Elementary Students' Learning of Materials Science Practices Through Instruction Based On Engineering Design Tasks

J Sci Educ Technol (2010) 19:580601
DOI 10.1007/s10956-010-9225-8
Elementary Students Learning of Materials Science Practices

Through Instruction Based on Engineering Design Tasks
Kristen Bethke Wendell Hee-Sun Lee
Published online: 14 May 2010

Springer Science+Business Media, LLC 2010
Abstract Materials science, which entails the practices of

selecting, testing, and characterizing materials, is an
important discipline within the study of matter. This paper
examines how third grade students materials science performance changes over the course of instruction based on
an engineering design challenge. We conducted a case
study of nine students who participated in engineering
design-based science instruction with the goal of constructing a stable, quiet, thermally comfortable model
house. The learning outcome of materials science practices
was assessed by clinical interviews conducted before and
after the instruction, and the learning process was assessed
by students workbooks completed during the instruction.
The interviews included two materials selection tasks for
designing a sturdy stepstool and an insulated pet habitat.
Results indicate that: (1) students significantly improved on
both materials selection tasks, (2) their gains were significantly positively associated with the degree of completion
of their workbooks, and (3) students who were highly
engaged with the workbooks reflective record-keeping
tasks showed the greatest improvement on the interviews.
These findings suggest the important role workbooks can
play in facilitating elementary students learning of science
through authentic activity such as engineering design.
K. B. Wendell (&)
Department of Education and Center for Engineering Education
and Outreach, Tufts University, Curtis Hall Lower Level,
474 Boston Ave., Medford, MA 02155, USA
e-mail: [email protected]
H.-S. Lee
Department of Education, Tufts University, Medford, MA, USA
123
Keywords Design-based science Elementary science

Design decisions Science workbooks
Clinical interviews Properties of materials
Objects, materials, and their physical properties are one

of the school science topics with which students have
the most firsthand, everyday experience (Spelke 1991).
Beginning in infancy, children continually attempt to make
sense of the objects and materials in the world around them
(Piaget 1930). By the time children reach the primary
grades, they already have a rich set of conceptual resources
and concrete experiences with which to reason about
objects, materials, and matter (Duschl et al. 2007).
Within the physical sciences, the study of matter constitutes a major conceptual field (American Association for
the Advancement of Science 1993; National Research
Council [NRC] 1996; Smith et al. 2006). For professional
scientists and engineers, the practice of specifying objects
and materials by measurable physical properties has led to
the creation of the discipline known as materials science
(Callister 2007). Expert materials scientists have deep
knowledge about the structure and composition of natural
and synthetic materials, and they are skilled at conducting
tests of materials, selecting appropriate materials, and
describing materials invariant, intensive properties. This
testing and characterization of materials is an important
subspecialty within the scientific study of matter, and
materials science may also play an important role in young
students physical science learning.
However, our knowledge of how to help students
develop materials science proficiency is limited. Many
studies on students conceptions of matter take the form of
clinical interviews in which they sort objects or material
samples into groups of similar samples (e.g., Dickinson
1987; Inhelder and Piaget 1964; Krnel et al. 1998).

Researchers often probe for specific modes of classification, such as sorting by constituent material, state of matter,
intensive properties like density or texture, or extensive
properties like weight or shape (e.g., Krnel et al. 2003).
They also explore how children differentiate object from
material kind (e.g., Dickinson 1987; Smith et al. 1985).
Although these studies have generated much knowledge
about how children think about objects and materials, they
have not addressed whether or how children apply that
thinking to select objects and materials for practical tasks.
For example, if a child can sort materials by their hardness,
does that mean she can select an appropriate material for a
table top? Research on students developing understanding
of material properties through the use of materials, rather
than on the classification and study of them, has been
limited. Existing research studies also provided limited
analysis of student work related to materials science during
instruction. Research that measures performance at only
one or two points in time is unable to capture fully the
instructional dynamic (Ball and Forzani 2007) of a
learning experience, which consists of constant interactions
between teacher, students, and curriculum materials.
The purpose of this article is to report on our investigation of what third-grade students learn about the
practices of materials science through an engineeringdesign-based curriculum module. Engineering design can
provide an authentic activity context for science learning;
authentic activities are those that require students to apply
scientific knowledge and reasoning to situations similar to
those they will encounter in the world outside the classroom, as well as to situations that approximate how scientists do their work (NRC 1996, p. 75). For our research,
we developed a materials science curriculum module
where elementary students learn to apply knowledge about
the properties of materials and objects to the engineering
design problem of constructing a stable, quiet, thermally
comfortable model house. The curriculum module uses
LEGOTM construction elements and electronic sensors as
tools for engineering design tasks.
In designing our study, we were guided by three
research questions: (a) What approaches do students use in
selecting, testing, and describing the materials for practical
engineering tasks before instruction? (b) How do their
approaches change after instruction? and (c) How can the
changes be traced to students engagement with the related
learning tasks during the instruction? To answer these
questions, we conducted a case study of nine-third-grade
students who represented varying degrees of academic
abilities and engagement with the learning tasks offered
during instruction. The first and second research questions
were answered using clinical interviews conducted before
and after the instruction, and the third research question
581
was answered using students workbooks completed during

instruction. The clinical interviews were situated in two
engineering design tasks: choosing the materials for a
sturdy stepstool and a warm pet habitat. In the workbooks,
we analyzed student work related to strength and thermal
insulation, which were critical concepts in the clinical
interviews.
In the first section of this paper, we characterize the
discipline of materials science and review previous
research related to childrens materials science learning.
Next, we outline the instructional intervention used in this
study and describe our research design, including subjects,
data sources, and data collection and analysis procedures.
We then present and discuss results by comparing them to
previous work and summarizing their contributions and
limitations.
Literature Review
Materials Science: The Study and Application
of Material and Object Properties
The major practical skills of materials scientists include
analyzing existing materials for their constituent parts and
structure, specifying materials by intensive properties that
are invariant to amount, identifying the properties needed
for particular tasks, and selecting existing materials and
creating new materials with desired properties (Ashby et al.
2007; Dowling 1998). The primary content knowledge of
materials scientists deals with how particular properties
relate to particular technological challenges, how materials
are composed and structured, and how macroscopic properties are determined by a materials structure and composition (Allen and Thomas 1999; Schaffer et al. 1999).
Among the properties considered most important by
materials scientists are the mechanical qualities of compressive strength, elastic modulus, and density, and the
thermal qualities of heat capacity, thermal conductivity,
and expansion coefficient. These properties are crucial to
the success of any physical technology that is subject to
mechanical and thermal stress (Ashby and Jones 2005).
The National Science Education Standards (NRC 1996)
state that by the upper elementary grades, students should
be able to describe and measure the observable properties
of materials and objects. In their proposed learning progression, Smith et al. (2006) similarly assert that by the end
of fifth grade, most students should be able to measure,
classify, and describe materials according to their properties, which include directly observable properties like color
and hardness as well as less obvious properties like density,
flammability, and conductivity. Smith et al. do not include
choosing materials and justifying material choices in their
123
582
Table 1 Key practices and content knowledge of expert and introductory materials science
Key scientific practices
Key scientific
content knowledge
Expert materials science
Materials science suitable for elementary school

students
a) Analyzing existing materials for their constituent

parts and physical properties
a) Testing materials for particular properties
b) Identifying the properties needed for particular

tasks and devices
b) Identifying the material properties important to a

given task
c) Selecting materials with those properties
c) Selecting materials with those properties
d) Uniquely specifying materials by intensive properties

that are invariant to the amount of material
d) Describing materials specifically via intensive and

extensive properties
e) Creating new materials with desired properties
e) Proposing combinations of materials that would

accomplish a task
a) How particular properties relate to particular

technological challenges
a) A few of the properties important for very common

mechanical tasks
b) How materials are composed and structured
b) How objects are different from their constituent

materials
c) How macroscopic properties generally relate to

materials structure and composition
d) How specific macroscopic properties are particularly
determined by a materials structure and composition
learning progression, but the NSES do elaborate that upper

elementary students should be able to choose suitable
materials for a particular task and justify their choice of
materials. Incorporating the NSES (1996) and Smith et al.
(2006), we summarzie our own ideal, full set of introductory materials science learning goals for elementary school
students in Table 1. Our set of materials science learning
goals is more comprehensive than existing lists of properties of materials standards because we have added the
goals of recognizing important properties for practical
tasks and articulating the reasons behind their importance.
and Samarapungavan 1999) such as wood is harder than

water, wood is heavier than water, or wire is denser
than wood. Students were more likely to explain fluidity
and malleability by discussing the samples hardness,
weight, strength, or density, than by proposing explanations related to material composition or particulate structure. Very few students used mechanistic explanations such
as, water is free to move because its molecules are spread
out, or wire is more flexible than wood because it has a
different atomic structure (Nakhleh and Samarapungavan
1999).
Previous Findings on Students Materials Science

Learning
Strength
Although previous studies of student achievement related

to materials science have not focused on selecting and
testing materials for everyday tasks, they have identified
numerous challenges involved in developing a scientific
perspective on the properties of materials and objects. Here
we describe the conceptual and practical challenges relevant to this study.
Explaining Properties
Elementary and middle school students often have difficulty providing a mechanistic cause for a macroscopic
property (Lee et al. 1993; Nakhleh and Samarapungavan
1999; Parkinson 2001; Snir et al. 2003). In one study, when
asked to explain waters fluidity and wires malleability
(compared to woods fluidity and malleability), most students simply invoked macroscopic characteristics (Nakhleh
123
Studies of elementary grade students ideas about how to

construct stable towers and bridges suggest that many
students model strength as directly varying with weight
(Gustafson et al. 1999; Parkinson 2001). For example, in
one study (Gustafson et al. 1999), most students proposals
for designing a tower that would not tip over related to
the weight of the tower materials. Some students suggested
adding weight to its base, support posts, or top, while
others proposed increasing the weight of the entire tower.
This inclination to associate weight with design improvement may be related to childrens tendency to attend to the
property of weight moreso than to other properties (Krnel
et al. 1998; Smith et al. 1985).
A related investigation of students ideas about how to
predict the strength of bridge designs found that students
need explicit instruction on strategies for comparing the
strength of multiple products (Gustafson et al. 2000). In
that study, some students suggested simply that each
designed bridge should be built and tested. Other students

described a specific action, such as adding weights or
shaking the bridges, but did not discuss how to compare
one bridge to another. A third group of students suggested
comparing the two bridges but did not specify any conditions to be held constant for both bridges. Very few
responses were both explicit about equivalent conditions
and specific about how to quantify the bridges strength,
such as counting how many weights could be added before
collapse. These findings suggest the need to distinguish
between two different barriers to student achievement in
materials science: limited experience with experimental
design (e.g., control of variables) and limited understanding of the physical phenomena related to particular material
properties.
583
and cotton) are good at keeping things hot (Lewis and Linn
1994).
Though the concepts of thermal conductivity and
strength are challenging for elementary students, our
decision to include these properties in this study was purposeful. If we task students with testing and selecting only
for simple perceptual properties like weight and color, we
cannot make strong arguments about their learning of
materials science practices. There are many correct commonsense ways to test and select for simple properties. By
asking students instead to test and select for complex
properties, we can differentiate more precisely among
levels of performance at these practices. A more novel
discovery can be made by asking students to test and select
for non-perceptually accessible properties such as strength
and insulation.
Thermal Insulation
One of the major conceptual challenges related to thermal
phenomena is the distinction between the concepts of
temperature and heat. The scientific understandings are
that heat is exchanged energy, and temperature is average
molecular kinetic energy (Wiser and Amin 2001). Many
preschool and early primary grade students, however,
conceive of heat as a gas or vapor (Paik et al. 2007), and
they tend to conflate heat with its effects, such as burning
and melting (Shayer and Wylam 1981). Most upper elementary students conceive of heat as perceived hotness,
so their concept of heat remains undifferentiated from
their concept of temperature (Wiser and Amin 2001). One
way to explore students understandings of heat and
temperature is to ask them to discuss the perceived temperature of different objects within an environment of
known ambient temperature (e.g., Clark 2006; Lewis and
Linn 1994; Paik et al. 2007). A scientifically normative
response would explain that all the objects are at the same
temperature, but they feel different to your hand because
they conduct heat at different rates. Childrens responses,
in contrast, typically indicate that the temperature of a
substance depends on its constituent material, size, or
thickness, and they do not differentiate between a substances speed of heat transfer and its temperature (Clark
2006; Erickson 1979).
A second major conceptual challenge related to thermal
phenomena is the equivalence between insulation that
maintains cold temperatures and insulation that maintains
hot temperatures. Children instead believe that materials
that can insulate hotness are fundamentally different
from materials that can insulate coldness (Lewis and
Linn 1994; Paik et al. 2007). They tend to think that
materials that feel cool to the touch at room temperature
(such as metal and glass) are good at keeping objects cold,
while materials that feel warm to the touch (such as wool
Curriculum Context
Based on a situative, social constructivist framework
(Driver et al. 1994; Greeno 1998), we characterize science
learning as both social enculturation into practices and
personal construction of ideas. Consistent with this
framework, we suggest that students can effectively learn
about object and material properties while engaged in
solving authentic problems through actual materials science practices. This involves writing with, talking with,
and physically using the cultural tools and symbolic
resources of materials science.
Engineering design is one kind of authentic activity that
requires the use of both practices and content knowledge
related to materials science. We consider design as the
activity of creating or proposing plans for a product that
will solve an open-ended and ill-structured problem (Dym
1994; Simon 1996). Engineering design is a particular
domain of design. For this study, we define engineering
design as the organized development and testing, through
the use of mathematical and scientific knowledge and
models, of artifacts that perform a desired function without
violating known constraints (Davis and Gibbin 2002; Dym
and Little 2004).
We developed a curriculum module called Design a
Model House: The Properties of Objects and Materials for
upper elementary grade students. The module situates
materials science learning within the activity of engineering design. It is an example of a design-based science
curriculum module (Fortus et al. 2005; Kolodner 2006;
Roth 1996). In design-based science, students are typically
challenged to design and construct a physical prototype
that meets a set of requirements while adhering to constraints. The design challenge is carefully structured so that
both science knowledge and practices are important to
123
584
success. Students conduct scientific investigations to deepen their understanding of the design problems requirements or constraints, and as they generate and implement
solutions to the design problem, they increase their
understanding of science concepts and practices. Our
approach to incorporating engineering design problems
into primary-level science is similar to the Learning by
DesignTM approach to middle-school science instruction
(Kolodner 2006).
In the opening lesson of the Design a Model House
module, students learn that their engineering design challenge is to create a miniature model house that is stable,
soundproof, and thermally insulated. Over the next six
lessons, guided by their teacher and their science workbooks (called Engineers Journals), the students conduct a
series of engineering tests to identify materials to meet
these design requirements. They use weights, sound sensors, and temperature sensors for their tests and craft
materials and LEGOTM construction elements for their
house surfaces and frames. As the students test materials
and begin prototyping, they are asked to make scientific
arguments about the best materials for each portion of the
house. They are encouraged to consider the material and
object properties of stability, strength, soundproofing,
reflectivity, and thermal insulation. To facilitate their
efforts throughout the module, science workbooks are used
to prompt students to reflect on experiments and observations. In the modules two concluding lessons, the students
employ their new understanding of object and material
properties to complete the design and building of their
miniature model houses. The nine-lesson module is intended for 8- to 9-year-old students and requires approximately 12 h of instructional time.
Multiple solutions exist to the model house design
challenge. Several different combinations of materials
enable the house to meet the criteria of being stable (it does
not collapse under pressure from a students palm), quiet (a
sound sensor in the house measures fewer decibels than
one adjacent to the outer wall), and thermally insulated (a
temperature sensor inside the house indicates slower temperature change than one exposed directly to a hot or cold
stimulus). Students are told that they will be evaluated on
how well their material choices can be justified with evidence from their investigations, rather than on the correctness of their choices.
The primary instructional goal of this curriculum module is to enable students to carry out three main introductory materials science practices: (1) describing objects and
materials by their extensive and intensive properties, (2)
conducting tests to determine whether materials exhibit
desired properties, and (3) selecting materials with the
properties that are most important for accomplishing a
specific design task. Therefore, we focus on students
123
learning of materials science through the practices of

describing, testing, and selecting. These encompass the first
four introductory materials science practices listed in
Table 1. The practices from Table 1 of identifying
important material properties and selecting materials with
those properties are now collapsed into a single practice
termed selecting. Therefore selecting includes both
choosing a material and justifying that choice with scientific knowledge.
Throughout the module, the following resources are
used to engage students in the practices of materials
science:
a)
physical tests demonstrated by the teacher and conducted by the students (compressive test of strength,
tap test for stability, insulation test with heat stimulus
and temperature sensor, sound absorption test with
decibel sensor),
b) graphical representations in the students science
workbooks (charts for recording test results, designated areas for sketching plans to use materials),
c) scientific language spoken by the teacher and written
in the workbook (material, property, test, pass, fail,
result, select, strong, stable, insulating, absorbing), and
d) physical processes of using materials to build and rebuild components of the model houses, as demonstrated by the teacher and conducted by the students.
We consider these resources to be the main affordances
of this module (Greeno 1998) because they enable children
to pick up the discourse, practices, and representations
the cultural toolsof materials science. Thus, for this
study, we hypothesized that in students post-interviews,
they would more frequently refer to intensive physical
properties to explain their selection of materials, and more
frequently propose material tests based on quantified
measures and fair comparisons. We also expected that
students performances on the interview design scenarios
would be traced to their engagement with the relevant
design tasks in the workbooks throughout the curriculum
module.
Previous teaching sequences related to materials science
have focused on determining and explaining material
properties (e.g., National Science Resources Center 2005;
Smith 2007), witnessing and explaining material transformations (e.g., Acher et al. 2007; Johnson 2000), and conducting scientific argumentation about material properties
and changes (e.g., Education Development Center 2003;
Lawrence Hall of Science 2005). These existing interventions do not emphasize the use of material samples in the
design, construction, and testing of artifacts that serve
authentic purposes. By including an engineering design
component, our instructional approach expands upon previously implemented approaches.
Method
We carried out an exploratory, descriptive case study of a
relatively small number of students. Our case was comprised of a group of nine students of varying academic
abilities from one third-grade classroom. In investigating
our case, we utilized both quantitative and qualitative
evidence (Yin 2003). A mixed-evidence case study
approach was chosen for several reasons. For one, the
phenomenon of interest, students science learning within
a single classroom, is best explored via multiple sources
of evidence, and this is a hallmark of case study methodology. Second, the phenomenon was naturally bounded and thus appropriately explored through case study
research (Merriam 1998). Indications of the boundedness of the phenomenon were the finite number of
students that could be sampled (the nine consenting students) and the fixed duration of the instruction (the nine
lessons in the curriculum). Third, case study methods are
particularly suitable for capturing process and development over time (Yin 2003), and our research questions
called for attending to changes in students science
practices over time. Quantifying some of the qualitative
data, such as workbook responses, helped to characterize
major patterns of change as well as identify students for
in-depth descriptions of learning. Qualitative evidence,
such as interview transcripts, highlighted the differences
among individual students learning using their own
accounts.
Setting and Participants
This case study took place in an urban K-8 public school in
the northeastern United States. The student body in the
school is 15% African American, 10% Asian, 40% Latino,
33% White, and 2% multiracial non-Latino. Of the student
population, 79% are eligible for free or reduced-price
lunch, and 60% learned English as a second language. The
third-grade classroom in which this study was conducted
was led by a teacher who volunteered to attend a summer
training workshop and implement the Design a Model
House science curriculum module. In the workshop, she
completed every writing prompt, drawing task, and designand-build challenge that students are asked to complete in
the module. She had no prior experience teaching engineering or using LEGO in her classroom, but she had
previously taught about material properties during a unit on
rocks and minerals. This teacher was the primary implementer of the curriculum. The first author of this paper was
present for half of the instructional sessions, provided
technical assistance with the LEGO electronics and construction materials, and conducted the interviews with
students.
585
The teacher initiated each lesson by explaining the

model house criteria to be investigated that day, and then
students wrote and drew in their workbooks for 5 min to
respond to an introductory exploration prompt related to
those criteria. Students then worked in pairs to complete
the days model house investigation. The teacher concluded each lesson by guiding the students through
reflective prompts (Davis and Linn 2000) in their
workbooks.
Eighteen students participated in the curriculum module
over a period of 4 weeks. Among them, two boys and
seven girls returned permission forms. These nine students
were included in this study. As shown in Table 2, these
participants represented a broad range of academic abilities. All nine students were present for all nine lessons of
the module.
Data Sources
To track students changes related to selecting, testing, and
describing materials, we conducted clinical interviews and
collected student workbooks. The interviews were conducted before and after the instruction, and the workbooks
were embedded assessments used during the instruction.
Clinical Interviews
The semi-structured clinical interview (Ginsburg 1997;
Piaget 1929) consisted of two material selection design
tasks. To begin each task, the interviewer stated a design
goal, displayed a design schematic, and gave the student
two material samples. The interviewer then asked the student to select, justify the selection of, propose testing
techniques for, and describe the material that would best
accomplish the goal. For the first design task, students
selected either bamboo pole or polyvinyl chloride (PVC)
plastic pipe for the legs of a stepstool that wouldnt collapse. This sturdy stepstool task elicited their ideas about
the property of strength in addition to probing for their
abilities to select, test, and describe materials. For the
second task, students selected either aluminum sheeting or
foam sheeting for the walls of a pet habitat that would
stay warm. This warm habitat task was designed to focus
students thinking on the property of insulation as they
again were asked to select, test, and describe a material.
Both strength and insulation were properties also addressed
during instruction; they were required characteristics of the
model houses.
Each interview lasted approximately 20 min. The preinterviews took place the week before the students began
working on the curriculum module. The post-interviews
took place over the two-week period after they completed
their model houses. The pre- and post-interview protocols
123
586
Table 2 Profiles of the student
participants
Name
Sex
Typical written science

test achievement
Currently designated
as english language learner?
Ava
Female
Middle (third quartile)
No
Brandon
Male
Middle (third quartile)
No
Chinelle
Female
Low (fourth quartile)
No
Elisa
Female
No
Julia
Female
Yes
Katie
Female
Middle (second quartile)
No
Macon
Male
Middle (second quartile)
No
Stella
Female
No
Viola
Female
High (first quartile)
No
were identical. A summary of the protocol is presented in

the Appendix.
Workbooks
Student workbooks were analyzed to investigate how the
changes in interview performance connected to students
engagement with materials science practices throughout
the curriculum module. The paper and pencil workbooks
were structured to provide students with introductory
questions, investigation instructions, data recording
prompts, and reflective prompts for each of the nine
instructional activities of the module. The 21-page workbooks included six pages with writing and drawing tasks
related to the property of strength, five pages with tasks
related to the property of insulation, and three pages with
tasks related to both strength and insulation.
Data Analysis
Interview Scoring
Interviews were transcribed in their entirety. Each transcript was then separated into six segments: the select, test,
and describe sub-tasks within the sturdy stepstool design
task, and the select, test, and describe sub-tasks within the
warm habitat design task. Each of the six segments was
assigned a set of descriptive codes (Glaser and Strauss
1967; Miles and Huberman 1994) as well as a numerical
score. The coding details are discussed below. Overall,
Table 3 presents how the coding and scoring scheme was
designed to probe high and low levels of student ability
related to materials science practices. All interview
responses were coded and scored by two independent raters. The interrater reliability for the numerical scores was
0.87 (Pearson correlation coefficient). We created a coding
scheme that emphasizes just three aspect of the students
responses: their selection rationales, proposed tests, and
123
descriptions of materials. Students did not earn or lose

points for choosing one or the other material.
For the select sub-tasks, students stated their rationales
for choosing a certain material. Each rationale was coded
for three factors: (a) relevance to design goal (Fortus et al.
2005), (b) accuracy for selected material sample (Minogue
2008), and (c) type, including intensive property, extensive
property (Smith 2007), material kind (Dickinson 1987), or
irrelevant to design goal. Responses earned 2 points for
each rationale that considered a relevant property with
scientifically accurate justification, such as I chose the
PVC because hard plastic is stronger than hollow wood.
Responses earned 1 point for each rationale that considered
a relevant property but used scientifically inaccurate justification, such as, I picked the hollow bamboo because it
is harder to push down than the PVC. (This rationale is
scientifically inaccurate because bamboo is more compressible than PVC. However, compressibility is a relevant
property to consider in designing a stepstool). By assigning
partial points to rationales built on consideration of relevant properties, we were able to credit a students ability to
note the properties most important for a task even when she
could not accurately identify a material characterized by
those properties. Zero points were earned by responses in
which all rationales were both irrelevant and scientifically
inaccurate.
For the test sub-tasks, students proposed techniques for
testing a given property (strength for the stepstool, insulation for the habitat). Each proposed technique was coded
for three factors: (a) relevancy to design goal (Fortus et al.
2005), (b) productivity for determining property (Smith
et al. 2006), and c) requirement for full prototyping, such
as building an entire stepstool or pet habitat (Dym and
Little 2004). Responses earned 2 points for each relevant,
productive test that did not require prototyping, such as I
would put a heavy weight on the PVC pipe and see if it
broke. Responses earned 1 point for each suggestion to
build and test a full prototype, such as, I would get more
587
Table 3 Construct analysis with examples of introductory materials science practices
Science Practices Related to Material and Object Properties
STUDENTS:
The student has
Highly proficient
abilities related to
material and object
properties
SELECTING:
To select a material for a
specific task, the
response
TESTING:
To discuss how to test for
a material property, the
response
DESCRIBING:
To describe a specific
material sample, the
response
Names an appropriate
material and provides a
rationale that is both
relevant to the task and
accurate for the material.
Proposes and justifies an

effective method that
does not require full
prototyping.
Accurately uses
multiple intensive and
extensive properties.
e.g., I choose foam for

the bath toy because it
has very low density and
is buoyant in water.
e.g., That cube of wood

is lightweight and
about 2 inches long on
each side. It is brown
and buoyant in water.
Proposes and justifies an

effective method, but it
requires full prototyping.
Primarily describes
surface features and
extensive properties.
e.g., Test for the strength

of that plastic by building
a bed out of it and seeing
if the bed collapses.
e.g., That cube of wood

has rough spots on it. It
could give splinters. It
feels heavy.
Provides no information
relevant to selecting
materials for the task.
Provides no information
relevant to measuring the
property.
Provides no
information relevant to
the material.
e.g., I would choose

wood for the bath toy
because its the best
choice.
e.g., I know that plastic

has a lot of strength.
e.g., I like how that

cube of wood feels.
Names a material and

provides a rationale that
is either relevant to the
task or accurate for the
material, but not both.
e.g., I choose clay for
the bath toy because
clay can float.
Very limited
abilities related to
material and object
properties
e.g., Test for the strength

of that plastic by
measuring its
compression distance
under a 1-kg weight.
PVC, build a stepstool, and stand on it. Responses earned

0 points if all proposed tests were both irrelevant to the
goal and unproductive for determining the property.
For the describe sub-tasks, students described the material
samples in as much detail as possible. Each descriptor was
coded as intensive property, extensive property (Smith
2007), surface appearance (Smith et al. 1985), material kind
(Dickinson 1987), function (Dickinson 1987), simile or
metaphor (e.g., it looks like a log), or inaccurate. Responses
earned 1 point for the first intensive property descriptor and 1
point for the first extensive property descriptor. A score of 0
points represents a description that included neither intensive
nor extensive properties.
We incorporated science content knowledge into the
materials science practices coding. First, for a selection
rationale to be coded as scientifically accurate, the student
needed to have at least some content understanding of the

material that she chose. Second, for a material test to be
coded as productive, the student needed to have understanding of the property for which he was proposing a test.
After coding and scoring each individual students
responses, we computed summary statistics for the entire
participant group, and we tabulated the frequency with which
each code was applied. We also computed the total number
of distinct (i.e., unique) ideas elicited from the group of
students for each selecting, testing, and describing sub-task.
Counting the number of unique ideas elicited by a task is
important because it indicates the generativity of the task as
well as the expertise of the students. Students are not always
able to generate relevant ideas in novel situations, and the
ability to generate multiple relevant ideas is an important
characteristic of experts (Linn and Eylon 2006).
123
588
Workbook Scoring
Each workbook page on strength or insulation was assigned
an activity completeness score that indicated how fully the
student attempted the tasks (Baxter et al. 2000). A response
earned 1 point if between one-third and two-thirds of the
page had been completed, and 2 points if two-thirds or
more of the page had been completed.
Each workbook page was also scored for performance
on the science practice it featured. As shown in the
sample scored workbook page in Fig. 1, each page consisted of multiple prompts for written or drawn responses.
If the student accurately selected, tested, or described
materials in all of the responses on a page, that page was
rated as showing full accuracy. If only some of the
responses showed accurate work, the page was rated as
showing partial accuracy. For each workbook page, a
specific scoring rubric described what could be considered partial or full accuracy at its key science practice.
For example, for the page shown in Fig. 1, the key science practice is selecting materials, and full accuracy
requires choosing column materials that are consistent
Fig. 1 Sample scoring of a

workbook page featuring the
property of strength and the
science practice of selecting
materials
123
with the recorded data as well as referencing strength in

the rationale for the choice. Two raters scored each
workbook response for both completeness and science
practice. Interrater reliability for the workbook scores was
0.86.
Results
The case study of nine students is reported in three steps.
First, we provide a summary of students responses to the
stepstool and habitat scenarios. Second, we utilize quantitative evidence to highlight the changes in interview performance from preinstruction to postinstruction. Finally,
we describe connections between students interview gains
and their engagement with workbook tasks during the
module. Based on the patterns observed across all nine
students, we use four students to illustrate how students
materials science practices changed and how these changes
can be traced to workbook efforts. When presenting these
four prototypical students, we identify the other students
who showed similar characteristics.
589
Selecting, Testing, and Describing Materials for Sturdy

Stepstools and Warm Habitats
Justifying Material Selections
As shown in Table 4, in selecting materials, students
responded with fewer relevant and accurate rationales for
their material selection for the warm habitat task than for
the sturdy stepstool task. However, both before and after
instruction, the students offered a greater total number of
selection rationales for habitat materials than for stepstool
Table 4 Material selection

rationales offered by all students
materials. This result means that while students were able

to engage in the practice of selecting materials for both
design tasks, they had greater conceptual understanding of
the phenonomena related to sturdiness than to temperature
maintenance. Though they offered an abundance of rationales for their habitat material choices (a sign of science
practice ability), the rationales for their stepstool material
choices were more relevant and accurate (a sign of conceptual understanding).
For example, Brandon stated only two reasons for
choosing PVC for the stepstool, but both reasons were
Rationales for material selection
List of all rationales
No. of distinct
rationales
Percentage relevant
and accurate
Relevant
and accurate
Irrelevant
or inaccurate
3/8 (38%)
Stronga
Heavy
Sturdy
Stable
Holds a lot of weight
Metal
Stepstool task
Pre-interview
Wooden
Paintable
Post-interview
12
5/12 (42%)
Strong
Heavy
Sturdy
Stable
Holds a lot of weight
Plastic
Hard
Splinter-free
Tear-resistant
Layered
Attachable
Thick
Habitat task
Pre-interview
10
2/10 (20%)
Retains heat
Blocks coldness
Heavy
Strong
Unbreakable
Hard
Thick
Cut-able
Lightweight
Feels warm
Post-interview
15
2/15 (13%)
Retains heat
Heavy
Proven as insulator
Strong
Unbreakable
Hard
Thick
Rationale statements made by

multiple students are counted
and listed only once.
Synonymous statements have
been collapsed into one entry;
e.g., the statements
unbreakable, not
breakable, and cant be
broken easily are collapsed
into the entry unbreakable.
Cut-able
Feels warm
Heats up quickly
Shiny
Metal
Soft
Melt-proof
Sound-proof
123
590
directly related to the goal of a collapse-proof stepstool:

PVC is strong and sturdy. In contrast, when Brandon
chose metal for the pet habitat, he gave three reasons, and
none of them was directly related to the goal of an insulated habitat wall. He explained that he chose metal
because it is hard, heavy, and the mouse cant bite
through it. Although Brandons rationales for his chosen
habitat material were not relevant to the explicit design
goal of insulation, we do note that his response shows
awareness of other important design features.
Proposing Material Tests
Table 5 illustrates that the students more frequently proposed scientifically productive tests for determining strength
than they did for determining insulation. However, both preand postinstruction, the students offered a greater number of
distinct proposals for testing insulation than they did for
testing strength. Taken together, these results indicate that
students conceptual understanding of strength enabled them
to conduct the practice of testing materials more productively for the stepstool task than for the habitat task.
For example, Julia suggested three appropriate, productive tests for determining PVCs strength: use a hammer to
bang it and see if it doesnt break, put a rock on it and see if
it breaks, and try dropping it on the floor. In contrast,
within the habitat task, Julias list of tests was more extensive, but fewer of the tests were relevant and productive to
testing for insulation. She said that to test for insulation, she
would touch [the metal] and see if it feels hot, use a rock
or a hammer, put something heavy in the middle and see if
it breaks, try to bend it, build a house and try it, and
make a little door to let air in and out.
This difference in performance between the stepstool
and habitat tasks also suggests that the design goal of an
insulated pet habitat was not well understood by all the
students; it was not a goal that fit easily into their conceptions of habitats. Even when the interviewer moved on
from the term insulation and reiterated the design goal as
choosing a wall material that would keep a warm temperature inside the habitat, the students chose to select
materials for and propose tests for other attributes of pet
habitats. In contrast, for the stepstool task, they remained
focused on the attribute of strength. The goal of a sturdy
stepstool made sense to the students, but the notion that
insulation (i.e., maintenance of a particular temperature) is
an important design goal did not resonate with their conceptions of adequate pet habitats. Thus, to complete the
habitat task, they drew from a variety of their personal,
everyday experiences with pets. In contrast, to complete
the stepstool task, they drew more from their school-science experiences with strength and stability and generated
a narrower range of responses.
123
To explore the results of the selecting and testing subtasks further, we examined students postinstruction
responses to interview questions about the definitions of
strength and insulation. In the post-instruction interview,
students were asked what it meant for something to be
strong and what it meant for something to be insulating,
before they were asked to propose tests for those properties. Students who were not able to define a term were
provided with its definition before being prompted with the
next interview question.
At postinstruction, all nine students produced accurate
definitions of strong. However, only two students,
Chinelle and Macon, produced an accurate definition of
insulating, which they described as keeping something
at the temperature you want. Despite being able to define
insulation, these two students showed below-average performance on selecting and testing materials for the habitat
in their postinstruction interviews. In contrast, five of the
students who could not accurately define the term insulating showed above-average performance on selecting and
testing materials for the habitat (after they were reminded
of the definition of insulating). The remaining two students,
Stella and Brian, could neither define insulating nor competently perform the materials science practices for the
habitat task, but they excelled at the materials science
practices when asked to apply them to the sturdy stepstool
task. These data highlight the distinction between having
vocabulary knowledge and applying that knowledge to
practical tasks. Although vocabulary knowledge is no
doubt crucial to long-term science achievement, in this
study it was neither necessary nor sufficient for students
success with materials science practices. In other words,
the data show no consistent relationship between vocabulary knowledge and science practice abilities.
However, what did emerge from the data was the
importance of specific conceptual understanding for students materials science success. Students materials science practices did not generalize across material
properties. For example, although Stella and Brian were
excellent at justifying selections and proposing tests for a
strong material, they were not able to select and test for
an insulating material. Whether or not they had mastered
the vocabulary term insulating, students were only able
to select and test for insulating materials if they had some
understanding of thermal phenomena. More important
than the definition of insulation were awareness of the
situations when it is important and of the variables that
allow it to be observed. Although the sophistication of the
term insulating contributed to the difficulty of the
habitat task, the conceptual challenges related to heat and
temperature played a larger role in limiting students
success in selecting and testing materials for an insulated
habitat.
591
Table 5 Material property tests proposed by the overall group of students

Tests of material properties
List of all proposed tests
Number
of distinct
ideas
Percentage
relevant
and productive
Relevant and productive
Irrelevant or unproductive
(7/8) 88%
Build prototype and step on ita
Weigh on scale
Stepstool task
Pre-interview
Build prototype and put rocks on it

Drop test
Bend until break
Tap until break
Impact with rock
Apply body weight
Post-interview
12
(10/12) 83%
Build prototype and step on it
Weigh on scale
Build prototype and do

compression test
Drop test
Measure thickness
Squeeze test
Impact with rock
Apply body weight
Apply test weight
Impact with hammer
Impact with golf ball
Flick test
Habitat task
Pre-interview
11
(2/11) 18%
Build and measure air temp over

time
Expose to heat and measure temp

[looking for warmth]
Build prototype and test with

mouse
Build and check for breathing

holes
Measure height
Measure thickness
Measure length
Weigh on scale
Check for melting
Impact with hammer
Try to poke holes
Post-interview
18
(4/18) 22%
Build and measure air temp over

time
Feel for warmth
Build and test with mouse
Expose to heat and measure temp

[looking for warmth]
Expose to light and measure temp

on each side over time
Put on overhead projector
Form into cylinder and expose to

light and measure temp
Build and check for breathing

holes
Weigh on scale
Try to break
Drop test
Impact with hammer
Apply test weight
Expose to light and check for light
transmission
Push test
123
592
Table 5 continued
Tests of material properties
List of all proposed tests
Number
of distinct
ideas
Relevant and productive
Percentage
relevant
and productive
Irrelevant or unproductive
Expose to light and feel for

warmth
Add door to house
Add tape to house
a
Tests proposed by multiple students are counted and listed only once. Synonymous tests have been collapsed into one entry; e.g., the proposals
to add a test weight, put weight on top, and see if it can hold a weight, are collapsed into the entry apply test weight.
Describing Material Properties

In describing materials for both the stepstool and habitat
tasks, students expressed accurate and prolific ideas. Before
instruction, 29% out of a total of 116 distinct material
descriptors were intensive properties, like weak, white,
and hard, while 31% were extensive properties, like
skinny, rectangular, and heavy. Only two of the
descriptors were inaccurate: Katie described the foam
sheeting by saying warm air goes through it, and it wont
burn down. After instruction, 25% of the 114 distinct
descriptors were intensive properties, and 30% were extensive properties. Only one of the descriptors was inaccurate: a
description of metal sheeting as being easy to melt.
Changes in Materials Science Practices from
Preinstruction to Postinstruction
For each interview performance, we computed three types
of scores: a total score, a score for each of the two design
tasks, and a score for each of the three materials science
practices of selecting, testing, and describing. For all six of
these measures, the postinstruction means were higher than
the preinstruction means (Table 6). However, out of the
three individual science practices, only the pre-post
increase for the testing sub-task was statistically significant
(p \ .05, 2-tailed Wilcoxon test). This means that after

instruction, the students improved most in proposing
methods to test for material properties. The students
exhibited greater understanding of the use of engineering
tests to determine a materials appropriateness for a design
goal, and this resulted in overall improvement on the
stepstool and habitat tasks.
Connecting Pre-Post Changes to Science Workbook
Performance
Workbook scores varied greatly across the nine students
and across tasks. As shown in Fig. 2, students interview
scores were positively correlated with their workbook
scores. Greater gains on the individual stepstool and habitat
interview tasks, respectively, were associated with higher
completion of and accuracy on the corresponding strength
and insulation workbook tasks (q = 0.478, p \ 0.05).
Throughout the instructional module, all students conducted the same number of engineering tests on house
materials and constructed the same number of house
components. Thus, they did the same amount of handson work. The difference among students was the amount
of information they recorded in the workbooks. Those who
recorded more in their workbooks learned more. When we
analyzed the students workbook performances using only
Table 6 Average scores on design-task interviews (n = 9)

Measure
Pre-instruction
mean (SD)
Post-instruction
mean (SD)
Average paired gain,

pre-post (SD)
Effect sizea,
cohens d
Total interview (out of possible 36)
8.8 (3.4)
13.8 (3.9)
5.0* (3.7)
1.34
Stepstool task (out of 18)
6.2 (2.4)
10.0 (3.8)
3.8* (3.6)
1.27
Habitat task (out of 18)
2.6 (1.5)
3.8 (1.8)
1.2* (1.1)
1.27
Select sub-tasks (out of 16)
2.7 (2.8)
4.3 (2.3)
1.7 (3.0)
0.56
Test sub-tasks (out of 16)
2.6 (1.4)
5.6 (2.4)
3.0* (1.7)
1.73
Describe sub-tasks (out of 4)
3.6 (0.7)
3.9 (0.3)
.33 (.87)
0.38
Effect size is calculated according to the Cohens d statistic, where d = (average of paired differences)/(standard deviation of paired
differences)
* p \ .05; 2-tailed Wilcoxon test
123
593
lowest total post-interview score. Both before and after

instruction, his responses comprised a narrow set of ideas
focused on weight. For the sturdy stepstool task, he
selected PVC because it felt heavier. At both interview
times he proposed weighing the PVC on a scale to test for
its strength:
M
I
M
I
M
I
Fig. 2 Students gains on the interview design tasks plotted against
their scores on the corresponding workbook exercises (Spearman
correlation = 0.48, p \ 0.05, n = 18); letters adjacent to the markers
identify individual students
their completeness scores, we found that greater gains on

the stepstool and habitat interview tasks were also significantly correlated with the completeness of the corresponding strength and insulation workbook exercises
(q = 0.473, p \ 0.05). Even if their recordings were
inaccurate, the students who recorded more in their
workbooks improved more on the interviews. This indicates that the importance of both task engagement and
record keeping for design-based science learning. If students ably experiment and design, but do not keep records,
then they gain less from their experimentation and design
activities.
Four Student Cases
We now use the work of four prototypical students to
illustrate qualitatively the nature of the connections
between the interviews and workbooks. Four representative
examples are those of Macon, who improved the least on
both interview tasks, Stella, who improved the most on the
stepstool interview task, Julia, who improved the most on
the habitat interview task, and Chinelle, who received the
highest total workbook score.
M
I
M
I
M
For the warm habitat task, Macon again focused on

weight as a priority in establishing the property of a
material. At both interview times, he chose aluminum
because it felt heavier, and he suggested weighing the
aluminum to determine its insulation ability. To Macon, the
property of weight could be used to determine both
strength and insulation:
M
I
M
I
M
I
Macon: No Improvement on Interview Tasks
Macon is the representative case for students who earned
low workbook scores and showed little or average gain on
the interview. Brian and Elisa also displayed these characteristics. Macons scores did not increase on either the
stepstool task or the habitat task, and he received the
[Chooses PVC pipe for as better material for stepstool

leg.]
Okay. Why do you think that would be the better
material for the stool?
Because its metal [incorrectly identifying PVC
plastic as metal]
Okay, because its metal. Why would that make it
better?
Because metals stronger than wood [3-s pause]
Can you tell me more about that? [5-s pause; M shrugs
shoulders] What, why is metal stronger than wood?
Because its heavier. And heavier stuff is stronger
Okay. What is something you could do to measure
exactly how strong this material is? So that you could
tell You could weigh it?.Um, you can put it on, like,
one of those weigh things that people stand on
And then what would that tell you about how strong it
is?
It would, it would, because heavier stuff makes it
stronger so, if its heavy, if its heavier than this [picks
up bamboo], then it [PVC] would be stronger
M
I
M
I
This one [points to metal tray]

Why do you think this would be the better material for
Cuz its heavy and stronger
And why would heavier and stronger make it better?
Because, if its thicker, then it might be heavy and
strong, and if its heavy and strong it makes stuff
warm
Okay, why do you think that something heavy and
strong makes stuff warm?
Because its thick
And what does the thickness do with the warmth?
It traps it all inside of something
So you picked this metal, the other engineers want
to know exactly how good it is at insulating. It
means how good it is at keeping the temperature the
123
594
M
I
M
same inside the house. How could you measure

how insulating this material is?
You could weigh it on a scale
And if it weighed more, what would that mean about
how insulating it is?
It would mean that it would be warmer
Macons workbook performance aligned with his lack of

improvement on the interview tasks. His workbook score,
18 out of 60, was the lowest of all nine students scores.
Although Macon and his partner did complete all of the
hands-on construction and testing activities with the rest of
their class, Macon responded to very few of the exploration
or follow-up prompts, and his workbook contains very few
complete phrases or drawings. It also contains very little
data from the classs tests of material properties, with one
major exception: Macon recorded all the results from the
sound absorption test and responded to the follow-up
prompt about soundproof materials. As a result, the one
new idea posed by Macon at postinstruction involved
sound absorption. Despite several reminders that the design
goal for the second task was a habitat that would stay
warm, Macon based his initial material selection for that
task on the samples soundproofing qualities. He declared
that the mouse should not be bothered by outside noises,
and foam was the best material because its sound absorption ability had been proved in class.
Stella: Greatest Improvement on Stepstool Task
Stella illustrates the case of showing a very high gain for one
interview task while showing a below-average gain for the
other interview task. Stella is a unique case, as no other
students fit this profile. Of all nine interviewed students,
Stella showed the greatest pre-post improvement on the
stepstool task. Both before and after the module, she
expressed many ideas for both design tasks, but only on the
stepstool task did her responses substantially improve in
their level of sophistication. At preinstruction, Stella chose
bamboo for the stepstool simply because of its material kind:
because its wooden. To test for the bamboos strength,
she proposed standing on it and trying to break it in half. To
describe the bamboo, she offered only four descriptive terms,
and none of them were extensive properties.
I
S
I want to know which you think would be the better

material to use to make the stepstool
This [bamboo] one because it wouldnt break
because this [bamboo] ones wooden This [PVC]
one would break because its not wooden
Yeah, okay, so what could you do to tell the
engineers exactly how strong this [bamboo] is? How
could you measure its strength?
123
S
I
S
I
S
You can stand on it

And what would that tell you about how strong it is?
Uh, um, tell you if its sturdy or not sturdy, and its like,
uh, if its like, uh, not sturdy to stand on it, and it like,
start standing on it and youll fall
Okay, any other tests you could do to find out how
strong it is?
Like if you can break it in half and see if you can stand
on it and it wont break apart
At postinstruction, Stella switched to choosing PVC for

the stepstool, and she based her choice on its intensive
properties: it was strong, sturdy, and hard. To test
for strength, she suggested a number of useful tests,
including hitting it with a golf ball, compressing it, building a prototype stepstool, and once again, standing on the
PVC. She described the bamboo in 10 different ways,
which included both intensive and extensive properties.
S
I
S
I
S
I
S
I
S
This one [PVC]

Okay, why would that be better?
Because, um this [bamboo] one like is wood, and this
[PVC] one um, is like stronger, cuz when you stand on
it, youre um, you wont fall
Why do you think this one [PVC] is stronger?
Because, um, it has, um, good materials, and its
sturdy Like, um, its hard, and um
What could you do to measure exactly how strong
this material is?
You can, um [3-s pause], put it like, um, on something,
and, um make sure it doesnt break with anything
You can try to stand on it
Are there any other tests that you could do with this
material, to see how strong it is?
You can um, make something out of it, and then you
can like um, like do it with, that thing we did for a golf
ball, you can We hit it with a ball
What did you do in that, is there anything you did in
that project where you tested for strength, that you
could use to test for the strength of this?
Um, you can like um, we pressed it down. [Holds PVC
upright, pushes on top] You can um, you can uh [6-s
pause] you can [12-s pause] you can make it out a
stepstool, and then you can try it
Like Macons workbook, Stellas workbook aligned

well with her interview performance. On the tasks related
to strength, she earned the second highest score. With the
exception of the tasks for the second lesson, Stellas
responses to prompts about strength are complete, and they
show proficiency at materials science practices. Her written
and pictorial inscriptions indicate that she was attempting
to connect materials characteristics to their appropriateness for certain tasks. In response to a prompt about useful
materials for support columns, she listed, brick, wood,

metal, gold, stone, silver, and cement, because these
materials are sturdy, strong, stable, able to support
weight.
Stellas much improved stepstool task performance and
her high level of workbook engagement indicate that she
learned a great deal about the property of strength, and that
the opportunity to write and draw about strength in several
different contexts enabled her to construct and explore new
ideas. However, her much more moderate improvement on
the habitat interview task suggests that students learning
about materials science practices does not generalize to all
material properties. Stella was able to apply her classroom
experiences to a novel interview task involving the property of strength, but not to a novel task focused on
insulation.
Julia: Greatest Improvement on Habitat Task
Julia represents the case of improving on both interview
tasks and earning workbook scores that were consistent
with both interview gain scores. Ava and Katie also
showed these characteristics. Julias gain on the habitat
task was the highest among all nine students. She also
improved substantially on the stepstool task, and she was
one of only three students who improved on all three subtasks: selecting, testing, and describing materials. Julia, an
English language learner, shifted from providing just a few
vague ideas in her pre-interview to offering many wellarticulated ideas at post-interview. Before instruction, she
chose metal for the habitat walls because it cant break,
and she proposed testing the metal with a hammer to see if
it was strong. Neither her selection rationale nor her proposed test was relevant to the design goal of insulation.
Further, in describing materials, she generated only three
descriptors for the metal. Below, she explains her choice
for the habitat:
I
J
I
J
I
J
I
J
Why would you pick that [metal] one as the better

material?
Because this [metal] one, it cant break and this [foam]
one is kind, could break, if you let it, do it like this
Mm, this [foam] white one could break
Yes
What if you had to think mostly about the job of
keeping the mouse house warm, would you still pick
this [metal] one, or would you pick this [foam] one?
Still pick this [metal] one
Okay, tell me about that
Because [3-s pause], um, if you like, made it with, you
could like, maybe you could like make a door with it?
At post-interview, Julia still cited metals strength as

part of her selection rationale for the habitat material, but
595
she also introduced new rationales and tests that did relate
to insulation. She reasoned in a relevant (though inaccurate) manner that metal wont let the cold in, and she
suggested testing the metal by touching it to see if it felt hot
and by building a prototype house to test its insulation
abilities. She also described the metal in seven different
ways.
I
J
I
J
I
J
I
J
So why would you pick this one [metal] as the better

material for keeping the temperature warm?
Because, um, when you put it in a house, this one
[metal] will make it hot, and this [foam] will not,
because it will like, cover the hotness and somehow its
gonna go in, the cold
So how could you measure exactly how good this
material [metal] is at keeping the temperature the
same?
Um, I wouldI would like put it on, and touch it, and
if its cold, and maybe when you build it, it will make it
hot because it lets in the
So when you touch it, it feels kind of cold
Yeah
But when you use it to build something
Yeah, inside
It would keep it warm?
Yeah. For the mouse
Like Macon and Stella, Julia engaged with her workbook tasks in a manner that aligned with her pre-post
interview changes. Her workbook earned the second
highest total score and the second highest score for
insulation tasks. In the workbook, Julia engaged in each
lessons beginning exploration question, which gave her
many opportunities to practice reasoning about material
and object selections. For example, as shown in Fig. 3,
she reasoned that a winter jacket keeps you warm because
it is thick and [it] keep[s] your body heated by sorta
stick onto you. Julia also completely and accurately
recorded results in her workbook for both tests related to
insulation (for the model house wall and roof materials).
Her computations of temperature changes were correct,
and her conclusions about the most insulating materials
were consistent with the data. Metal was not a material
used during the instructional module. Although Julias
workbook answers were correct, she was not considering
metal as a potential insulating material when she composed those answers.
Chinelle: Highest Workbook Score
In the previous three examples, each students workbook
scores were aligned with his or her interview performance.
Chinelles example is less straightforward. She earned the
highest total workbook score, 51 out of 60 points. If her
123
596
Fig. 3 Pages from the

workbooks of Macon, Julia, and
Chinelle; left-hand samples are
exploration questions about
insulation from the middle of
the module; right-hand samples
are from the concluding
workbook page
results followed the pattern of Macon and Julia, the quality

of her workbook would predict a great improvement on her
interview responses. But instead, Chinelles 4-point interview gain was below average. Because her pre-interview
score of 6 points was among the lowest, this small pre-post
gain was not due to a ceiling effect. We use Chinelle to
represent the case of earning workbook scores that would
predict much higher interview gains than were actually
achieved. Viola is the other student whose work followed
this pattern.
One potential explanation for Chinelles lack of pre-post
improvement is that she somehow failed to engage in a
critical instructional activity. Indeed, one-third of Chinelles missing workbook points were on the very last
workbook page, which prompted the students to review
123
how they had treated a given property during their model

house design. The prompts were, How did you test for the
property of ____? and Why is the property of ___
important to your house? Each individual student was
assigned a different property by the teacher. Chinelle was
assigned the property of insulation, but as shown in Fig. 3,
her page was only partially complete; she wrote only, The
white is a good insulator. This work gave no indication of
accurate science practices and earned only 1 out of 4
possible points. In contrast, Julia and Stella earned 4 and 3
points, respectively, on their work for this page. Stella
wrote that to test for strength, you have to push down and
see if it pushes, and that strength is important for a
house because you have to keep it up. Julia recalled
that for her reflectivity test, she had got foam, [and]
cardboard of the color black or white, and that reflectivity is important because one needs to let the heat out
not in the house.
Julia used her final workbook page to write about temperature regulation for the model house, and she performed
well on the habitat task in her post-interview. Likewise,
Stella concluded her workbook by writing accurately about
the property of strength, and she expressed sophisticated
ideas about strength in her post-interview. Chinelle, on the
other hand, did not take advantage of the last workbook
task, and her subsequent post-interview performance
showed relatively little improvement from pre-interview.
Macon, similarly, earned 0 points on his last workbook
page, and he showed no improvement between the two
interviews. Besides Macon and Chinelle, one other student
earned less than 2 of the closing activitys points, and his
interview gain was just above the class average. Though
additional investigation is needed, these patterns suggest
that the closing instructional activity, which tasked students
with reflecting on the actions and reasoning that led to their
model house design decisions, may have been a critical
factor in helping students attain their newly learned
materials science practices.
Discussion
This case study focused on third-grade students learning of
materials science practices. During both the curriculum
module and the interviews, students had multiple opportunities to select, test, and describe materials. The model
house, sturdy stepstool, and warm habitat design scenarios provided authentic contexts (Brown et al. 1989; Roth
1996) for materials science practices and for reasoning
about properties of materials and objects. Several intriguing patterns emerged from the workbook performances and
interview responses elicited by these design scenarios.
Third-Graders Approaches to Materials Science
The students relied on perceptually accessible properties
(Smith 2007), including thickness, texture, felt weight,
and felt temperature, as they constructed their rationales
for material selection. They also preferred perceptually
accessible properties as they proposed methods for testing
strength and insulation, two properties that are not immediately observable. This preference for perceptual accessibility is consistent with findings in other studies
(Gustafson et al. 1999; Lewis and Linn 1994), and it can
help to explain why, despite the substantial gains between
pre- and post-interviews, the highest post-interview scores
were still less than 60% of the maximum score. The
properties that the children were asked to consider were
597
complex physical properties determined by underlying

molecular structure, bonding, and manufacturing methods.
Primary grade students lack knowledge about these invisible determinants of complex properties (Nakhleh and
Samarapungavan 1999). As a result, when the students
tried to select and propose tests for materials, they relied on
immediate perception and based their ideas on directly
observable properties, rather than on underlying intensive
properties.
Changes in the Third-Graders Approaches to Materials
Science
Despite their preference for perceptually accessible properties, all the students included at least one intensive
property in their post-interview rationales for material
selection, and all but one student proposed at least one
useful material test for an intensive property. These results
indicate that elementary students can learn to recognize and
test intensive properties, even if they cannot generate
mechanistic explanations for those properties. Smith et al.
(1985) conducted a study in which most 6- to 9-year-olds
simply invoked material kind to explain why a clay ball is
heavier than a wax ball; they said one ball was heavier
because it was made of different stuff than the other ball.
This finding suggests that the students in our study would
simplistically invoke material kind and state that the
bamboo was weaker just because it was made of wood, and
that the aluminum was less insulating just because it was
made of metal. However, after instruction, very few of the
students appealed to material kind to explain why one
sample was weaker or less insulating than the other.
Rather, they made serious attempts to construct a causal
relationship between immediately accessible material and
object properties and the less accessible properties of
strength and insulation.
Another pattern in the data was the emergence of
synthetic models in the postinstruction responses to the
warm habitat task. When instruction causes a student to
find incoherencies in her commonsense theory for making
meaning of something, then she might be inclined to
create what Vosniadou (2002) calls a synthetic model.
This happens when information received through
instruction seems to become assimilated to the initial
explanatory framework creating synthetic or internally
inconsistent models (Vosniadou 2002, p. 62). We identified two examples of synthetic models. The first was
Katies idea that a materials ability to block light transmission determines its ability to block heat transfer. In
both her workbook and postinstruction interview, Katie
suggested that if she held a material up to a light and saw
light passing through, she would know the material was a
poor insulator. This model of equating transparency with
123
598
thermal conductivity could have been based on the curricular activity in which a lamp provides the heat for an
insulation test. Katie possibly viewed the lamps primary
function as a light source rather than a heat source. The
second synthetic model was the idea, put forth by both
Stella and Julia, that air flow into a space is equivalent to
heat input into the space. At postinstruction, both of these
students proposed methods for insuring that adequate air
could enter the pet habitat. Stella wanted to make holes in
the habitat walls, and Julia wanted to create an open door.
They reasoned that if the air could flow into the habitat,
then the habitat would be warm. This model for warmth
maintenance could be based on home experiences of
opening doors and windows to regulate the indoor temperature, and it is consistent with nonnormative ideas
uncovered by previous research (Clark 2006; Erickson
1979; Paik et al. 2007).
Connecting the Third-Graders Interview Approaches
to Their Workbook Performances
When studying an instructional intervention with the aim
of improving student learning outcomes, it is important to
examine the black box, or instructional dynamic (Ball
and Forzani 2007), of the interactions between students,
teacher, and curriculum materials. In this study, we focused
on analyzing the students interactions with the curriculum
materialsthe science workbooks. Overall, the workbooks
served as a tool that enabled the students to track the tests
they conducted on potential materials as well as the
selection decisions they made based on the test results. In
this way, the workbooks may have eased the cognitive
effort required to connect test results back to experimental
procedures, and to link test results forward to decisions
about courses of action (Shepardson and Britsch 2001).
This means that students who take great advantage of
opportunities to record their observations and reasoning
(like Julia and Stella) are likely to show the greatest change
in their approaches to the selection and testing of materials.
This is consistent with the finding by Olson (1996) that
writing about abstract notions, rather than only speaking
about them, has helped humans to advance new ideas and
concepts.
However, because most third-grade students have not
yet developed differentiated concepts of heat and temperature, third-grade student performance on the habitat
interview task might not improve even with tireless effort
on writing and drawing exercises related to insulation.
Even after hours spent writing and drawing about thermally comfortable model houses, there may be a conceptual cap on childrens ability to select, test, and
describe materials in the context of designing an insulated
structure.
123
Moreover, there are also specific kinds of writing and

drawing tasks that may do more to foster the building of
proficiency at materials selection, testing, and description.
The closing workbook activity used in this study prompted
the student to reflect on the scientific reasoning behind
design decisionse.g., to explain how and why particular
properties were tested and particular materials were chosen. This kind of reflection activity appears to be a critical
exercise for learning materials science practices. As shown
in Chinelles example, if students engage in almost all
learning tasks but do not respond to reflective prompts
(Davis and Linn 2000), they may miss a key chance to
consolidate newly constructed ideas with newly learned
practices.
Study Limitations
Several factors limit the generalizability of this studys
findings. First, since the interviews used particular material samples and posed particular design scenarios, they
did not comprise an exhaustive assessment of introductory
materials science. We did not collect information on the
students initial commonsense theories of matter. This
information can help to examine whether a students
existing theory of matter, including the presence or
absence of material/object differentiation, is predictive of
ability to select, test, and describe materials. Second,
although nine students can be an appropriate number for a
descriptive, exploratory case study, the small sample size
prohibited the use of statistical analyses to generalize
from patterns observed in the target student group.
Finally, our analysis of the relationship between workbook and interview performance was intended to explore
the role of instructional experiences in student learning,
but workbook tasks are a single element of the multifaceted instructional environment. Other indicators such
as discourse patterns among students and between students and teacher can be used to further examine the
instructional dynamic of the classroom.
Conclusions
In this paper, we described a case study of elementary
students who developed materials science practices within
a science learning environment focused on the use of
materials, rather than on the classification and study of
them. The results of this study showed that elementary
students are capable of learning the practices of materials
science, even though it is a discipline typically limited
to college-level engineering students. The results also
showed that the act of recording materials science decisions during instruction is associated with improved
ability to make and explain materials science decisions

during interviews. Thus, engagement not just with prototype building and testing, but also with reflecting upon
and inscribing decisions about that prototype, is critical
for learning. Using science workbooks, we were able to
connect individual variations in learning gains directly to
students varying completion of record-keeping tasks.
Finally, this study demonstrated that while design-based
instruction can improve students abilities to make and
test materials selection decisions, students full achievement is hindered by limitations in their conceptual
understanding of particular material properties. Thus,
even when applied to the elementary school level, designbased instruction in the physical sciences needs to pay as
599
much attention to conceptual supports as it does to

practical experiences.
Acknowledgements This study was funded by National Science
Foundation Grant DRL-0633952 but does not necessarily represent the
policies or positions of the NSF. We would like to thank the readers and
reviewers who have offered helpful comments on earlier drafts of this
paper. We also wish to thank the students who took part in the study and
their teacher, who generously devoted time and energy to the project and
enthusiastically taught two new science units in her classroom in 1 year.
Appendix
See Table 7.
Table 7 Summary of the clinical interview protocol

Property
Practice
Scenario and props presented to

students
Examples of additional spoken

prompts
Strength
Selecting materials and

objects for a particular
task
Your job is to pick the material for

this stepstool design. The
stepstool shouldnt collapse
when someone steps on it.
Imagine you have to choose
between these two materials:
Which material would you pick as

the better material for the
stepstool?
(Compressive yield strength)
Why do you think this is better for

the stepstool?
12 length of PVC, 1.5 in

diameter
12 length of bamboo, 1.5 in
diameter
Testing for a particular
property
Insulation
(Inverse of thermal conductivity)
Now the other engineers want to

know exactly how strong these
two materials were. What does it
mean for something to be strong?
(Provide definition if student
does not know.)
Tell me how you would measure

how strong the materials are.
And are there any other tests you
would do to figure out how
strong the materials are?
Describing objects and

materials
Now the other engineers need to

find the material that you picked
in a big supply room.
Describe the strong material to me

with as many details as you can
think of.
Selecting materials and

objects for a particular
task
Now imagine you have a new

engineering job. Your job is to
pick the material for the walls of
this mouse habitat. The walls
should keep the inside of the
habitat warm. Imagine you have
to choose between these two
materials.
Which material do you think you

would pick as the better material
for the mouse house walls?
Why do you think this is better for
the walls?
6 9 9 Styrofoam tray
6 9 9 aluminum tray
Testing for a particular
property
Describing objects and

materials
Now imagine the other engineers

want to know exactly how
insulating these two materials
are. What does it mean for
something to be
insulating?(Provide definition if
student does not know.)
Tell me how you would measure

how insulating the materials are.
Now the other engineers need to

find the material that you picked
in a big supply room.
Describe this material to me with

as many details as you can think
of.
And are there any other tests you

would do to figure out how
insulating the materials are?
123
600
References
Acher A, Arca` M, Sanmart N (2007) Modeling as a teaching learning
process for understanding materials: a case study in primary
education. Sci Educ 91:398418
Allen SM, Thomas EL (1999) The structure of materials. John Wiley,
New York
American Association for the Advancement of Science (1993) Benchmarks for science literacy. Oxford University Press, Inc, New York
Ashby MF, Jones DRH (2005) Engineering materials 1: an introduction to properties, application, and design. Elsevier, Amsterdam
Ashby M, Shercliff H, Cebon D (2007) Materials: engineering,
science, processing, and design. Butterworth-Heinneman,
Burlington
Ball DL, Forzani FM (2007) What makes education research
educational. Educ Res 36(9):529540
Baxter G, Bass K, Glaser R (2000) An analysis of notebook writing in
elementary science classrooms. CSE education technical report
533. University of California, Los Angeles
Brown JS, Collins A, Duguid P (1989) Situated cognition and the
culture of learning. Educ Res 18(1):3242
Callister WD Jr (2007) Materials science and engineering: an
introduction, 7th edn. John Wiley, New York
Clark DB (2006) Longitudinal conceptual change in students
understanding of thermal equilibrium: an examination of the
process of conceptual restructuring. Cognit Instr 24(4):467563
Davis LA, Gibbin RD (eds) (2002) Raising public awareness of
engineering. National Academies Press, Washington
Davis EA, Linn MC (2000) Scaffolding students knowledge integration: prompts for reflection in KIE. Int J Sci Educ 22:819837
Dickinson DK (1987) The development of a concept of material kind.
Sci Educ 71(4):615628
Dowling NE (1998) Mechanical behavior of materials: engineering
methods for deformation, fracture, and fatigue, 2nd edn.
Prentice-Hall, Inc., Upper Saddle River
Driver R, Asoko H, Leach J, Mortimer E, Scott P (1994) Constructing
scientific knowledge in the classroom. Educ Res 23(5):512
Duschl RA, Schweingruber HA, Shouse AW (eds) (2007) Taking
science to school: learning and teaching science in grades K-8.
National Academies Press, Washington
Dym CL (1994) Engineering: a synthesis of views. Cambridge
University Press, New York
Dym CL, Little P (2004) Engineering design: a project-based
introduction. John Wiley, New York
Education Development Center (2003) Insights: mysterious powder
teachers guide, 2nd edn. Kendall/Hunt Publishing Company,
Dubuque
Erickson GL (1979) Childrens conceptions of heat and temperature.
Sci Educ 63(3):221230
Fortus D, Krajcik JS, Dershimer RC, Marx RW, Mamlok-Naaman R
(2005) Design-based science and real-world problem-solving. Int
J Sci Educ 27(7):855879
Ginsburg HP (1997) Entering the childs mind: the clinical interview
in psychological research and practice. Cambridge University
Press, NY, pp 115158
Glaser BG, Strauss AL (1967) The discovery of grounded theory:
strategies for qualitative research. Aldine de Gruyter, New York
Greeno JG (1998) The situativity of knowing, learning, and research.
Am Psychol 53(1):526
Gustafson BJ, Rowell PM, Rose DP (1999) Elementary childrens
conceptions of structural stability: a three year study. J Technol
Educ 11(1):2744
Gustafson BJ, Rowell PM, Guilbert SM (2000) Elementary childrens
awareness of strategies for testing structural strength: a three
year study. J Technol Educ 11(2):522
123

Inhelder B, Piaget J (1964) The early growth of logic in the child,
classification and seriation. Routledge & Kegan Paul, London
Johnson P (2000) Childrens understanding of substances, part 1:
recognising chemical change. Int J Sci Educ 22:719737
Kolodner JL (2006) Case-based reasoning. In: Sawyer KL (ed) The
Cambridge handbook of the learning sciences. Cambridge
University Press, Cambridge, pp 225242
Krnel D, Watson R, Glazar SS (1998) Survey of research related to
the development of the concept of matter. Int J Sci Educ
20(3):257289
Krnel D, Glazar SS, Watson R (2003) The development of the
concept of matter: a cross-age study of how children classify
materials. Sci Educ 87:621639
Lawrence Hall of Science (2005). Chemical interactions teachers
guide (Full Option Science System). Delta Education
Lee O, Eichinger DC, Anderson CW, Berkheimer GD (1993)
Changing middle school students conceptions of matter and
molecules. J Res Sci Teach 30(3):249270
Lewis EL, Linn MC (1994) Heat energy and temperature concepts of
adolescents, adults, and experts: implications for curricular
improvements. J Res Sci Teach 31(6):657678
Linn MC, Eylon B-S (2006) Science education: integrating views of
learning and instruction. In: Alexander PA, Winne PH (eds)
Handbook of educational psychology. Lawrence Erlbaum Associates, Mahwah, pp 511544
Merriam SB (1998) Qualitative research and case study applications
in education. Josey-Bass Publishers, San Francisco
Miles MB, Huberman AM (1994) Qualitative data analysis, 2nd edn.
SAGE, Thousand Oaks
Minogue J (2008) A lemon of a lesson. Science and Children
42(1):2527
Nakhleh MB, Samarapungavan A (1999) Elementary school childrens beliefs about matter. J Res Sci Teach 36:777805
National Research Council/NRC (1996) National science education
standards. National Academy Press, Washington
National Science Resources Center (2005) Solids and liquids teachers
guide. Science and Technology for Children, Kendall/Hunt
Publishing
Olson DR (1996) The world on paper: the conceptual and cognitive
implications of writing and reading. Cambridge University Press,
Cambridge
Paik S-H, Cho B-K, Go Y-M (2007) Korean 4- to 11-year-old student
conceptions of heat and temperature. J Res Sci Teach 44(2):
284302
Parkinson E (2001) Teacher knowledge and understanding of design
and technology for children in the 311 age group: a study
focusing on aspects of structures. J Technol Educ 13(1):4458
Piaget J (1929) The childs conception of the world. Harcourt, Brace
Jovanovich, NY
Piaget J (1930) The childs conception of physical causality. Kegan
Paul, London
Roth W-M (1996) Art and artifact of childrens designing: a situated
cognition perspective. J Learn Sci 5(2):129166
Schaffer JP, Saxena A, Antolovich SD, Sanders TH, Warner SB
(1999) The science and design of engineering materials, 2nd edn.
WCB/McGraw-Hill, Boston
Shayer M, Wylam H (1981) The development of the concepts of heat
and temperature in 1013 year-olds. J Res Sci Teach 18(5):
419434
Shepardson DP, Britsch SJ (2001) The role of childrens journals in
elementary school science activities. J Res Sci Teach 38(1):4369
Simon HA (1996) The sciences of the artificial, 2nd edn. MIT Press,
Cambridge
Smith CL (2007) Bootstrapping processes in the development of
students commonsense matter theories: using analogical

mappings, thought experiments, and learning to measure to
promote conceptual restructuring. Cognit Instr 25(4):337398
Smith C, Carey S, Wiser M (1985) On differentiation: a case study of
the development of the concept of size, weight, and density.
Cognition 21:177237
Smith C, Wiser M, Anderson C, Krajcik J (2006) Implications of
research on childrens learning for standards and assessment: a
proposed learning progression for matter and atomic-molecular
theory. Measurement 14(12):198
Snir J, Smith C, Raz G (2003) Linking phenomena with competing
underlying models: a software tool for introducing students to
the particulate model of matter. Sci Educ 87:794830
601
Spelke ES (1991) Physical knowledge in infancy: reflections on
Piagets theory. In: Gelman R, Carey S (eds) The epigenesis of
mind. Essays on biology and cognition. Lawrence Erlbaum,
Hillsdale, pp 133169
Vosniadou S (2002) On the nature of nave physics. In: Limon M,
Mason L (eds) Reconsidering conceptual change: issues in theory
and practice. Kluwer Academic Publishers, Dordrecht, pp 6176
Wiser M, Amin T (2001) Is heat hot? Inducing conceptual change
by integrating everyday and scientific perspectives on thermal
phenomena. Learn Instr 11:331355
Yin RK (2003) Applications of case study research, 2nd edn. Sage,
Thousand Oaks
123

Elementary Students' Learning of Materials Science Practices Through Instruction Based On Engineering Design Tasks

Uploaded by

Copyright:

Available Formats

Elementary Students' Learning of Materials Science Practices Through Instruction Based On Engineering Design Tasks

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Elementary Students' Learning of Materials Science Practices Through Instruction Based On Engineering Design Tasks

Uploaded by

Copyright:

Available Formats

J Sci Educ Technol (2010) 19:580601