. 3). These results are in agreement with reports by Response

Crouch and Mazur (7). These authors reported that it is much Stroke volume Decrease
easier to change the mind of someone who is wrong than it is End-diastolic volume Increase
to change the mind of someone who has selected the correct Cardiac output Decrease

Advances in Physiology Education • VOL 30 • DECEMBER 2006

Downloaded from www.physiology.org/journal/advances (184.149.000.204) on July 9, 2019.
How We Teach
172 COLLABORATIVE PROBLEM SOLVING

5. Predict (increase/decrease/no change) what would ular capillary pressure during efferent arteriolar vasoconstric-
happen to the volume of blood returning to the heart through tion:
the veins if right atrial pressure (central venous pressure)
increased: Response
A. Increase
B. Decrease (correct) Glomerular filtration rate Increase
Renal blood flow Decrease
C. No change Glomerular capillary pressure Increase
6. Predict (increase/decrease/no change) what would happen
to tissue fluid formation if you provide a medication causing 2. Predict (higher/lower/the same) how urine osmolarity
smooth muscle contraction in veins: would be compared with plasma osmolarity if you provide a
A. Increase (correct) medication blocking the 2Cl⫺-Na⫹-K⫹ cotransporter (i.e., fu-
B. Decrease rosemide):
C. No change A. Higher (correct)
B. Lower
Appendix B: Samples of Assessed Respiratory Questions C. The same
3. Predict (higher than 1/lower than 1/equal to 1) the frac-
1. Predict (higher/lower/the same) how the tidal volume of a tional excretion of a drug that is both filtered and secreted (with
horse immediately after a race would compare with its tidal no reabsorption).
volume at rest: A. Higher than 1 (correct)
A. Higher (correct) B. Lower than 1
B. Lower C. Equal to 1
C. The same 4. Predict (increase/decrease/no change) what would happen
2. Predict (increase/decrease/no change) what will be the to the urine concentration capacity in a dog given a low-protein
effect on alveolar ventilation of breathing at higher frequency diet:
while keeping the same respiratory volume: A. Increase
A. Increase B. Decrease (correct)
B. Decrease (correct) C. No change
C. No change 5. Predict (increase/decrease/no change) what would happen
3. Predict (increase/decrease/no change) what would happen to plasma Na⫹ concentration, total body Na⫹ content, plasma
to functional residual capacity in a lung disease leading to K⫹ concentration, and total body K⫹ content if you provide a
emphysema: medication with an aldosterone antagonistic effect:
A. Increase (correct)
B. Decrease
C. No change Response

4. Predict (increase/decrease/no change) what would happen ⫹

Plasma Na concentration No change
to airflow resistance if you provide a medication causing Total body Na⫹ content Decrease
smooth muscle relaxation: Plasma K⫹ concentration Increase
Total body K⫹ content Increase
A. Increase
B. Decrease (correct) 6. Predict (increase/decrease/no change) what would happen
C. No change to urine elimination of tritratable acids if the organic load of
5. Predict (increase/decrease/no change) what would happen metabolic acids increases:
to arterial PO2 in a dog breathing 100% oxygen (oxygen A. Increase (correct)
therapy): B. Decrease
A. Increase (correct) C. No change
B. Decrease 7. Predict (increase/decrease/no change) what would happen
C. No change to both the volume and osmolarity of extra- and intracellular
6. Predict (increase/decrease/no change) what would happen fluid compartments if you provide a hypertonic saline solution
to ventilation and perfusion in a dog having its right pulmonary (i.e., 7.5% NaCl) intravenously:
artery blocked:
Response
Response
Volume Osmolarity
Right lung ventilation Decrease
Left lung ventilation No change Extracellular fluid compartment Increase Increase
Left lung perfusion No change Intracellular fluid compartment Decrease Increase

Appendix C: Samples of Assessed Renal Questions

GRANTS
1. Predict (increase/decrease/no change) what would happen This work was supported, in part, by The American Physiological Society
to the glomerular filtration rate, renal blood flow, and glomer- Teaching Career Enhancement Award (to M. J. Giuliodori).

Advances in Physiology Education • VOL 30 • DECEMBER 2006

Downloaded from www.physiology.org/journal/advances (184.149.000.204) on July 9, 2019.
 How We Teach
COLLABORATIVE PROBLEM SOLVING 173

REFERENCES 11. Lymna F. The responsive classroom discussion. In: Mainstreaming Di-
gest, edited by Anderson AS. College Park, MD: Univ. of Maryland
1. Bonwell C and Eison J. Active Learning: Creating Excitement in the College of Education, 1981.
Classroom. Washington, DC: George Washington Univ. Press, 1991.
12. Mazur E. Peer Instruction: a User’s Manual. Upper Saddle River, NJ:
2. Byrnes JP. Cognitive Development and Learning in Instructional Con-
Prentice Hall, 1997.
texts. Boston, MA: Allyn and Bacon, 1996.
13. Mazur E. Peer instruction: getting students to think in class. Proc ICUPE
3. Chi M, de Leeuw N, Chiu M, and La Vancher C. Eliciting self-
399: 981–988, 1997.
explanations improves understanding. Cogn Sci 18: 439 – 477, 1994.
4. Committee on Developments in the Science of Learning, Committee 14. Michael J and Modell H. Active Learning in Secondary and College
on Learning Research and Educational Practice, and National Re- Classrooms: a Working Model for Helping the Learner to Learn. Manwah,
search Council. How People Learn: Brain, Mind, Experience and School. NJ: Erlbaum, 2003.
Washington, DC: National Academy Press, 2000. 15. Michael J and Rovick A. Problem Solving in Physiology. Upper Saddle
5. Cortright RN, Collins HL, and DiCarlo SE. Peer instruction enhanced River, NJ: Prentice Hall, 1999.
meaningful learning: ability to solve novel problems. Adv Physiol Educ 16. Rao SP, Collins HL, and DiCarlo SE. Collaborative testing enhances
29: 107–111, 2005. student learning. Adv Physiol Educ 26: 37– 41, 2002.
6. Cortright RN, Collins HL, Rodenbaugh DW, and DiCarlo SE. Student 17. Rao SP and DiCarlo SE. Peer instruction improves performance on
retention of course content is improved by collaborative-group testing. quizzes. Adv Physiol Educ 24: 51–55, 2000.
Adv Physiol Educ 27: 102–108, 2003. 18. Rao SP and DiCarlo SE. Active learning of respiratory physiology
7. Crouch C and Mazur E. Peer instruction: ten years of experience and improves performance on respiratory physiology examinations. Adv
results. Am J Physics 69: 970 –977, 2001. Physiol Educ 25: 55– 61, 2001.
8. DiCarlo SE. Cell biology should be taught as science is practised. Nat Rev 19. Rowe M. Pausing principles and their effects on reasoning in sciences. In:
Mol Cell Biol 7: 290 –296, 2006. Teaching the Sciences: New Directions for Community Colleges, edited by
9. Johnson D, Johnson R, and Smith KA. Active Learning: Cooperation in Brawer F. San Francisco, CA: Jossey-Bass, 1980.
the College Classroom. Edina, MN: Interaction Book, 1991. 20. Thorndike EL and Woodworth RS. The influence of improvement in
10. Lujan HL and DiCarlo SE. First-year medical students prefer multiple one mental function upon the efficiency of other functions. Psychol Rev 8:
learning styles. Adv Physiol Educ 30: 13–16, 2006. 247–261, 1901.

Advances in Physiology Education • VOL 30 • DECEMBER 2006

Downloaded from www.physiology.org/journal/advances (184.149.000.204) on July 9, 2019.