2003 Loverude Et Al

You might also like

Download as pdf or txt
Download as pdf or txt
You are on page 1of 11

Helping students develop an understanding of Archimedes’ principle. I.

Research on
student understanding
Michael E. Loverude, Christian H. Kautz, and Paula R. L. Heron

Citation: American Journal of Physics 71, 1178 (2003); doi: 10.1119/1.1607335


View online: http://dx.doi.org/10.1119/1.1607335
View Table of Contents: http://scitation.aip.org/content/aapt/journal/ajp/71/11?ver=pdfcov
Published by the American Association of Physics Teachers

Articles you may be interested in


A Spring, Hooke's Law, and Archimedes' Principle
Phys. Teach. 43, 516 (2005); 10.1119/1.2120379

A beam chopping experiment to help students develop research skills


Am. J. Phys. 72, 695 (2004); 10.1119/1.1632490

Helping students develop an understanding of Archimedes’ principle. II. Development of research-based


instructional materials
Am. J. Phys. 71, 1188 (2003); 10.1119/1.1607337

Reconsidering Archimedes' Principle


Phys. Teach. 41, 340 (2003); 10.1119/1.1607804

Archimedes’ principle without the king’s crown


Phys. Teach. 36, 557 (1998); 10.1119/1.880136

This article is copyrighted as indicated in the article. Reuse of AAPT content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
158.49.113.199 On: Mon, 13 Oct 2014 16:04:42
Helping students develop an understanding of Archimedes’ principle.
I. Research on student understanding
Michael E. Loverude,a) Christian H. Kautz,b) and Paula R. L. Heron
Department of Physics, University of Washington, Seattle, Washington 98195-1560
共Received 4 February 2002; accepted 18 July 2003兲
This paper is the first of two that describe how research on student understanding of Archimedes’
principle is being used to guide the development of instructional materials on this topic. Our results
indicate that standard instruction on hydrostatics leaves many science and engineering majors
unable to predict and explain the sinking and floating behavior of simple objects. A number of
serious and persistent difficulties with the concepts and principles used to analyze such behavior are
identified. Although some of these difficulties are specific to the concept of the buoyant force, many
others seem to reflect lingering confusion about concepts that are widely assumed to be understood
by students before the study of hydrostatics begins. © 2003 American Association of Physics Teachers.
关DOI: 10.1119/1.1607335兴

I. INTRODUCTION has developed to supplement instruction in a standard intro-


ductory course.5,6 Approximately half the students take an
The Physics Education Group at the University of Wash- optional associated laboratory course. Also included were
ington 共UW兲 is engaged in an ongoing investigation of stu- UW students taking a second-year course that covers hydro-
dent understanding of hydrostatic pressure and buoyancy.1 statics and thermal physics. This course has an enrollment
Prior to the current phase of the investigation, we had devel- that ranges from fewer than 20 to more than 60 students, of
oped instructional materials on buoyancy for Physics by In- whom typically about 30% are physics majors. There is no
quiry, a set of laboratory-based modules primarily intended laboratory component. A year of introductory calculus-based
for use in special physics courses for prospective and prac- physics 共which does not include hydrostatics兲 is a prerequi-
ticing K–12 teachers.2,3 site. In both the algebra-based and second-year courses, the
This article describes the extension of our investigation to amount of time devoted to hydrostatics ranged from a few
science and engineering majors in typical university courses. days to about two weeks, depending on the instructor. This
Students who have studied hydrostatics in introductory phys- investigation also involved courses at other universities, in-
ics are expected to be able to account for sinking and floating cluding algebra-based physics at the University of Maryland,
in terms of net force and Newton’s laws. In a typical course, College Park, and calculus-based physics at Purdue Univer-
the instructor and textbook explain that the buoyant force is sity.
due to the variation of pressure between the top and bottom
surfaces of an object. The students are explicitly taught that
the buoyant force on an object is equal to the weight of the III. INSIGHTS FROM INITIAL INVESTIGATION
fluid displaced and the equation B⫽ ␳ gV disp is derived. How-
ever, we found that many students cannot apply Archimedes’ We began our study with the assumption that university
principle even in very simple situations. science majors would be able to give simple physical expla-
Lucid explanations by the instructor and practice in solv- nations to account for everyday phenomena involving sink-
ing standard homework problems are apparently insufficient ing and floating. Therefore, we based our initial efforts to
to help students overcome the trouble that they have with the probe their understanding of hydrostatics on a more chal-
underlying concepts and lines of reasoning. Under such cir- lenging phenomenon: the behavior of a Cartesian diver.7
cumstances, it is imperative to try to identify precisely the We conducted individual demonstration interviews with
basic difficulties so that they can be systematically ad- seven volunteers from the second-year course after all in-
dressed. This article provides the information about student struction on hydrostatics was completed. The students had
thinking that we found necessary in our efforts to design final course grades at or above the mean. The interviews
instruction that would more effectively accomplish this ob- were audiotaped and videotaped for later analysis. The stu-
jective. The companion article describes the development dents were asked to predict what would happen to a barely
and assessment of instructional materials.4 floating diver when the sealed plastic bottle in which it was
contained was compressed.8 Although all the students recog-
nized that the pressure in the container would increase, all
II. CONTEXT FOR INVESTIGATION but two predicted that the diver would rise. None of the
students could account for the subsequent observation that
Our investigation involved more than 2000 students. The the diver sank. After seeing the diver sink, some students
largest group consisted of UW students in the second quarter drew free-body diagrams that showed both an upward
of introductory algebra-based physics, which covers fluids, ‘‘buoyant force’’ and an additional, downward ‘‘pressure
thermal physics, and electricity and magnetism. Mechanics is force’’ 共sometimes described as the ‘‘weight of the water
a prerequisite. Enrollment in individual lecture sections of above the diver’’兲.9 Several students predicted that the sink-
this course ranges from about 50 to more than 200. The ing diver would come to rest before reaching the bottom.
course has no small sections, and, in particular, does not They seemed to assume that the phenomenon of objects
involve Tutorials in Introductory Physics, which our group floating underwater is relatively common. Although this be-

1178 Am. J. Phys. 71 共11兲, November 2003 http://aapt.org/ajp © 2003 American Association of Physics Teachers 1178
This article is copyrighted as indicated in the article. Reuse of AAPT content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
158.49.113.199 On: Mon, 13 Oct 2014 16:04:42
Fig. 1. The five-block problem.

havior is possible, we suspected that the students did not tion that the density of block 3, while greater than that of
have a firm grasp of the conditions under which this might block 2, is less than or exactly equal to that of water were
occur. also considered correct. For example, the response: ‘‘If 2
barely floats then 3 probably sinks, but it could be exactly
IV. STUDENT UNDERSTANDING OF ARCHIMEDES’ neutrally buoyant and stay where it’s released’’ was consid-
PRINCIPLE ered correct 共see Fig. 2兲.
This problem has been given in two sections of the intro-
The degree of confusion in the students’ responses during ductory course in which instruction on buoyancy had not yet
the Cartesian diver interviews was such that we were unable begun. About 25% of the students answered correctly (N
to identify the nature of their difficulties. To probe their un- ⫽218). In two classes in which standard lecture instruction
derstanding further, we developed a series of written prob- had been completed, about 10% answered correctly (N
lems that involved much simpler situations. The problems
were administered as either nongraded quizzes 共of about 10 ⫽151) 13 共given the expected size of the random fluctuations
minutes兲 or examination questions.10 In such settings it is and the large number of variables involved, we do not assign
difficult to estimate the effect of, much less control, many of significance to the apparent difference in results obtained be-
the variables. However, the data provide a measure of the fore and after instruction兲.14 About 50% of the students in the
prevalence of certain difficulties and can provide guidance UW second-year course gave a correct answer (N⫽101) af-
for setting priorities in instruction.11 ter standard instruction. In the introductory course at Purdue,
the success rate after standard instruction was about 40%
A. Predicting sinking and floating (N⫽765).
The problem describes five blocks of identical size and In all the classes, the most common incorrect response
shape but varying mass, with block 1 being the lightest and showed blocks 3 and 4 at intermediate depths in the water
block 5 the heaviest 共see Fig. 1兲. Each block is held approxi- 共see Fig. 3兲. We refer to this as the ‘‘descending line’’ re-
mately half-way down in the water in an aquarium and re- sponse. This answer could only be considered correct if it
leased. The final positions of blocks 2 and 5 are shown in an was assumed that the density of the water varies so that
accompanying figure and described in the text 共block 2 objects of different density might each match the average
barely floats and block 5 rests on the bottom of the tank兲.12 density of the surrounding water at different depths. How-
Students were asked to sketch the final positions of the other ever, water had been treated as incompressible in the courses
three blocks and to explain their reasoning. in which the students were enrolled.15 Moreover, although
A correct answer to this problem requires an inference some students explicitly related the final positions of the
from the given information that block 2 barely floats, as il- blocks to their relative densities, none referred to the density
lustrated by the student quote: ‘‘Since m 1 ⬍m 2 , and m 2 has of the surrounding liquid. In all but the initial version of this
a density just about equal to that of H2 O, m 1 must have a problem, we were careful to remind students to treat water as
lighter density and thus will float more. Block 3’s density incompressible. This warning has had no apparent effect on
will be more than that of H2 O and m 2 and so it will sink.’’ the outcome.
The problem does not specify in detail how the masses of the Explanations in which forces are mentioned are somewhat
blocks vary. Therefore, answers consistent with the assump- more prevalent after formal instruction than before. How-

Fig. 2. Possible correct answers to the


five-block problem.

1179 Am. J. Phys., Vol. 71, No. 11, November 2003 Loverude, Kautz, and Heron 1179
This article is copyrighted as indicated in the article. Reuse of AAPT content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
158.49.113.199 On: Mon, 13 Oct 2014 16:04:42
the bottom. None mentioned a density gradient or compress-
ibility. When shown the outcome of the experiment, they
expressed surprise and were unable to resolve the inconsis-
tency between their prediction and observation.

B. Relating the buoyant force on submerged objects to


their volume
In their responses to the five-block problem, some students
explicitly linked the buoyant force on a block to its depth or
mass. For example, one wrote: ‘‘Assuming that the masses
increase in equal intervals, it seems likely that the buoyant
forces would also increase linearly.’’ On a version in which
Fig. 3. An example of the incorrect ‘‘descending line’’ response to the we asked students to draw free-body diagrams for blocks 2
five-block problem. and 5 just after being released, about 60% of those who drew
reasonable diagrams 共that is, with no additional or missing
forces兲 clearly indicated that the buoyant forces on the two
ever, even after instruction, only about 10% to 20% of the blocks would be different (N⬃200). 16 Most suggested that
UW introductory students referred to forces, while about the buoyant force on the lighter block would be greater.
20% to 30% referred to the densities of the blocks 共the frag- The responses to the five-block problem led us to design a
mentary, incoherent nature of many of the student responses series of questions to examine the ability of students to relate
permits only a rough categorization兲. The majority of stu- the buoyant force on a submerged object to its volume. One
dents simply referred to the mass or weight of the blocks. version is shown in Fig. 4. Three blocks of equal volume are
For example, one student wrote, ‘‘Block 4 is heavier than 3 suspended from strings and completely submerged in a tank
but lighter than 5 so it descends somewhere halfway between of water 共the text states that all the blocks have been ob-
3 and 5.’’ No obvious correlation exists between the success served to sink兲. In other versions, blocks that have been ob-
rate and the type of explanation offered. served to float are attached to strings connected to the bottom
The five-block task also was used as the basis for indi- of the tank. In each case, a pair of blocks of different mass is
vidual interviews with 12 students from the second-year at the same depth and/or a pair of blocks of the same mass
course after all relevant instruction had been completed. The are at different depths. In some versions, a block of the same
students were given five small cubes of equal volume but volume as one of the others floats freely 共see Fig. 5兲. Stu-
different material and asked to sort them by weight. The dents are asked to rank the magnitudes of the buoyant forces
interviewer then placed blocks 2 共nylon兲 and 5 共iron兲 in a 共some versions ask for a pairwise comparison兲. To answer
beaker of water and asked each student to predict the final correctly, students need to recognize that the submerged
positions of the other blocks 共wood, aluminum, and lucite兲 blocks displace the same volume of water. Therefore, the
and to explain their reasoning. About half of the students buoyant forces are all equal. The floating block 共if present兲
gave the descending line response. These students seemed to displaces a smaller volume and therefore experiences a
be strongly committed to their prediction. Despite often smaller buoyant force.
lengthy discussions, all of the students who initially gave this After all instruction, a correct comparison of the buoyant
response maintained that blocks 3 and 4 would ‘‘float’’ above forces on all three submerged blocks was given by 20% of

Fig. 4. An example of a written problem in which three


blocks of equal volume are suspended underwater by
strings. There are two blocks of the same mass at dif-
ferent depths 共A and B兲 and two blocks of different
mass at the same depth 共A and C兲. Students are asked to
rank the buoyant forces and/or the tensions in the
strings.

1180 Am. J. Phys., Vol. 71, No. 11, November 2003 Loverude, Kautz, and Heron 1180
This article is copyrighted as indicated in the article. Reuse of AAPT content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
158.49.113.199 On: Mon, 13 Oct 2014 16:04:42
Fig. 5. A written problem on the buoy-
ant force that involves several com-
parisons. Students are first asked to
compare the buoyant force on a block
that is held underwater 共B兲 to the
buoyant force on an identical block
that is floating 共A兲 and a block of dif-
ferent mass submerged to the same
depth 共C兲. All of the blocks are then
described as being moved to a differ-
ent liquid. Students are asked how the
buoyant forces exerted on blocks A
and B compare to those exerted on the
same blocks when in water. 共The ques-
tion about block A is referred to in the
text as version 1 of the floating blocks
question.兲

the students in the introductory course (N⫽198) and about 2. Blocks of the same mass submerged to different depths
60% in the second-year course (N⫽83). The errors made by
students in both courses were similar in nature. There was About 55% of the introductory students and 80% of the
very little variation among the different sections of each second-year students answered that the buoyant forces on
course 共with one exception兲.17 identical blocks submerged to different depths would be the
same. About 30% of the introductory students and 15% of
the second-year students claimed that the buoyant force is
1. Blocks of different mass submerged at the same depth greater on the block at greater depth. Interpreting these re-
About 35% of the introductory students and 65% of the sponses is complicated. The written explanations were often
second-year students correctly compared the buoyant forces too vague to rule out the possibility that some students, al-
on blocks of different mass submerged at the same depth. though not explicitly stating so, were assuming a significant
The most common incorrect answer, given by 45% of the density gradient, in which case their answers would be cor-
introductory students and 20% of the second-year students, rect. To gain further insight, we administered some versions
was that the more massive block would have a greater buoy- in which students were also asked to compare the tensions in
ant force. the strings holding the blocks. About 35% of the introductory
Failure to distinguish mass, volume, and density: A small students stated that the tensions would be the same (N
but significant fraction of students referred to the masses of ⫽517).
the objects, not their volumes. For example, in answering Confusion about the relationship between pressure and
the problem in Fig. 4, one student wrote: ‘‘The buoyant force buoyant force: Vague references to the depth dependence of
on C is less than that of A & B—it displaces less water due pressure were common. For example, one student’s explana-
to its smaller mass.’’ 共This tendency is discussed in detail in tion consisted of ‘‘F B of B is greatest because the pressure on
Sec. V.兲 B is the greatest.’’ Most students seemed aware that the
Misuse of formulas: Other students used a formula that is forces exerted by the fluid on the individual surfaces of an
appropriate only for floating objects to support the claim that object increase with increasing depth. They may have as-
a heavier object experiences a greater buoyant force. For sumed that the vector sum of these forces must also increase.
example, one wrote: ‘‘F B⫽mg thus A & B are equal and However many answers suggested that students did not rec-
since C has less mass, smaller F B .’’ ognize that the term ‘‘buoyant force’’ refers to this vector
Failure to distinguish buoyant force from buoyancy: Many sum.
students who claimed that a lighter block experiences a By comparing the responses of individual students to the
greater buoyant force apparently failed to distinguish this questions on the buoyant forces and the string tensions, we
well-defined physical concept from the more general notion were better able to distinguish between a misuse of terminol-
of ‘‘buoyancy.’’ For example, one student wrote, ‘‘Buoyancy ogy and a misunderstanding about the forces exerted on a
can be looked at as the tendency for an object to be able to submerged object. For example, a student who associated the
float. Since C has the lowest density it will float the easiest.’’ term buoyant force with the force exerted upward on the

1181 Am. J. Phys., Vol. 71, No. 11, November 2003 Loverude, Kautz, and Heron 1181
This article is copyrighted as indicated in the article. Reuse of AAPT content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
158.49.113.199 On: Mon, 13 Oct 2014 16:04:42
bottom surface of a block might claim that the buoyant force
is larger on the deeper block. That student could still dem-
onstrate an understanding of the physics by taking into ac-
count the downward forces exerted on the upper surfaces of
the blocks and concluding that the tensions in the strings are
the same. Of the students who claimed that the buoyant force
is greater at greater depth, about 20% answered that the ten-
sions are equal. About 45% stated that the tension would be
greater in the string supporting the lower block. This answer
also was given by 35% of the students who claimed that the
buoyant forces on the two blocks are the same. Most expla-
nations referred to the greater amount of water above the
deeper block. One explained: ‘‘Since all buoyant forces are Fig. 6. A written question in which students are asked to compare the
equal, I will ignore them. Since B is the lowest, it has the buoyant forces exerted on two identical blocks floating in different liquids.
most downward force on it from the water. T B ⬎T A .’’ Many This question is referred to in the text as version 2 of the floating blocks
question.
student responses are consistent with thinking that both a
buoyant force and an additional force are exerted on an ob-
ject by the surrounding fluid. These results suggest that the of the students claimed that the buoyant forces are the same
problem is more serious than confusion about the compress- in both liquids, typically noting that the block is at the same
ibility of water. depth in both.

3. Identical blocks fully and partially submerged D. Relating the buoyant force on floating objects to
After all instruction, 60% of the students in the introduc- their mass „or weight…
tory course (N⫽792) and 80% of the students in the second- As noted, many students applied the formula F B⫽mg to
year course (N⫽49) correctly stated that the buoyant force submerged objects. However, others failed to apply this re-
on a floating block is less than that on an identical block held lation even in cases in which it is appropriate. For example,
below the surface. About 30% of the introductory students on an examination question in the introductory course, stu-
claimed that the buoyant forces are the same. dents were asked to determine the magnitude of the buoyant
Misuse of formulas: Some students who claimed that the force acting on a piece of ice floating in a bucket of water.
buoyant forces would be equal noted that the blocks have the The mass of ice was given as 1000 g. Only about 50% of the
same mass and referred to the equation F B⫽mg. This rela- students answered correctly (N⫽65). Many tried to find the
tionship applies only to objects that are floating freely. volume displaced by the ice.
Failure to recognize that the relevant volume is that of the We designed some questions to gauge the ability of stu-
fluid displaced: Other students noted that the volumes of the dents to relate the buoyant force on a floating object to its
blocks are the same. The following explanation was typical: weight. Version 1 concerns the floating block in the problem
‘‘They both will displace the same amount of water. F B shown in Fig. 5. The block is described as being moved from
⫽ ␳ gV⇒same volume same buoyant force.’’ Many students a tank of water to a different fluid, where it clearly floats
apparently did not realize that the relevant volume is that of lower. In version 2 of this question, two identical blocks are
the liquid displaced, not that of the block itself.18 More wor- shown floating in different liquids in side-by-side containers.
risome is the possibility that the students did not recognize The volumes displaced by the blocks are clearly different
the difference between these two quantities. 共see Fig. 6兲. In both versions, Newton’s second law can be
used to deduce that, because the blocks are identical and at
rest, the buoyant forces supporting the blocks must be of
C. Relating the buoyant force on submerged objects to equal magnitude 共knowledge of hydrostatics is not even re-
the fluid density quired兲.
One written problem included a question on the role of These questions were asked in the UW introductory course
fluid density in determining the buoyant force on submerged before and after instruction. About 20% of the students an-
objects. A block is shown attached by a string to the bottom swered version 1 correctly (N⫽285) while about 40% an-
of a tank of water 共see Fig. 5兲. The block is then attached by swered version 2 correctly (N⫽424). Among the sections
the same string to the bottom of a tank containing a different given each version, very little variation was seen. Therefore,
fluid. The students are asked how the buoyant forces exerted we attribute the higher success rate on version 2 to the fact
on the block by the two fluids compare 共most students cor- that it involves a more straightforward side-by-side compari-
rectly inferred from the behavior of a floating object that the son than version 1. On an examination in the introductory
density is less than that of water兲. This problem was admin- algebra-based course at Maryland, about 20% of the students
istered after all instruction in the introductory course.19 answered version 2 correctly (N⫽161).
About 75% of the students correctly stated that the buoyant Failure to consider all the variables: A number of students
force is greater in the more dense liquid (N⫽164). However, mentioned only the different volumes displaced in the differ-
many of these students also claimed that the buoyant force ent liquids. For example, in answering the question in Fig. 6,
on a floating object would be greater in the more dense liq- one student wrote, ‘‘We know the buoyant force on D is
uid. 共They typically noted that the block floats at a higher greater because more amount of liquid is displaced.’’ Other
level in the more dense liquid. This difficulty is discussed students mentioned only the fluid densities and concluded
below.兲 The tendency to associate the buoyant force with that the buoyant force is greater on the block in the more
depth also appeared in responses to this question. About 20% dense liquid.

1182 Am. J. Phys., Vol. 71, No. 11, November 2003 Loverude, Kautz, and Heron 1182
This article is copyrighted as indicated in the article. Reuse of AAPT content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
158.49.113.199 On: Mon, 13 Oct 2014 16:04:42
Fig. 7. A student response to the written problem in
which blocks of equal volume and different mass are
lowered into cylinders of water.

Use of ‘‘compensation’’ reasoning: A number of students A. Determining the volume displaced by objects
argued that because the displaced volume is greater in one submerged in water
case, while the fluid density is greater in the other, the buoy-
ant forces on the two blocks must be the same. For example, We administered a written problem in which students are
one student wrote: ‘‘It depends on the density of the liquid. told that two blocks of the same size and shape are lowered
Since in the diagram block E appears to be less submerged into graduated cylinders that have identical initial water lev-
than block D it is displacing a smaller amount of liquid, but els. One block is aluminum; the other brass. The final water
assuming the liquid is more dense the differences should level in the cylinder containing the aluminum block is
cancel each other out, creating equal buoyant forces.’’ This shown. Students are asked how it will compare to the final
argument seems to be based on the assumption that the two water level in the cylinder containing the brass block.
variables 共fluid density and displaced volume兲 are inversely This problem differs from the version described above in
proportional and will thus compensate for each other to en- that the two objects are suspended at different levels 共the
sure that the product remains constant. This type of argument aluminum block is lowered farther than the brass block兲.
has been observed in other contexts.20 In this case, that the Informal discussions, in which some students seemed to in-
fluid density and displaced volume of each liquid are propor- terpret ‘‘volume displaced’’ as the volume of water above a
tional can be considered to be a consequence of the equality submerged object, prompted this modification. We were also
of the buoyant forces. aware that some students fail to equate the change in water
Tendency to relate the positions of objects to the forces level in a container with the volume of water displaced.
exerted on them: The most common incorrect answer was Therefore, in some versions students are asked to compare
that the buoyant force is greater on the cube that floats at a both the final water levels and the volume of water displaced
higher level. Most students who gave this answer related the by the two blocks. These issues proved to affect fewer than
positions of the blocks to the forces exerted on them. For 10% of the students.
example, one wrote: ‘‘The weight of the cubes are the same, The problem was administered in several sections of the
which means it has to be the buoyant force which is causing UW introductory course, in some cases immediately prior to
the difference in depth of the block. Since E is higher, it is the first laboratory experiment. In all, about 75% of the stu-
being pushed up more which means its buoyant force is dents correctly stated that the final water levels would be the
greater.’’ 共This tendency is discussed in greater detail in Sec. same (N⫽395). At Purdue, about 65% of the students have
VI.兲 answered correctly (N⫽250). Although many students sup-
plied correct explanations, a few gave glib answers indicat-
ing that the answer is too obvious for explanation. At both
UW and Purdue, about 20% of the students have stated that
V. STUDENT UNDERSTANDING OF RELATED the water level would rise more in the cylinder containing
CONCEPTS: VOLUME DISPLACEMENT the heavier block.
Failure to distinguish the concepts of mass, volume, and
Many students’ responses to the problems described above density from one another: Most of the students who claimed
seem to reflect lingering confusion about concepts that are that the final water level would be higher in the cylinder
widely assumed to be understood by university students be- containing the brass block explicitly cited its greater mass:
fore the study of hydrostatics begins. In a number of cases ‘‘Since B has a higher mass, the volume of water displaced
students explicitly linked the amount of water displaced by a will be more, so the water level will be higher.’’ One student
submerged object to the mass of the object. We had observed drew the sketch shown in Fig. 7. Other students focused on
similar tendencies among elementary and middle school density: ‘‘The volume displaced by A is less than volume
teachers and nonscience majors: when asked what would displaced by B because water displacement is proportional to
happen to the water level in two identical graduated cylin- the object’s density.’’
ders when two metal balls of the same volume but different
mass were submerged, the majority predicted that the level B. Determining the volume displaced by objects
would rise more for the heavier ball. This response can be submerged in other fluids
seen as a failure to distinguish between mass and
volume.21,22 We decided to explore the possibility that some The majority of students recognized that the volume of
of the introductory physics students in the present investiga- water displaced by a submerged object is equal to the volume
tion might have similar difficulties. of the object itself. However, in informal discussions, a few

1183 Am. J. Phys., Vol. 71, No. 11, November 2003 Loverude, Kautz, and Heron 1183
This article is copyrighted as indicated in the article. Reuse of AAPT content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
158.49.113.199 On: Mon, 13 Oct 2014 16:04:42
Fig. 8. Free-body diagrams for blocks 2 and 3 for a
student who gave the descending line response to the
five-block problem. The diagrams indicate that the two
blocks are subject to the same buoyant force, which is
equal to the weight in the case of block 2 共which barely
floats兲 but less than the weight for block 3, which the
student predicts will remain at rest near the center of the
tank.

of these students mentioned the density of water, implying ⬍W4⬍W5 and all are submerged ⬖V disp are all equal ⬖FB 3
that it had some significance in this situation. We therefore ⫽FB 4 ⫽FB 5 ⬖W 4 ⬎FB 4 and W 5 ⰇFB 5 so block 4 will be
administered a variation of the problem described above in higher than block 5.’’
which the two blocks are identical. One cylinder contains In one version of the problem, students were asked to
water; the other contains oil. The first part of the question draw free-body diagrams for blocks 2 and 3 at their final
asked students to compare the liquid levels in two identical positions. Several students who predicted that block 3 would
cylinders if each contains 500 ml of liquid. The second part remain suspended beneath the surface clearly indicated a
of the question asks students to compare the final liquid lev- nonzero net force in their diagram. The student diagram in
els if the two blocks are lowered to the same level in their Fig. 8 provides an example. Block 2, which barely floats, is
respective cylinders. This question was asked in the intro- depicted as having equal upward and downward forces while
ductory lab before any instruction (N⫽311). block 3, also at rest, is depicted as having a net downward
About 85% of the students answered the first part cor- force. This diagram suggests that the block’s level is related
rectly. The remainder cited the different densities of the liq- to the degree of imbalance between the weight and the buoy-
uids. Of the students who recognized that the initial liquid ant force.
levels are the same, about 60% went on to answer correctly For the floating blocks questions, many students who re-
that the final levels will also be the same. 共The responses of lied on the formula F B⫽ ␳ gV made no reference to the mo-
students who answered the first part incorrectly proved too tions or positions of the blocks. The inconsistency of their
difficult to interpret.兲 answers with Newton’s second law apparently went unno-
Tendency to relate the volume displaced to the density of ticed. Other students explicitly attributed the different posi-
the liquid: Most of the students who answered incorrectly tions of the blocks to a difference in the forces exerted on
claimed that a greater volume of the less dense liquid would them. Typically, the higher level of one of the blocks was
be displaced. The students typically argued that a greater taken as evidence of a larger upward force.
volume was necessary to compensate for the lower density,
implying that the same mass of liquid must be displaced in
both cylinders.
B. Applying principles from Newtonian dynamics to
mechanical equilibrium problems
VI. STUDENT UNDERSTANDING OF RELATED
PRINCIPLES: NEWTON’S LAWS Many student responses on the problems described above
reflected a poor understanding of Newtonian dynamics. In an
All of the students in our study had completed a previous
effort to determine if the same issues would arise in contexts
course in mechanics.23 Nevertheless, in their answers to the
free of the complications of hydrostatics, we decided to ad-
hydrostatics problems, many students failed to apply New-
minister some mechanical equilibrium problems analogous
ton’s laws correctly, if at all.
to those described above. The problems either require stu-
A. Recognizing the need to apply „or being able to dents to predict the motions of identical objects subjected to
apply… principles from Newtonian dynamics to the same forces or to draw inferences about the forces ex-
hydrostatics erted on identical objects that are at rest. In both cases, the
objects are 共at least initially兲 at different levels.
In responding to the five-block problem, many students The Atwood’s machine problem was suggested by a task
apparently assumed that a block would move from the level used to probe student understanding of gravity22 in which
at which it was released to some higher or lower level. Ref- students tended to assume that an imbalance in weights on
erences to ‘‘equilibrium’’ were common.24,25 Many explana- the two sides of an Atwood’s machine causes the weights to
tions directly or indirectly contradicted Newton’s laws. Even hang at different levels, rather than to accelerate. In our ver-
students who recognized that all the blocks are initially sub- sion, two identical blocks are connected by a light, inexten-
ject to the same buoyant force failed to conclude that some sible string that runs over a frictionless pulley 共or a smooth
would accelerate upward while others would accelerate peg兲. A diagram shows the blocks being held in place, with
downward. For instance, one student wrote: ‘‘Block 4 will be one lower than the other 关see Fig. 9共a兲兴. The students are
lower than block 3 and higher than block 5 because W 3 asked to predict what will happen to the blocks when they

1184 Am. J. Phys., Vol. 71, No. 11, November 2003 Loverude, Kautz, and Heron 1184
This article is copyrighted as indicated in the article. Reuse of AAPT content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
158.49.113.199 On: Mon, 13 Oct 2014 16:04:42
Fig. 9. Mechanical equilibrium problems. 共a兲 The At-
wood’s machine problem. 共b兲 The springs problem.

are released. They need to recognize that because the weights students who gave incorrect answers to the Atwood’s ma-
are the same, there will be no acceleration and the blocks chine problem referred to ‘‘equilibrium’’ in their explana-
will not move. tions. For example, one student wrote: ‘‘Block A will move
In the springs problem, students are shown two identical up 共and B will move down兲 until they reach the same level.
blocks that are supported at different levels on nonidentical Having the same mass, the blocks will reach an equilib-
vertical springs 关see Fig. 9共b兲兴. Students are asked to com- rium.’’ Some students who referred to forces came to the
pare the forces exerted on the blocks by the springs. Because same conclusion: ‘‘Block B will fall down as Block A rises
the blocks are identical and at rest, the forces exerted by the until they rest at an equal height. The same forces are acting
springs must be equal in magnitude. on equal objects therefore they will rest at equal positions.’’
The problems were given in several sections of the UW The underlying assumption—that objects of the same weight
introductory course and one section of the second-year or subject to the same forces should be at the same level—
course. All of the students had completed a previous course was widespread. The corresponding idea, that objects of dif-
in mechanics. In most cases, the two problems were admin- ferent weight should be at different levels, is reminiscent of
istered at the same time. Correct answers to both problems many student responses to the five-block problem. In that
were given by about 45% of the introductory students and case, a number of students who gave the incorrect descend-
about 65% of the second-year students. ing line answer simply stated that a heavier block would sink
About 65% of the introductory students correctly an- to a lower level.
swered the Atwood’s machine problem (N⫽327). 26 In the Assumption that different forces are required to keep iden-
second-year course, about 70% made a correct prediction tical objects at different levels: In their answers to the
(N⫽46). The most common incorrect answer in both springs problem, many students attributed the different
courses was that the blocks would move until they reach the heights of the blocks to a difference in the applied forces. For
same level 共a few students said that the blocks would oscil- example, one wrote: ‘‘I would suspect that since the springs
late about this new level兲. The springs problem was an- are not yielding to the ground, then the spring forces are
swered correctly by 55% of the students in the introductory greater than the gravitational force, and that the spring force
course (N⫽424) and 80% of the students in the second-year on the right spring is greater than the one on the left.’’ This
course (N⫽46). 27 About 25% of the introductory students explanation suggests that a non-zero net force directed up-
claimed that the force exerted by the spring on the higher ward is exerted on both blocks and that the magnitude of the
block must be greater than that on the lower block. This was net force is greater on the higher block.
also the most common incorrect answer in the second-year Failure to consider all the variables: Some students
course. claimed that a greater force is exerted by the spring on the
Assumption that objects of the same weight (or subject to lower block. Most stated that this spring is ‘‘more com-
the same forces) should be at the same level: Many of the pressed,’’ but did not mention a spring constant. These stu-

1185 Am. J. Phys., Vol. 71, No. 11, November 2003 Loverude, Kautz, and Heron 1185
This article is copyrighted as indicated in the article. Reuse of AAPT content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
158.49.113.199 On: Mon, 13 Oct 2014 16:04:42
dents seem to have used Hooke’s law, but failed to take into The majority of introductory students in this investigation
account that the springs are different, despite a comment to failed to demonstrate an understanding of the concepts re-
this effect in the problem statement. quired for a correct, coherent argument. Therefore, we be-
lieve that for instruction on this topic to be successful, the
specific incorrect ideas expressed by students need to be ad-
dressed. In Ref. 4, we describe the instructional materials we
VII. CONCLUSION designed to help students deepen their understanding.
Standard instruction on hydrostatics apparently leaves
many students unable to predict and explain the sinking and ACKNOWLEDGMENTS
floating behavior of simple objects. In particular, many stu-
dents were unable to identify the forces exerted on an object This investigation has benefited from the efforts of many
by a fluid and to recognize the factors that govern the mag- present and past members of the Physics Education Group.
nitudes of those forces. It was clear that students had learned The authors are deeply grateful for the important contribu-
tions of Lillian C. McDermott to the preparation of this paper
the formulae F B⫽mg and F B⫽ ␳ gV. However, when no
and to the investigation that it describes. Special thanks are
numbers were given, students frequently used these equa-
due to Peter S. Shaffer and Matthew Cochran and to the
tions inappropriately, incorrectly, or not at all. The failure of
instructors at the University of Washington, the University of
students to recognize when their answers contradicted funda-
Maryland, College Park, and Purdue University in whose
mental principles was widespread.
classes the written problems were administered. The authors
Many students did not recognize the crucial role of dis-
gratefully acknowledge the support of the National Science
placed volume in determining the buoyant force. The as-
Foundation through Grant Nos. DUE 9354501 and DUE
sumption that the buoyant force on a submerged object de-
9727648.
pends on its mass was common. For some students, this
belief can be interpreted as a failure to distinguish clearly a兲
Department of Physics, California State University–Fullerton, Fullerton,
between the concepts of mass and volume. Although intro- CA 92834.
ductory students who exhibited this difficulty were a minor- b兲
Department of Physics, Technical University of Hamburg-Harburg.
1
ity, for them it may have been a critical barrier to the devel- This research is described in greater detail in M. E. Loverude, ‘‘Investiga-
opment of an understanding of sinking and floating tion of student understanding of hydrostatics and thermal physics and of
共generally, these students were not struggling academically兲. the underlying concepts from mechanics,’’ Ph.D. dissertation, Department
of Physics, University of Washington, 1999.
Many students attributed a dependence on depth to the 2
L. C. McDermott and the Physics Education Group at the University of
buoyant force that does not appear to stem from an exagger- Washington, Physics by Inquiry 共Wiley, New York, 1996兲.
ated view of the compressibility of water. Rather, it seems to 3
These materials are discussed in the companion article: P. R. L. Heron, M.
reflect confusion about the relationship between pressure, E. Loverude, P. S. Shaffer, and L. C. McDermott, ‘‘Helping students de-
which increases with increasing depth, and buoyant force, velop an understanding of Archimedes’ principle. II. Development of
which only increases if the pressure gradient is nonlinear. research-based instructional materials,’’ Am. J. Phys. 71, 1188 –1195
共2003兲, following paper.
Our results demonstrate that the confusion is not simply lin- 4
See Ref. 3.
guistic. 5
L. C. McDermott, P. S. Shaffer, and the Physics Education Group at the
Any attempt to solve the five-block problem on the basis University of Washington, Tutorials in Introductory Physics 共Prentice-
of forces requires an understanding of Newtonian dynamics. Hall, Upper Saddle River, NJ, 2002兲.
6
Ample evidence shows that many students successfully com- Although the tutorials are used in algebra-based courses in several colleges
plete introductory physics without having developed such an and universities, the algebra-based course at the University of Washington
understanding.28 We found that difficulties with mechanical does not have small sections at present.
7
A Cartesian diver is an object of variable density that can be made to sink
equilibrium were serious and widespread. Many students or float in response to pressure changes in a fluid.
made statements in their explanations for the descending line 8
The diver used in the interviews is made from the transparent bulb of a
prediction that betrayed a failure to apply the second law. plastic dropping pipette. A weight is attached to its open end. The bulb is
Previous research has documented the tendency to associate placed with its neck down in a two-liter plastic soda bottle nearly filled
the net force on an object with its velocity, rather than its with water. The bulb contains water with a bubble of air above the water.
acceleration.29 In our study, students often appeared to be The amount of air is adjusted so that the diver initially barely floats. When
the soda bottle is squeezed, the increase in pressure leads to a decrease in
using an incorrect analogue of Newton’s second law in the volume of the air bubble inside the bulb. Water enters through the neck
which position, rather than acceleration, is assumed to be of the diver, thus increasing its density so that the diver sinks. 共The stu-
proportional to net force. Thus although previous studies dents had seen a lecture demonstration involving a Cartesian diver in class,
have shown that students often assume that any object hav- but the apparatus differed significantly in appearance from that used in the
ing zero velocity 共even at an instant兲 has zero net force, we 9
interviews.兲
found that students often assumed that an object would re- Evidence that students confuse the concepts of pressure and weight can be
found in Ref. 1, as well as in E. Engel Clough and R. Driver, ‘‘What do
main at rest in spite of, or perhaps because of, a nonzero net
children understand about pressure in fluids?,’’ Res. Sci. Tech. Ed. 3, 133–
force. 143 共1985兲 and in P. Kariotoglou and D. Psillos, ‘‘Pupils’ pressure models
On the five-block problem, many of the written explana- and their implications for instruction,’’ ibid. 11, 95–108 共1993兲.
tions lacked a logical link between the ideas expressed 10
In most cases, we have found that the results for a given problem are
共whether correct or incorrect兲 and their predictions. In some essentially the same whether administered on a non-graded quiz or on a
cases, written statements appeared to contradict the predic- graded course examination. For related results, see C. Henderson, ‘‘Com-
tion. Moreover, essentially all of the incorrect ideas ex- mon concerns about the force concept inventory,’’ Phys. Teach. 40, 542–
547 共2002兲.
pressed were associated with the same prediction. Thus, it is 11
In our experience, the variation in results from class to class is usually
possible that many explanations were attempts to justify a very small. In particular, for large classes (N⬃100 or more兲 we have
prediction based on intuition, rather than attempts to articu- found that the success rates of individual classes usually fall within about
late a line of reasoning that leads to a prediction. 5% of the mean taken over several classes. Therefore, we have reported all

1186 Am. J. Phys., Vol. 71, No. 11, November 2003 Loverude, Kautz, and Heron 1186
This article is copyrighted as indicated in the article. Reuse of AAPT content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
158.49.113.199 On: Mon, 13 Oct 2014 16:04:42
percentages rounded to the nearest five percent. In the second-year course standing of the work-energy and impulse-momentum theorems,’’ Am. J.
the number of students is typically much smaller and we collected data Phys. 55, 811– 817 共1987兲 and M. E. Loverude, C. H. Kautz, and P. R. L.
from fewer sections. Our experience has been that the variation is compa- Heron, ‘‘Student understanding of the first law of thermodynamics: Relat-
rable to that in the case of larger classes. However, the primary function of ing work to the adiabatic compression of a gas,’’ ibid. 70, 137–148 共2002兲.
21
the data from the second-year course is to illustrate the extent to which Difficulties in distinguishing mass and volume are described in M. L.
similar conceptual issues surface among students who typically have a Rosenquist, ‘‘Improving preparation for college physics of minority stu-
stronger mathematical background and who have had more instruction on dents aspiring to science-related careers,’’ Ph.D. dissertation, Department
the topic. of Physics, University of Washington, 1982, p. 74.
12 22
In many of the written questions discussed in this article, blocks are de- For related results, see J. McKinnon and J. W. Renner, ‘‘Are colleges
picted as floating with surfaces parallel to the water surface. In practice concerned with intellectual development?,’’ Am. J. Phys. 39, 1047–1052
these blocks would float with one corner up. We have not found this to be 共1971兲 and J. McKinnon, ‘‘Earth science, density, and the college fresh-
a source of confusion for students nor did we wish to explore the issue of man,’’ J. Geol. Educ. 19, 218 –220 共1971兲.
23
stability with these questions. We have revised the questions so that cur- We note that the mechanics courses did not include relevant tutorials from
rent versions refer to cylinders instead of blocks. Ref. 5. However, even if they had, it may be unrealistic to assume that
13
Only classes in which we can be certain that none of the students had students would gain enough facility during their initial study of dynamics
participated in the relevant lab are included. In one case the students were to apply their knowledge of forces in the more complicated environment
not asked to write their names and so the lab students cannot be separated of a liquid.
24
from the rest of the class. Therefore we have excluded these results. Also For examples in mechanics in which students inappropriately refer to equi-
excluded are results from a class in which about half the students had librium positions, see R. F. Gunstone, ‘‘Student understanding in mechan-
participated in the lab, and there was evidence that a number of students ics: A large population survey,’’ Am. J. Phys. 55, 691– 696 共1987兲 and R.
changed their answers while taking the ungraded quiz 共an unusual occur- F. Gunstone and R. T. White, ‘‘Understanding of Gravity,’’ Sci. Educ. 65,
rence兲. Results obtained after the lab are discussed in detail in Ref. 4. 291 共1981兲.
14 25
For instances in which results have been similar before and after instruc- Similar difficulties with the concept of equilibrium have been documented
tion, see, for example, L. C. McDermott, ‘‘Millikan Lecture 1990: What in the context of thermal physics and rigid-body dynamics. See C. Kautz,
we teach and what is learned—Closing the gap,’’ Am. J. Phys. 59, 301– ‘‘Investigation of student understanding of the macroscopic and micro-
315 共1991兲 and L. C. McDermott, ‘‘Guest comment: How we teach and scopic behavior of an ideal gas,’’ Ph.D. dissertation, Department of Phys-
how students learn—a mismatch?,’’ ibid. 61, 295–298 共1993兲. ics, University of Washington, 1999 and L. G. Ortiz, ‘‘Identifying and
15
Compressibility had not been an issue in our investigation of student un- Addressing Student Difficulties with Rotational Dynamics,’’ Ph.D. disser-
derstanding of hydrostatic pressure. In that study, students attributed pres- tation, Department of Physics, University of Washington, 2001.
26
sure gradients to a number of different variables. Density, however, was On a similar question asked on an end-of-high-school examination in Aus-
not one of them. tralia, Gunstone reported that about 45% of the students answered cor-
16
The prevalence of the descending line response was essentially the same rectly, with about 35% giving the answer that the blocks would return to
as on versions in which no free-body diagrams were required. It seems that the same level 共see Ref. 24兲.
27
additional questions did not help students arrive at correct answers. In some cases, students were instructed to draw free-body diagrams for
17
After collecting the students’ responses to the five-block problem, the in- each block. The results were apparently unaffected.
28
structor in one section of the introductory course conducted a class discus- In addition to the references already cited, see L. C. McDermott, ‘‘Re-
sion in which he explained the correct answer and discussed the common search on conceptual understanding in mechanics,’’ Phys. Today 37,
incorrect descending line answer. This discussion may have had an effect 24 –32 共1984兲; D. Hestenes, M. Wells, and G. Swackhamer, ‘‘Force con-
on the students’ subsequent ability to answer other questions. cept inventory,’’ Phys. Teach. 30, 141–158 共1992兲; R. R. Hake,
18
For other examples in which students apply formulae without understand- ‘‘Interactive-engagement versus traditional methods: A six-thousand-
ing the terms, see B. S. Ambrose, P. S. Shaffer, R. N. Steinberg, and L. C. student survey of mechanics test data for introductory physics courses,’’
McDermott, ‘‘An investigation of student understanding of single-slit dif- Am. J. Phys. 66, 64 –74 共1998兲. Besides this representative sample, an
fraction and double-slit interference,’’ Am. J. Phys. 67, 146 –155 共1999兲; extensive bibliography can be found in L. C. McDermott and E. F. Redish,
L. C. McDermott and P. S. Shaffer, ‘‘Research as a guide for curriculum ‘‘Resource letter: PER-1: Physics education research,’’ ibid. 67, 755–767
development: An example from introductory electricity, Part I: Investiga- 共1999兲. The emphasis in this resource letter is on university students and
tion of student understanding,’’ ibid. 60, 994 –1003 共1992兲; 61, 81共E兲 on references readily available to physicists.
共1993兲. 29
See, for example, L. Viennot, ‘‘Spontaneous reasoning in elementary dy-
19
In this case, the students had also performed a laboratory experiment on namics,’’ Eur. J. Sci. Educ. 1, 205–221 共1979兲; A. Champagne, L. Klopfer,
buoyancy in which they measured the buoyant force on the same block and J. Anderson, ‘‘Factors influencing the learning of classical mechan-
suspended in two different liquids. ics,’’ Am. J. Phys. 48, 1074 –1079 共1980兲 and J. Clement, ‘‘Students’ pre-
20
See, for example, R. A. Lawson and L. C. McDermott, ‘‘Student under- conceptions in introductory mechanics,’’ ibid. 50, 66 –71 共1982兲.

1187 Am. J. Phys., Vol. 71, No. 11, November 2003 Loverude, Kautz, and Heron 1187
This article is copyrighted as indicated in the article. Reuse of AAPT content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
158.49.113.199 On: Mon, 13 Oct 2014 16:04:42

You might also like