Promoting Advanced Writing Skills in an Upper-Level Engineering Class

Promoting Advanced Writing Skills
in an Upper-Level Engineering Class
B. YALVAC
for all disciplines. The traditional approach of instructing students
Learning Sciences Program
how to produce written text (e.g., transcribing language in written
Northwestern University
form, learning spelling and grammatical conventions, learning the
principles of good style by examining exemplary models) was re-
H.D. SMITH
placed by an approach that included a variety of cognitive activities
Department of Psychology
such as planning, translating, and reviewing written text to satisfy
Northwestern University
communication needs such as purpose and audience [5]. WAC
programs grew out of pioneering efforts at such schools as Carleton
J.B. TROY
College, Beaver College (now Arcadia University), Michigan
Department of Biomedical Engineering
Technological University, and George Mason University. Support-
Northwestern University
ed by faculty workshops and grounded in learning science theory,
WAC is embraced now by schools at every level. Writing intensive
P. HIRSCH
courses may be found in such disparate areas as history, sociology,
The Writing Program
economics, biology, and even mathematics. WAC and WID, for
Northwestern University
Writing within the Disciplines, have become familiar terms, and
books about what constitutes good writing as a professional in spe-
cific disciplines have proliferated [6–9]. Thus, faculty members at
ABSTRACT
many schools, even at research universities where WAC programs
have not always been supported institutionally [10], have benefited
This paper summarizes the design and evaluation of an instruc-
from WAC practices and now consider it their responsibility to
tional approach aimed at improving the writing skills of a group
help their students improve at writing.
of undergraduate engineering students. We sought to determine
Engineering is certainly one field where proficiency in written
whether student performance in difficult writing skills such as
communication is valued [11]. ABET EC2000 mandates that mas-
argumentation and synthesis could be improved by integrating a
tery of skills such as written communication, teamwork, and design
single writing exercise into an upper level engineering course. In
be acquired progressively throughout the undergraduate curriculum
designing the exercise, we relied heavily on recommendations for
[12]. Since 1996, the engineering school at Northwestern University
best practices from the learning science community, specifically
has integrated writing instruction thoroughly into a required fresh-
those codified in the National Academy text How People Learn
man design course taught jointly by writing instructors and engi-
[1]. We found reliable improvement in student performance in
neering faculty [13, 14]. In recent years, written communication has
many of the areas targeted, demonstrating that the approach
also been included vertically in Northwestern’s design curriculum in
taken was effective. Since we modified the exercise a few times
the Institute in Design and Engineering Applications [15]. At Rice
before meeting our objectives for student learning, we could
University, the Cain Project in Engineering and Professional Com-
compare the effectiveness of different implementations of our
munication offers communication instruction integrated into the
approach. Our success and failures provide guidance for others
curriculum through courses, workshops for faculty and students,
seeking to improve the competence of engineering undergradu-
online resources, and other support [16]. At the Massachusetts
ates in writing.
Institute of Technology and a related international consortium,
communication resources are offered through their CDIO
Keywords: How People Learn, role play, writing
(Conceiving–Designing–Implementing–Operationalizing) initia-
tive [17]. The NSF-funded VaNTH (Vanderbilt-Northwestern-
Texas-Harvard/MIT) Engineering Research Center in
I. INTRODUCTION : APPROACHES TO WRITING
Bioengineering Educational Technology has also developed work-
THROUGHOUT THE CURRICULUM
shops and a taxonomy of professional skills designed to help
engineering educators clarify learning objectives as they design inte-
Since its inception in the 1960s and 70s, the Writing Across the
grated communication assignments and work across departments to
Curriculum (WAC) movement has promoted the idea that writing
coordinate integrated instruction [18]. Many similar efforts are
should be taught as a mode of learning and not merely a means to
underway elsewhere.
remediate deficiencies in writing skill. Before, instructors in English
Much of the research in engineering writing has focused on the
assumed responsibility for student writing, generally through
pedagogical approach taken. Thompson et al. [19] studied inter-
freshman composition courses [2–4]. However, as educators came
active communication among learners, peers, and non-peers and
to pay more attention to the relationship between language and
encouraged students to take responsibility for their own learning,
learning, writing gained acceptance as an integral part of learning
finding that when undergraduates talked to other people about
April 2007
Journal of Engineering Education
117

---

their research, they improved their skill at problem-solving as well as
undergraduate program would experience continuity in training in
their ability in written, oral, and graphical communication. Jensen and
written communication that is in line with EC2000 recommenda-
Fisher [20] studied how the evaluations that students received en-
tions, while still focusing on learning crucial content material.
hanced their writing: comparing the effectiveness of reviews received
Teaching and learning in the course takes place in three one-hour
from peers and those received from teaching assistants and instructors
lectures and one one-hour discussion session per week, along with
on laboratory reports, they determined that peer-evaluated students
three laboratory exercises and five homework assignments. The quar-
performed better. Some research has also focused on sociological
ter-length (11 weeks) course is given once per year. Approximately
factors that influence effective writing in engineering. Two ethno-
half of the eleven weeks are devoted to discussing the organization of
graphic longitudinal studies led Winsor [21, 22] to conclude that, for
the nervous system at the systems level. Emphasis is on the so-
effective communication, students must learn about the rhetorical
matosensory system, including pain, the visual system, and the motor
nature of engineering writing, particularly how power relationships
system. The other half of the course is devoted to discussing the struc-
influence and are influenced by writing in hierarchical organizations.
ture of the neuron, its interaction with glial cells, its membrane prop-
In spite of the movement to incorporate student writing contex-
erties, including the action potential, passive and active propagation
tually in courses within one’s discipline and the research supporting
of neural signals, and synaptic processes. The required textbook is
this endeavor (succinctly summarized by Bazerman et al. [4]), faculty
Purves et al., Neuroscience [26], used in conjunction with the simula-
are often on their own when it comes to designing integrated writing
tion software package Moore and Stuart, Neurons in Action [27].
instruction. Even those faculty members who appreciate the merit of
Since 1999 students in the course have been asked to write team-
offering writing in content areas may be reluctant to add writing ex-
based research papers that force them to explore the literature in
ercises because of the additional work such exercises entail. Many
neural systems that go beyond the textbook. Emphasis is placed on
faculty members are also concerned about what they might have to
helping students work collaboratively within an authentic context.
remove from a course to make room for a writing exercise. In
The research topics selected (one or two per year) were ones where
addition, some engineering faculty may be unconvinced about the
biomedical engineers have made significant contributions to neuro-
effectiveness of integrating writing instruction into their classes since
science. One topic fit well with the half of the course devoted to the
most assessment of learning in WAC programs and of the programs
cellular properties of neurons. The topic was the currently unre-
themselves has been qualitative rather than quantitative [4]. In fact,
solved question of how neurons encode information. Students
many factors contribute to faculty resistance to incorporating writing
wrote papers on this topic every year. The other topic fit well with
exercises into content courses for a discipline [10, 23].
the half of the course devoted to neural systems. In different years,
In this paper, we describe the results of one attempt to improve
this paper concerned either the restoration of neural function by bi-
student writing in a junior and senior level biomedical engineering
ological or synthetic replacement (1999, 2003), the representation
course in neural systems physiology. The strategy we employed was
of images by the visual system (2000, 2004), or the control of loco-
to assign one or two team-based research papers per year in the
motion by neural circuits (2000). Students were given the criteria on
course. Although the assignments typified a WAC approach and
which their written work would be graded in advance, and one or
accepted WAC assumptions about the usefulness of teaching
two class sessions were devoted to writing instruction. The writing
writing within disciplines, we found it necessary to optimize the in-
exercises were done in groups of three to six students.
structional environment before the learning objectives of the
All instructional modules developed by VaNTH, whether for
instructor could be met. It is likely that this experience is a common
mastery of domain skills or core competencies, aim to help students
one faced in instructional design and thus of general interest. Here
understand the deeper concepts that underlie skills development,
we summarize the characteristics of our study participants, the de-
thus facilitating their ability to transfer skills and knowledge from
sign process and reliability issues of our study instrument, and the
one problem to another and to more comfortably handle ambiguity,
analysis. After presenting the findings, we discuss the study impli-
solve novel problems, and communicate at a high level. One design
cations followed by the lessons learned from the 2005 intervention.
criterion is organizing learning exercises around challenging topics
with open-ended solutions like those described earlier. In develop-
ing the communications module reported in this paper, the instruc-
II. THE STUDY CONTEXT: GENERAL COURSE
tor also used the VaNTH taxonomy of core competency skills in
DESCRIPTION AND PREVIOUSLY REPORTED FINDINGS
writing [28] to define five specific learning objectives for accom-
plishing the goal of helping students improve their capability in
As reported in Troy et al. [24], writing assignments were added
written technical communication:
to a junior/senior course in neural systems physiology that was
●
writing concisely
being redesigned as part of the VaNTH effort to improve biomed-
●
using figures, tables and equations, along with text, to explain
ical engineering instruction in physiology [25]. Additionally the
ideas (in other words, realizing that engineering communica-
aim was to fill a gap in the biomedical engineering curriculum at
tion is multifaceted)
Northwestern University. As previously noted, all engineering stu-
●
synthesizing ideas from multiple research papers
dents at Northwestern receive substantial instruction in written
●
using headings and so forth to add structure to reports
technical communication during their first year. Biomedical engi-
●
citing others’ work appropriately
neering majors also receive significant instruction in engineering
The instructor chose these five objectives in 2003 because an analy-
writing as seniors in required courses. However, exposure to writing
sis of student papers from earlier runs of the writing exercise
as engineers between freshman and senior years was less controlled.
(1999–2002) revealed student deficiencies in these areas.
Adding the writing assignments to a core course in systems
In the 2003 offering of the course, students wrote two research
physiology ensured that all students in the biomedical engineering
papers, so we assessed the change in student writing from papers
118
Journal of Engineering Education
April 2007

---

one to two. After the course concluded, the student papers were
see the importance of synthesizing information from different
evaluated by trained, independent coders using a rubric that had
sources in order to develop an argument. We believed we could im-
been developed to assess the learning objectives noted above. For
prove student learning in the areas where we had failed previously
the most part, student writing was assessed with regard to two over-
by making two key changes. We would make the instruction more
arching dimensions, which we called lower level skills, such as gram-
learner-centered and more community-centered by (a) recasting
mar and mechanics, and higher level skills, such as synthesis and
the challenge as a well structured role-play and (b) replacing one of
argumentation. The rubric used to score the papers (Appendix A)
the two class sessions devoted to writing with a more student-cen-
broke these broad dimensions into sub-categories that were easy to
tered event.
define and measure. We found that students showed more im-
provement over time in the lower level skills [24]. Average gains
A. Re-Casting the Challenge as a Structured Role-Play
were typically greater on grammar and mechanics, style, and orga-
Our first change in the writing intervention was to modify the
nization, despite the fact that students started the term with greater
challenge so that students could more easily imagine themselves in a
mastery of them. While these are necessary skills for proficient
professional role. Although the assigned research topics were based
writing, an additional and more important goal for us was to im-
on authentic, unsolved problems in the field (how information is
prove students’ higher level writing skills; that is, to help students
encoded in the brain, whether neural function can be recovered
develop the ability to synthesize research findings in a meaningful
biologically or artificially, etc.), we believed that the students were
way, represent data effectively in graphs and tables, and craft a co-
unable to imagine themselves as researchers who cared about these
herent, persuasive argument. These skills are precisely what profes-
problems because the material was difficult to master and the re-
sional engineers need in order to write effectively in academia and
searchers who confront these issues are much more experienced
industry.
than juniors and seniors in college. This speaks to the problem of
engagement for students, who tend to be influenced in their writing
by the classroom setting and the knowledge that writing assign-
III. GOAL OF THE CURRENT STUDY
ments are part of the course grade. Herrington [29] found that
AND REVISED INSTRUCTIONAL APPROACH
learning to write in chemical engineering courses was difficult for
students because they had to learn not only the conceptual activities
Writing instruction in the 2005 offering of neural systems
that dominate the discipline but also the roles of the writer and au-
physiology was designed to address the shortfall between learning
dience and the purpose of the written document. In a school setting,
objectives and outcomes discovered in our analysis of the 2003 pa-
students know they are writing for the instructor, and if the material
pers; the goal was specifically to help students improve in writing
is difficult for them to master, they tend to see their research papers
areas like synthesis and argumentation. Since the course was being
as just another assignment to complete for a grade. In our assign-
re-designed under the auspices of VaNTH, which encourages the
ment, the complexities of the neural coding issue contributed to the
use of a How People Learn (HPL) pedagogical approach [1], a re-
students’ difficulty in synthesizing information from different
newed effort was made to apply HPL techniques more deliberately
sources. Thus, the assignment was frustrating for students, and not
to the writing interventions.
sufficiently learner-centered for undergraduates.
To design an instructional approach that would more specifically
Herrington proposed structuring classroom settings to help
target the higher level writing skills desired by the instructor, our fac-
enact a professional context. Following this line of thinking, we re-
ulty team returned to the core principles that underlie all VaNTH
cast the challenge as a scenario in which the students were to see
instructional modules. As mentioned previously, all of these mod-
themselves as entry-level graduate students—a more attainable
ules, or segments of courses, are designed around challenges intend-
goal—working in one of two biomedical research labs. In the sce-
ed to engage the students in an authentic disciplinary activity, and
nario, each lab espouses one of the two major theories about neural
all modules should reflect the following four key dimensions [1].
coding: (1) that neurons encode information in the rate at which
●
Knowledge-centeredness: to help students (a) learn with under-
they discharge action potentials and (2) that neurons encode infor-
standing by organizing knowledge around key concepts and
mation in their pattern of action potentials discharged. We placed
(b) move from novice to expert problem-solving methods.
one-half of the student groups randomly in a lab that ascribes to the
●
Learner-centeredness: to (a) take into account the knowledge,
former theory and the other half in one that ascribes to the latter
skills, preconceptions, and learning styles of all students and
theory. In both cases, we gave the students a more easily imaginable
(b) start with what students know when they enter the class.
and streamlined task: we asked them to imagine that they were
●
Community-centeredness: to (a) encourage students and facul-
helping the lab director prepare for a speech at a prestigious confer-
ty to learn from one another and (b) situate learning within
ence. The two labs view each other as competitors and thus each
real-world (authentic) challenges.
director is keen to be well versed in the arguments that the other
●
Assessment-centeredness: to (a) provide frequent opportunities
might present. See Appendix B for the entire scenario and related
for students to make their thinking visible and receive appro-
assignment.
priate feedback and (b) revise teaching and learning activities
We hypothesized that students who could see themselves as a
after measuring student learning.
young member of a research team in someone’s lab would write
Analysis of the assignments and student papers from 2003
more effectively because they would better understand their role.
within this context suggested that even though the instructor had
Moreover, this role dictated the side of the issue they should argue
designed the challenges around authentic issues in the field and had
in the debate. Thus, the role-play made argumentation more promi-
tried to help students see how he, as an expert, would approach the
nent in the exercise at the same time that it made the writing assign-
problem, the students did not sufficiently embrace the challenge or
ment more learner-centered and more community-centered. All of
April 2007
Journal of Engineering Education
119

---

these changes, we hoped, would achieve the main goal of HPL
●
“It would be better to combine (synthesize) the results from
interventions: to help students learn with greater understanding.
several experiments.”
●
“You need more than one experiment to support this idea.”
B. Using a More Student-Centered Event to Teach
●
“Where did you learn all of this?”
Argumentation and Synthesis
After the presentations, students asked the teaching assistant spe-
In 2003 and 2004, two class sessions were devoted to writing;
cific questions about argumentation and documentation:
however, as we reviewed the course from a new perspective, we
●
“Should we mention the positive and the negative sides from
decided these classes were not sufficiently learner- or community-
all perspectives?”
centered. In one class, the instructor showed students models of
●
“Do we use N-citation or another type?”
professional articles that were effectively organized, well argued,
●
“Do we need to include everything about the experiment in
and properly documented. This was a knowledge-centered class in
the paper or just the conclusion or results?”
which the instructor, using himself as a model of the expert writer,
●
“Do we have to say Temporal Coding is completely
showed students how he would approach the research and begin to
wrong?”
develop a draft. We decided to retain this class session in 2005,
●
“What if I list four experiments for Rate Coding and four
which uses a cognitive apprenticeship approach to teaching [30], but
experiments for Temporal Coding? Is this a good format?”
to select readings for the class session that would more dramatically
Students often responded to one another’s questions and engaged
illustrate stronger and weaker examples of argumentation.
in a dialogue where they identified strategies to better synthesize the
In the second session from 2003 and 2004, students were shown
literature and write a convincing argument. In those discussions,
examples of information that was synthesized in professional arti-
each student had the chance to talk about his or her specific needs
cles. Synthesizing literature was explained to students by explicating
and ask questions that we as educators might not have thought
its difference from analyzing the literature. The concepts were pre-
about. For example, one asked whether there is a limit to the number
sented in mini-lectures, followed by time for students to discuss
of experiments a good argument may involve. For questions like
models and their own paper planning in small groups.
these, students received feedback from the teaching assistant and
While the class sessions were meant to be discussion sessions
from their peers. They left the session looking more confident about
rather than lectures, students rarely asked or responded to questions.
the writing assignment.
The small group discussions also lacked engagement. As a follow-up
For a few discussion points, no consensus was reached, e.g.,
to these class sessions, students were asked to edit drafts of their peers’
whether to summarize all the evidence from all points of view or
papers and to get feedback on their drafts from the university’s writing
whether to hide some of the opposing evidence while highlighting
center. However, few groups took advantage of these opportunities,
all the supporting evidence. After students discussed this issue, the
even though a reward was offered in the form of bonus points.
teaching assistant commented that talking about the opposing
In 2005, since the instructor did not believe he could devote
evidence may help convince the reader that the writer is knowledge-
additional class time to writing instruction, we needed to make this
able about the subject, which in turn helps to promote the validity of
second class session much more interactive. We decided to struc-
an argument.
ture it as a mini-presentation in which two randomly selected
In summary, linking the class format to the scenario and having
groups (one from each side of the debate) would present their ar-
students present their ideas orally in the group setting made the as-
guments to the rest of the class. Following their presentations, the
signment more learner-centered and more community-centered.
teaching assistant who moderated the debate would hold a ques-
tion and answer session. This event was scheduled to take place a
week before the final papers were due so that students would be
IV. METHODS OF INQUIRY
motivated to attend, ask questions, and revise their drafts following
the session. The instructor did not attend in case his presence
To assess the effectiveness of the new, more deliberately HPL
would be an inhibiting factor. However, beforehand, the teaching
interventions, we analyzed and compared the group papers written
assistant was made fully aware of the learning objectives of the ex-
in 2004 and 2005. The first year reflected what might be consid-
ercise, particularly the importance of argumentation and synthesis.
ered the less HPL-inspired approach to teaching writing, similar
We hoped that this structure would encourage more peer-to-peer
to what was described in Troy et al. [24]. The 2005 effort reflected a
learning.
targeted attempt to improve the teaching of higher level skills and
One researcher attended the class sessions to observe. Below is a
was based upon what appeared to be lacking after assessment of the
summary of this debate and discussion session, showing much
2003 data.
greater involvement on the part of the students.
Presenters discussed their groups’ main arguments, supporting
A. Participants
evidence, and counter evidence. After each presentation, the
As mentioned earlier, study participants were the undergradu-
teaching assistant posed questions to the group members and
ates enrolled in an upper level Biomedical Engineering course at
commented on the arguments. Typical questions and comments
Northwestern University in 2004 and 2005. For each year, there
included the following:
were 10 groups of students and thus ten papers. Each group com-
●
“What is your main point in this paper?”
prised three to five students. All students were asked to consent to
●
“Is this [paragraph/section] one example of your point or
participate in this study. Because a few individuals did not provide
everything you want to say?” (The teaching assistant ex-
consent to review their in-class materials, we eliminated one paper
plained that a good argument requires more than one exam-
from 2004 and two papers from 2005. Overall, there were nine
ple or experiment for supporting evidence.)
papers from 2004 and eight papers from 2005 in our analysis.
120
Journal of Engineering Education
April 2007

---

B. Instrument
sion of the course, it helps to rule out the possible explanation of
As in our earlier study, we used a scoring rubric to assess students’
practice effects accounting for differences. Additionally, it more re-
writing skills, but the rubric was more detailed. The rubric covered
alistically reflects how much writing, and thus improvement, can
seven areas: (a) organization and formatting, (b) mechanics, (c) style,
consistently be expected in a course such as this.
(d) clarity and quality of content, (e) synthesis, (f) argumentation,
and (g) visual thinking (charts and tables). For each of these seven
areas, papers written in groups were rated on a five point scale (1 to
V. FINDINGS
5). The maximum possible score for a paper was 35 and the
minimum score was seven.
Given the small sample size and resulting low statistical power,
1) Rationale for the rubric division: To maintain some continuity
ANOVA results were supplemented by measures of Effect Size
with the rubric of our previous study, we grouped the first
(ES), representing the standardized difference between groups in
three areas, which students seem to find easier to improve, as
standard deviation units. While it is advisable to interpret ES in
basic level writing skills and the later four, which seem to pre-
light of established norms in a particular domain, one suggested rule
sent more difficulty for students, as advanced level writing
of thumb suggests that 0.2 should be considered small, 0.5 consid-
skills.
ered medium, and 0.8 considered large effects [31]. ANOVA re-
2) Inter-rater reliability: To improve the inter-rater reliability of
sults revealed that students who received the 2005 intervention
the rubric, we put it through several iterations based upon a
performed better than students who received the 2004 intervention
small sample of the papers. Three researchers—a learning
in the following components of their written reports: “Organization
scientist, a cognitive psychologist, and a writing instructor—
and Formatting,” F(1, 16)
9.23, p
0.01, ES
1.50, “Style,”
in addition to another outside person whose field of study is
F(1,16)
5.31, p
0.05, ES
1.13, “Clarity and quality of con-
not engineering, blind-scored the papers. We also utilized
tent,” F(1, 16)
9.65, p
0.01, ES
1.53, and “Argumentation,”
the engineering instructor’s grading for the papers in our
F(1,16)
14.46, p
0.01, ES
1.85. Differences on one compo-
comparisons. Based on the weaknesses in some of the com-
nent, “Synthesis,” approached statistical significance but did not
parisons, we revised the sub-scores of each component in the
meet the p
0.05 minimum criterion, F(1, 16)
4.01, p
0.064,
rubric, adding more detailed explanations, and then re-scored
ES
0.99. Students’ paper scores were not found significantly dif-
the papers. With the revised rubric, we achieved a high corre-
ferent in the “Mechanics,” F(1,16)
0.27, p
0.61, or “Visual
lation (r
0.89) among the three researchers’ ratings. All
thinking (charts and tables),” F(1,16)
0.69, p
0.41, compo-
scoring was done blind to the class year of the paper.
nents of the rubric. Figure 1 represents the means of students’ paper
scores for each component of the rubric.
C. Analysis
After achieving a sufficient level of inter-rater reliability, two of
the researchers read the papers from 2004 and 2005 and blind-
VI. DISCUSSION
scored them. The correlations of those two researchers’ ratings
among the rubric components were between 0.71 and 0.96. Table 1
In 2004 and 2005, HPL strategies were used more deliberately
lists correlation coefficients. Because they were fairly high, we aver-
to help improve students’ engineering writing skills. As discussed
aged the two raters’ scores for our analysis.
earlier, the 2004 instruction encompassed an in-class presentation
We compared the averaged student paper scores of 2004 and
and guidance from the instructor. Students performed relatively
2005 data using the SPSS software package. Several analysis of
well after receiving this instruction, but they improved more in
variance measures (ANOVA) were computed to determine if the
“lower level” writing skills than in “higher level” skills such as syn-
student writing skills in 2005 differed from those in 2004. This is a
thesis and argumentation.
different strategy from the one described in Troy et al. [24] in that
Hence, in the design stages of the 2005 intervention, we focused
it compares groups of different students as opposed to the same
on the advanced level writing skills. In 2005, we implemented a dif-
students on two successive occasions. While this approach was
ferent strategy, described above, more fully informed by the How
taken simply because it was all that was possible, given the new ver-
People Learn framework [1]. Because our aim was to help students
improve advanced writing skills such as synthesis and argumenta-
tion, extra emphasis was placed on learner- and community-
centeredness in the instructional strategy to encourage student
engagement with the challenge and with each other.
Our results suggest that the “HPL informed approach” benefit-
ed students in a variety of ways. Although not all differences were
statistically significant, which should be expected from such a small
sample, improvements were seen on every dimension, including
lower level skills. The largest gains appear in the area of argumenta-
tion (ES
1.85) and the smallest in mechanics (ES
0.25). Gen-
erally, the larger gains are seen in the areas described above repre-
senting more advanced skills, with ES of 1 or higher on synthesis
and argumentation. This represents quite large differences between
Table 1. Correlations between rater 1 and rater 2 across the
years, with the high numbers suggesting the strength of the new
rubric components.
approach.
April 2007
Journal of Engineering Education
121

---

Figure 1. The mean scores of students’ papers across 2004 and 2005 (ES stands for Effect Sizes. “*” indicates that the p values were less than
0.05. “^” indicates that the p value was very close to 0.05).
We believe that the learner-centered and community-centered
our rubric prioritized students’ understanding of the mechanical
emphasis of the exercise was responsible for helping students
and conventional handling of graphics, which was not emphasized
overcome the difficulties posed by the challenging goal of devel-
in class and thus may partially account for the lack of improvement
oping advanced writing skills. Argumentation and synthesis seem
in this category.
to require illustrative examples and personal commitment. For ex-
Similarly, mechanics in writing were not stressed in class, and
ample, in order to synthesize the literature, students need to know
students may not have put much of an effort into making their final
the content very well but also to develop strategies for illustrating
drafts as professionally finished as possible, especially since they
ideas and statements in a concise and meaningful way. Without
would not be publishing these papers. If this hypothesis is true, then
knowing specific techniques, students may not be able to synthe-
the student behavior mimics the behavior of most professionals,
size the literature effectively regardless of their content mastery.
who do not overly worry about mechanics when they are not pub-
Yet, writing about that literature is an ideal way to learn how to
lishing their work. Thus, we do not find the lower numbers in me-
synthesize it because, as Emig [32] noted, writing is a unique way
chanics to be a concern.
of learning that is inherently integrative and requires the use of
both analysis and synthesis. Similarly, writing good arguments re-
quires not just knowledge of content but a convincing style aug-
VII. CONCLUSION
mented with the logic of argumentation. Role-playing, personal
commitment, and peer-to-peer learning (reader feedback) be-
In summary, this study suggests that an HPL approach to peda-
come crucial in writing good arguments. The techniques for visual
gogy in a WAC or WID context is just as beneficial to promote
thinking also require practical illustrations and perhaps a more
learning in writing as it has been shown to be for mastery of science,
focused instructional strategy, something that we may not have
mathematics, and engineering. As WAC researchers have shown,
emphasized sufficiently since only a modest improvement, which
and our previous experience [24] confirms, simply integrating writing
did not reach statistical significance, was observed in this area. Al-
instruction into a disciplinary course like systems physiology is insuf-
ternatively, it may be that our rubric was not well designed for
ficient to help students acquire the more advanced writing skills that
measuring growth in visual thinking or sufficiently coordinated
they will need as experts in a professional field. Rather, to teach high-
with our pedagogy.
er level skills like synthesis and argumentation, instruction must be
Perhaps visual thinking as a category needs to be subdivided,
learner-centered and community-centered. Moreover, writing in-
with one part measuring students’ understanding of conventions
struction benefits from the assessment-centered dimension of an
and best practices related to data graphics (e.g., where labels belong,
HPL approach. Assessment of student performance in the early years
line thickness, style of graphic presentation, how to discuss the
of the exercise laid the groundwork for formulating well-defined
point of a graphic in the text), and another part measuring whether
learning objectives. Analysis from the study of Troy et al. [24] showed
students understand when and how to use a graph to present evi-
us where our instructional intervention was failing, and analysis of the
dence in an argument or to write concisely about data that would be
papers from 2005 has shed considerable light on how we may need to
confusing if presented in text. If so, the first category might be
refine our teaching and assessment of visual thinking in writing.
thought of as a lower level, mechanical skill, whereas the second
Following are some lessons learned from the 2005 intervention
would be more closely related to clear writing and effective argu-
about teaching advanced writing skills:
mentation. Neither our instruction nor our analysis of student pa-
●
Faculty must do more than embed writing assignments into their
pers took this complexity about visual thinking into account. In fact,
class. They need to create challenge-based assignments, like
122
Journal of Engineering Education
April 2007

---

the scenario in our 2005 assignment, which are well suited to
[2] Bazerman, C., and D.R. Russell, (Eds.), Landmark Essays on
a student’s developmental abilities and sense of identity. By
Writing Across the Curriculum, Vol. 6, Davis, CA: Hermagoras Press,
the time students write papers in our upper level courses, they
1994.
have spent 14 years developing their identity as a student. If
[3] Townsend, M.A., “Writing Across The Curriculum,” In A.C.
we want them to develop the writing skills that are useful for
Purves, (Ed.), The Encyclopedia of English Studies and Language Arts, Vol. II,
a professional, we have to help them imagine themselves in a
1994, pp. 1299–1302, http://cwp.missouri.edu/cwpinfo/aboutwac.htm.
professional role. If we make the leap too great in an assign-
[4] Bazerman, C., J. Little, L. Bethel, T. Chavkin, D. Fouquette, and
ment, students will not be able to accept the challenge. If they
J. Garufis, Reference Guide to Writing Across the Curriculum, West Lafayette,
cannot imagine the assignment’s authenticity, they will fall
IN: Parlor Press, 2005.
back on what they do know: how to complete an assignment
[5] Galbraith, D., and G. Rijlaarsdam, “Effective Strategies for the
for a due date in school.
Teaching and Learning of Writing,” Learning and Instruction, Vol. 9,
●
The writing assignment must be given with time for students to
No. 2, 1995, pp.93–108.
receive feedback, reflect on their learning, and revise their drafts.
[6] McMillan, V.E., Writing Papers in the Biological Sciences, Boston,
Feedback need not always be written and need not be individ-
MA: Bedford Books, 1997.
ual, but it must be well timed. Students are most open to writ-
[7] Paradis, J.G., and M.L. Zimmerman, The MIT Guide to Science and
ing instruction—and most able to verbalize their preconcep-
Engineering Communication, 2nd ed., Cambridge, MA: The MIT Press,
tions and misconceptions—when they are in the middle of
2002.
working on a paper. In our 2005 intervention, students for-
[8] Miller, J.E., The Chicago Guide to Writing about Numbers, Chicago,
mulated their own questions about the writing assignment.
IL: University of Chicago Press, 2004.
We believe this had a direct bearing on the improvement we
[9] Hedengren, B.F., A TA’s Guide to Teaching Writing in All Disciplines,
noted in the papers of 2005 over the ones produced in 2004.
Boston, MA: Bedford/St. Martin’s, 2004.
●
Writing requires an interactive, coaching pedagogy—like that
[10] Fulwiler, T., “How Well Does Writing Across The Curriculum
advanced by HPL techniques—because writing mastery is both
Work?,” in C. Bazerman and D.R. Russell, (eds.), Landmark Essays on
conceptual and performance-based. Students do not really un-
Writing Across the Curriculum, Vol. 6, Davis, CA: Hermagoras Press, 1994,
derstand writing until they can do it, and “doing it” has just as
pp. 51–64.
much, if not more, to do with understanding key concepts
[11] Shuman, L.J., M. Besterfield-Sacre, and J. McGourty, “The
such as audience, purpose, and context as with mastering
ABET ‘Professional Skills’-Can they be taught? Can they be assessed?,”
skills in style, conventions, and technologies.
Journal of Engineering Education, Vol. 94, 2005, pp. 41–55.
We believe accomplishing the HPL framework with the three
[12] Accreditation Board for Engineering and Technology (2006),
strategies outlined above—the use of embedded assignments, ade-
Engineering Change: A Study of the Impact of EC2000, Executive Summary,
quate time for student reflection, and interactive and coaching ped-
http://www.abet.org/papers.shtml.
agogy for teaching writing—will encourage students to improve
[13] Hirsch, P., B. Shwom, J. Anderson, G. Olson, D. Kelso, and J.E.
their writing skills. Instructional designs similar to the one present-
Colgate, “Engineering Design and Communication: Jump-Starting the
ed in this paper can be implemented in other learning environ-
Engineering Curriculum,” Proceedings, ASEE Annual Conference and
ments. For example, in teaching research ethics in a science course,
Exposition, Seattle, WA, 1998.
students in groups can be confronted with real ethical dilemmas on
[14] Engineering Design and Communication (EDC), Northwestern
the use of human and/or animal subjects for scientific research and
University, http://www.edc.northwestern.edu.
asked to present their cases to one another in an interactive manner.
[15] Institute for Design and Engineering Applications (IDEA),
If students are given useful feedback and time to reflect upon it, they
Northwestern University, http://www.idea.northwestern.edu.
will most likely write more effective case summaries. Most impor-
[16] Cain Project in Engineering and Professional Communication,
tantly, writing instruction in upper level science and engineering
Rice University, http://www.owlnet.rice.edu/~cainproj/about.html.
courses can—and should—successfully target the skills and applica-
[17] The CDIO Initiative, http://www.cdio.org/index.html.
tions that students are unlikely to receive in general courses in
[18] VaNTH Engineering Research Center for Bioengineering Educa-
English or composition. Writing needs to be taught both by writing
tional Technologies, www.vanth.org.
instructors and writers in the disciplines.
[19] Thompson, N.S., E.M. Alford, C. Liao, R. Johnson, and M.A.
Matthews, “Integrating Undergraduate Research into Engineering: A
Communications Approach to Holistic Education,” Journal of Engineering
ACKNOWLEDGEMENT
Education, Vol. 94, No. 3, 2005, pp. 297–3007.
[20] Jensen, W., and B. Fischer, “Teaching Technical Writing
The VaNTH ERC is supported by the Engineering Research
Through Student Peer Evaluation,” Journal of Technical Writing and Com-
Centers Program of the National Science Foundation under Award
munication, Vol. 35, No. 1, 2005, pp. 95–100.
Number EEC-9876363.
[21] Winsor, D., Writing Like An Engineer: A Rhetorical Education,
Hillsdale, NJ: Erlbaum, 1996.
[22] Winsor, D., Writing Power: Communication in an Engineering Cen-
REFERENCES
ter, Albany, NY: State University of New York Press, 2003.
[23] McLeod, S.H., “Writing Across the Curriculum: The Second
[1] Bransford, J., A. Brown, and R. Cocking, How People Learn: Brain,
Stage, And Beyond,” in C. Bazerman, J. Little, L. Bethel, T. Chavkin, D.
Mind, Experience, and School, Washington, D.C.: National Academy Press,
Fouquette, and J. Garufis (eds.), Reference Guide to Writing Across the Cur-
1999.
riculum, West Lafayette, IN: Parlor Press, 2005, pp. 79–86.
April 2007
Journal of Engineering Education
123

---

[24] Troy, J., P. Hirsch, H.D. Smith, and B. Yalvac, “Team-Based Writ-
B.A. in English from the University of Michigan and an M.A. and
ten Communication Exercises for Biomedical Engineering Juniors: Where To
Ph.D. in English from Northwestern University.
Do It and What Works,” Proceedings, ASEE Annual Conference and Exposition,
Address: The Writing Program, Northwestern University, 1860
Salt Lake City, UT, June 2004, http://www.vanth.org/docs/Troy.pdf.
Campus Drive, Evanston, IL 60208; telephone: ( 1) 847.491.4969;
[25] Troy, J.B. and R.A. Linsenmeier, “Physiology Instruction in Bio-
fax: ( 1) 847.491.4840; e-mail: phirsch@northwestern.edu.
medical Engineering through the Vanth ERC: Optimizing the Delivery of
Content,” IEEE Engineering in Medicine and Biology Magazine, July/
H. David Smith is a senior lecturer in the Department of Psy-
August 2003, pp. 80–87.
chology at Northwestern University. He specializes in assessment
[26] Purves, D., G.J. Augustine, D. Fitzpatrick, W.C. Hall, A.S.
and evaluation of educational interventions. He has been on the
LaMantia, J.O. McNamara, and S.M.Williams, (Eds.), Neuroscience, 3rd
faculty at the University of Michigan and Middlebury College
ed., Sunderland, MA: Sinauer Press, 2004.
and holds a Ph.D. in Psychology from Virginia Commonwealth
[27] Moore, J.M., and A.E. Stuart, Neurons in Action, Version 1.43,
University.
Sunderland, MA: Sinauer Press, 2004.
Address: Department of Psychology, Northwestern University,
[28] Core Competencies, VaNTH ERC, 2006, http://www.vanth.org/
2029 Sheridan Road, #303C, Evanston, IL 60208; telephone: ( 1)
curriculum/curr_taxon_comm.asp.
847.491.4669; e-mail: hdsmith@northwestern.edu.
[29] Herrington, A.J., “Writing in Academic Settings: A Study of the
Contexts for Writing in Two College Chemical Engineering Courses,” in
John B. Troy, professor of Biomedical Engineering at Northwest-
C. Bazerman, J. Little, L. Bethel, T. Chavkin, D. Fouquette, and J. Garufis
ern University, chaired the undergraduate program there for 12 years
(eds.), Reference Guide to Writing Across the Curriculum, West Lafayette, IN:
and has taught Neural Systems Physiology for more than 15 years.
Parlor Press, 2005, pp. 97–121.
An active researcher in this area, he is domain leader for systems
[30] Ben-David Kolikant, Y., D. Gatchell, P. Hirsch, and R.A.
physiology in the VaNTH ERC. He holds undergraduate degrees
Linsenmeier, “A Cognitive-Apprenticeship-Inspired Instructional Ap-
from the Universities of Reading and London and a D.Phil. in Ex-
proach for Teaching Scientific Writing and Reading,” The Journal of College
perimental Psychology (Neuroscience) from the University of Sussex.
Science Teaching, Vol. XXXVI, No. 3, 2006, pp. 20–25.
Address: Biomedical Engineering Department, McCormick
[31] Cohen, J., Statistical Power Analysis for the Behavioral Sciences, 2nd
School of Engineering and Applied Science, 2145 Sheridan Road,
ed., Hillsdale, NJ: Lawrence Erlbaum Associates, 1988.
Evanston, IL 60208-3107; telephone: ( 1) 847.491.3822; fax:
[32] Emig, J., “Writing as a Mode of Learning,” in C. Bazerman,
( 1) 847.491.4928; e-mail: j-troy@northwestern.edu.
J. Little, L. Bethel, T. Chavkin, D. Fouquette, and J. Garufis (eds.), Refer-
ence Guide to Writing Across the Curriculum, West Lafayette, IN: Parlor
Bugrahan Yalvac is an assistant professor of Science Education
Press, 2005, pp. 89–96.
at Texas A&M University in the Department of Teaching, Learn-
ing, and Culture. Previously, he spent three years as a post-doctoral
research fellow in learning science for the VaNTH ERC at North-
AUTHORS’ BIOGRAPHIES
western University. He holds a Ph.D. in Science Education from
the Pennsylvania State University, and an M.S. degree from the
Penny L. Hirsch, Ph.D., is associate director of the Writing
Middle East Technical University (METU), Ankara, Turkey. His
Program at Northwestern University and a Charles Deering Mc-
specialty is learning environment design and research methods in
Cormick University Distinguished Lecturer. A principal in her own
science and engineering education.
communications consulting firm since 1986, she is also the
Address: 444 Harrington Tower, College Station, TX 77843;
VaNTH project leader for core competency instruction. She has a
telephone: ( 1) 979.862.1713; e-mail: yalvac@tamu.edu.
124
Journal of Engineering Education
April 2007

---

APPENDIX A
Dr. Wilson has asked your team to help her prepare her speech.
She has chosen you, a recent graduate in biomedical engineering
Scoring Rubric Used to Grade 16 Papers from 2003
from Midwestern University, to be on the team because of your
strong background in mathematics and engineering. What you are
Lower Level Skills
called upon to do is research the literature on neural coding, paying
●
Tables, Figures, Equations: yes/no
special attention to other models of neural coding than the rate
●
Organization
code. There is no need for you to search out and read Dr. Smythe’s
o Divides information into useful categories and/or para-
papers; Dr. Wilson herself will look them over.
graphs
o Effectively uses headings and sub headings
Even Number Groups
o Subdivides long sections
Meanwhile, over in Dr. Smythe’s lab, preparations for his
●
Grammar and Mechanics
speech are also underway. Dr. Smythe has long promoted the
o Proper referencing (in bibliography and body of text)
view that neural messages are encoded in patterns of action poten-
o Errors in usage and grammar
tial discharge, and it is the case for such coding that will form the
●
Style
core of his presentation at the National Academy. Dr. Smythe is
o Avoids repetition and narrative
very keen to make a strong impression at this meeting. He has
o Avoids lengthy sentences
spent a lot of time over the years courting National Academy
o Is easy to read
members and now that he feels he is on the verge of seeing the
o Uses words precisely
fruits of that effort, he does not want to let it all go to waste with a
Higher Level Skills
below par performance in the citadel he so earnestly wishes to
●
Synthesis
grace with his membership. Dr. Smythe has been irritated ever
o Draws on literature
since he found out that Dr. Wilson would be offering the oppos-
o Integrates ideas
ing viewpoint. While he is quite sure that he has the sharper
o Combines and summarizes evidence from other sources
mind, one must worry about having an off day.
●
Organization/Argumentation
You are also a recent graduate of Midwestern University, with a
o Makes a clear argument
B.S. in biomedical engineering. You chose Dr. Smythe’s lab be-
o Provides evidence for claims
cause it is renowned. Dr. Smythe has assigned you the task of mak-
o Considers alternate viewpoints
ing the case for the alternative rate coding hypothesis. He is deter-
o Provides rationale for paper
mined to be ready for whatever curve-balls Dr. Wilson might
●
Presentation of Information, Concepts, and Equations in
throw out in her presentation. He has not asked you to look at
Charts and Figures
Dr. Wilson’s papers. He will review them thoroughly himself, but
o Effectively utilizes multiple ways of presenting information
he would like you to research the rest of the literature for him. It is
important that he be seen to have a full grasp of the field as it
stands today.
APPENDIX B
Both Groups
Scenario and Assignment for the Team-Based Research Paper,
It does not matter whether you like Dr. Smythe more than
Fall 2005
Dr. Wilson or vice versa; you must act professionally and not let
your personal feelings weaken your resolve to do the best job you
The following story is fictional. It is an exaggerated but more or
can. A library of neural coding papers has been provided on the
less realistic example of a situation that could arise in the profession-
course Blackboard website that you can use. You are also encour-
al life of an academic biomedical engineer.
aged to search out your own sources. You are discouraged from rely-
ing too heavily on summaries you may find on the web. They often
Odd Number Groups
contain inaccuracies. People who are active in a research field rarely
The National Academy is sponsoring a workshop on the topic of
have time to write web tutorials. An important point of the exercise
Neural Coding in Washington, D.C. Your new advisor, Dr. Wilson,
is to help you become familiar with original source material. You
has published a number of papers advancing the case that the mes-
can download recent papers by accessing the library’s electronic
sages carried by neurons are encoded in their spike trains in terms of
resources (http://er.library.northwestern.edu), but you will have to
the rate at which they fire action potentials. She has been asked to de-
familiarize yourselves with the journal stacks in the library to find
liver a keynote speech making the case for neural rate coding. Al-
some of the classic papers.
though it is a great honor to be asked to make such a speech at such
an important event in such a prestigious location, Dr. Wilson is
Mechanics
anxious that she will not do herself justice and be embarrassed in
For the homework exercise, you have been assigned to a group
front of her colleagues. At the root of Dr. Wilson’s anxiety is the
that will write a report making the case either “for neural rate
fact that her main protagonist, Dr. Smythe, will also be at the meet-
coding” (even number groups) or “for spike pattern coding” (odd
ing, making another keynote speech advancing an alternative model
number groups). You should summarize the evidence in favor of
of how neurons encode information. Dr. Wilson is concerned be-
and against your position, making the strongest case you can for
cause Dr. Smythe is so respected in the field. After all, it is widely
your side of the argument. In reality, one would generally seek to
rumored that he will soon be inducted into the National Academy.
write a balanced view of the debate, but since we are seeking to
April 2007
Journal of Engineering Education
125

---

hone your ability to write a convincing argument, the emphasis of
2002). For two author papers, give both names. In the bibli-
the exercise is slanted in this direction. You have two weeks in
ography, all authors’ names should be given for every paper.
which to finish the assignment (due October 26). You will score
3) You can scan figures from books and papers (or download
very well on the paper if you have found some additional sources
pictures from the web) to include in your paper as long as
other than those you were given (new original papers counting
they are cited appropriately. Bear in mind though that you
more than information from the web) and have effectively syn-
should be using figures to make a point, not just to include
thesized information from those papers into your group’s argu-
them for their own sake.
ment for or against rate encoding. You will also be assessed on
On October 19, the full class-time will be used so that some
the completeness of your research, the strength of your argu-
groups (drawn at random) can make preliminary reports to the class
ments, the paper’s overall organization, its clarity and grammati-
on the case for and against rate coding.
cal correctness, and how effectively you have used tables and dia-
Posted on Blackboard you should find some of the papers you
grams to communicate ideas that are difficult to explain with
are assigned. The assigned papers are listed in the document titled
words alone.
Homework 2 2005 Neural Coding Papers. All groups are expected
Instructions for the paper
to read the general papers and the papers assigned to their group.
1) The paper should be written so that it has no more than five
To access some of the papers, you will need to use the journal stacks
pages of double-spaced lines of text with one inch margins
in Northwestern University’s Library. This is intentional. One ob-
top and bottom, left and right. The five-page limit does not
jective of the exercise is to familiarize you with accessing original
include a bibliography at the end, figures or tables. A title
papers. You should be able to determine how to do this yourselves.
page does not count against the page limit. Use Arial 11 point
Also, posted on Blackboard are the following supplementary
as the font.
documents, intended to help you with technical writing:
2) Use a consistent form of referencing when citing papers.
●
Rationale, Advice, and Scoring–Homework 2 2005
When you refer to a multi-author paper, use either all names
●
Supplemental Document (Writing for Excellence)
or et al. (e.g., Kumar, Smith, and Wu, 2002 or Kumar et al.,
●
Guidelines for Writing an Effective Group Report
126
Journal of Engineering Education
April 2007

---